Search is not available for this dataset
article
stringlengths 4.36k
149k
| summary
stringlengths 32
3.35k
| section_headings
listlengths 1
91
| keywords
listlengths 0
141
| year
stringclasses 13
values | title
stringlengths 20
281
|
|---|---|---|---|---|---|
Fusion proteins composed of the histone methyltransferase mixed-lineage leukemia ( MLL ) and a variety of unrelated fusion partners are highly leukemogenic . Despite their prevalence , particularly in pediatric acute leukemia , many molecular details of their transforming mechanism are unknown . Here , we provide mechanistic insight into the function of MLL fusions , demonstrating that they capture a transcriptional elongation complex that has been previously found associated with the eleven-nineteen leukemia protein ( ENL ) . We show that this complex consists of a tight core stabilized by recursive protein–protein interactions . This central part integrates histone H3 lysine 79 methylation , RNA Polymerase II ( RNA Pol II ) phosphorylation , and MLL fusion partners to stimulate transcriptional elongation as evidenced by RNA tethering assays . Coimmunoprecipitations indicated that MLL fusions are incorporated into this complex , causing a constitutive recruitment of elongation activity to MLL target loci . Chromatin immunoprecipitations ( ChIP ) of the homeobox gene A cluster confirmed a close relationship between binding of MLL fusions and transcript levels . A time-resolved ChIP utilizing a conditional MLL fusion singled out H3K79 methylation as the primary parameter correlated with target expression . The presence of MLL fusion proteins also kept RNA Pol II in an actively elongating state and prevented accumulation of inhibitory histone methylation on target chromatin . Hox loci remained open and productive in the presence of MLL fusion activity even under conditions of forced differentiation . Finally , MLL-transformed cells were particularly sensitive to pharmacological inhibition of RNA Pol II phosphorylation , pointing to a potential treatment for MLL . In summary , we show aberrant transcriptional elongation as a novel mechanism for oncogenic transformation .
Mixed-lineage leukemia ( MLL ) is a particularly aggressive subtype of acute leukemia with a very dismal prognosis . This disease is caused by chromosomal aberrations , mostly translocations , affecting Chromosome 11 at band q23 . This chromosomal locus contains the gene for the histone H3 lysine 4–specific methyltransferase MLL [1]–[4] . As a corollary of these genomic rearrangements the 5′ portion of MLL is fused in frame to a variety of different and mostly unrelated partner genes . The translation of the chimeric RNAs transcribed from the altered locus results in the production of fusion proteins . In these fusions , the original MLL methyltransferase activity is replaced by biological properties provided by the fusion partner . This creates novel oncoproteins that are potently transforming hematopoietic cells ( for reviews , see [5]–[7] ) . MLL fusions are aberrant transcription factors that induce ectopic expression of their respective target genes , and as a consequence , they block hematopoietic differentiation . Critical targets for MLL-induced transformation are the clustered HOXA homeobox genes and the gene for the HOX-dimerization partner MEIS1 [8] . Accordingly , a relative overexpression of HOXA and MEIS1 transcripts is the characteristic hallmark of the MLL-specific gene expression profile [9] , [10] . Despite this predominance of HOX expression , however , it has been shown by genome-wide chromatin precipitations that MLL fusion proteins occupy several thousand binding sites [11]–[13] . As it has been noted some time ago , transcriptional activation by MLL fusions is accompanied by a conspicuous and dramatic increase of histone H3 lysine 79 dimethylation across the HOXA locus [14] , and this phenomenon has been confirmed also for many of the other MLL fusion target loci [12] . The only known histone methyltransferase that is capable of introducing the H3K79 mark is DOT1L , a protein conserved from yeast to man [15] , [16] . Indeed , it could be shown for the MLL fusion partner AF10 that a direct interaction with DOT1L was instrumental for the oncogenic function of the respective fusion protein [17] . First hints for a shared function of several MLL fusion partners came from studies performed by Bitoun et al . [18] . These authors conducted overexpression studies and published data to support a model of multiple MLL fusion partners being involved in a transcriptional elongation complex that includes the MLL partner proteins AF4 , AF9 , ENL , and AF10 , as well as DOT1L , and positive transcription elongation factor b ( pTEFb ) . A direct interaction between proteins of the AF4 and AF9/ENL families had been noted before by Erfurth et al . [19] , as well as our own group [20] . A somewhat contradictory interaction of AF9 and DOT1L has also been described to be necessary for aldosterone-induced gene silencing [21] . To elucidate the function of normal ENL , we recently purified wild-type ENL from mammalian nuclei [22] . It could be shown that endogenous ENL was also present in a large macromolecular protein complex similar to the one described by Bitoun et al [18] . Although the complex was initially termed ENL associated proteins ( EAP ) , we now propose to redefine EAP as “elongation assisting proteins” to better reflect the function of this protein complex . In addition to DOT1L , EAP contained pTEFb , a cyclin-dependent kinase 9/cyclin T dimer that phosphorylates the C-terminal repeat domain ( CTD ) of RNA Polymerase II ( RNA Pol II ) [23] . CTD phosphorylation is necessary to ensure productive transcriptional elongation . Next to AF4 , the AF4 homologs AF5 and LAF4 were also present in EAP , confirming the results of Estable et al . [24] , who had copurified AF5 with pTEFb . AF4 itself is the most frequently encountered MLL fusion partner , and in a recent survey , approximately 50% of all MLL cases in infants and adults carried a MLL-AF4 translocation [25] . EAP was ubiquitously expressed , and interference with EAP assembly affected transcriptional elongation of many genes . However , it was not clear whether EAP activity was important for the respective MLL fusion proteins . In the fusion , a bulky 180-kDa MLL moiety is added to an ENL protein of approximately 70 kDa . This type of modification might substantially alter or even destroy the EAP complex . Here , we present a comprehensive picture of MLL fusion biology , demonstrating that EAP has a very stable core that is capable of also accommodating MLL fusion proteins . The constitutive recruitment of EAP to MLL target loci is responsible for persistent target transcription through stimulation of transcriptional elongation . This mechanism resists differentiation stimuli and therefore causes a maturation arrest . Finally , MLL fusion transformed cells were sensitive to EAP inhibition , pointing to a potential lead for pharmaceutical intervention .
In previous studies , the total molecular weight of all proteins coprecipitating with ENL amounted to more than 1 MDa , whereas the bulk of ENL eluted on sizing columns with an apparent molecular weight of approximately 400 kDa to 500 kDa [22] . To explain this discrepancy and to further elucidate the molecular architecture of the EAP assembly , we performed two-hybrid assays to test for mutual protein–protein interactions . A large deletion series of existing [20] , [22] , [26] , [27] and newly constructed two-hybrid bait clones for ENL , AF4 , CYCT2A—the cyclin component of pTEFb—and Dot1l was probed for interaction with full-length versions of the same proteins . As reported previously [22] , only the mouse homolog of DOT1L was available in cDNA repositories , and therefore , mouse Dot1l was used throughout this study . A total of 78 potential interaction pairs were interrogated ( Figure S1 ) . These experiments showed that EAP contained a tight core stabilized by a recursive set of direct protein–protein interactions ( Figure 1 ) . Each protein tested was able to interact with two other proteins , thus linking ENL , AF4 , Dot1l , and CYCT2/CDK9 ( pTEFb ) in a tight “circular” network . In this way , histone H3 methylation catalyzed by Dot1l can be coordinated with RNA Pol II phosphorylation introduced by pTEFb . The total calculated molecular weight of the EAP core components was 481 kDa , and this number was very close to the previously determined [22] value for the EAP complex eluting from gel filtration . Of note , ENL , Dot1l , and CYCT2 utilized a single domain to interact with both of their binding partners , whereas the two binding interfaces were separated in AF4 . The AF4 N-terminal homology domain provided contact to CYCT2 , whereas sequences further C-terminal formed the interface with ENL . It is important to note that the respective interaction domains are highly conserved between the homologous MLL fusion partners ENL and AF9 and between AF4 and the related AF5 , LAF4 , and FMR2 proteins as well . In two-hybrid assays , AF4 sequences could be replaced with the corresponding AF5 regions , yielding identical results ( unpublished data ) . In the cellular environment , EAP , therefore , is likely to exist in different configurations , explaining the large number of proteins that have been identified in ENL precipitates . Next , we wanted to know whether recruitment of MLL fusion partners to specific genes would promote transcriptional elongation . For this purpose , we used an RNA tethering assay to detect elongation activity . This test places a luciferase reporter gene downstream of a modified HIV-1 LTR promoter that grafts the stem loop IIb from the HIV-1 Rev response element ( RRE ) onto the TAR ( transactivation response RNA ) double-stranded RNA ( Figure 2A ) . RNA Pol II stalls after the TAR element . LTR regulation is achieved by binding of the transactivator TAT that regulates promoter output via recruitment of pTEFb to stimulate elongation [28] . The hybrid IIb/TAR loop allows tethering of any protein of interest to the LTR RNA by fusing it to the RNA binding protein Rev . Luciferase levels , therefore , will reflect the ability to recruit pTEFb elongation activity [29] . Rev fusions with ENL , AF5 , AF4 , or deletion derivatives of these proteins were transiently expressed in 293T cells in the presence of the TAR/IIb luciferase reporter . Correct expression was verified by an anti-Rev immunoblot ( Figure 2B ) . Rev alone and a Rev-CDK9 chimera served as negative and positive controls , respectively . Because pTEFb is ubiquitously expressed , 293T cells are a suitable environment for this assay . Attaching ENL to Rev induced an approximately 6-fold increase in luciferase levels comparable to the effect of a Rev-CDK9 fusion ( Figure 2C ) . Because ENL does not directly interact with pTEFb ( compare to Figure 1 ) , this contact must have been made through endogenous AF4 , Dot1l , or both proteins , respectively . Consequently , small deletions in the C-terminal AF4/Dot1l interaction domain eliminated Rev-ENL activity . As expected , Rev-AF5 and Rev-AF4 induced overall higher luciferase outputs because these molecules should be able to recruit pTEFb directly through CYCT binding and , in addition , indirectly via ENL . Deletion of the CYCT binding domain in AF5 should allow EAP interaction only through ENL . Indeed , a corresponding Rev-AF5 mutant had reduced luciferase activities comparable to the values achieved with Rev-ENL alone . No effect on elongation could be recorded with AF5 lacking both CYCT and ENL interaction regions . These results provided strong evidence that recruitment of MLL fusion partners induced elongation activity . Next , we tested whether the elongation activity of ENL and AF5 persisted after fusion with MLL . MLL-ENL and MLL-AF5 , as well as two mutants with a deletion in the respective EAP interaction domains , were joined to Rev and tested in the RNA tethering assay ( Figure 2D ) . In these experiments , Rev-MLL-ENL and Rev-MLL-AF5 could activate the luciferase reporter to a similar degree as Rev-CDK9 . In contrast , luciferase levels were close to background for the Rev-MLL fusions that had lost the capability to recruit EAP . This indicated a functional association also of MLL fusion proteins with EAP . In the past , it has been noted by So and Cleary [30] and by our group [31] that a heterologous fusion of MLL with the strong transactivator VP16 had transforming ability . In contrast , chimeras of MLL and the acidic transactivation domain AD42 ( derived from a mammalian two-hybrid system ) had no oncogenic activity despite the fact that AD42 was a more powerful transactivator than ENL . Later , it was shown that VP16 recruits pTEFb [32] , whereas no such activity is known for AD42 . To further strengthen the correlation between elongation and functional MLL fusion proteins , we determined the overall transactivation capability of ENL , AD42 , and VP16 in a conventional GAL4-based reporter assay and compared it to the elongation activity as Rev fusion ( Figure S2 ) . GAL4-ENL was 30-fold weaker than GAL4-VP16 and 5-fold less active than GAL4-AD42 in SV40 core promoter-based reporter assays . In stark contrast , Rev-ENL induced almost the same elongation activity as Rev-VP16 on the TAR-reporter , whereas Rev-AD42 showed almost no elongation stimulation in this test . MLL fusion proteins add a large 180-kDa MLL moiety to the respective fusion partner . Therefore , it was not clear whether these huge proteins could be accommodated within the EAP core . To answer this question , a series of immunoprecipitations were performed . Because sensitive antibodies that detect their cognate antigen at endogenous levels were only available for ENL and CDK9 ( pTEFb ) , HA-tagged versions of Dot1l , AF4 , and AF5 were utilized for these experiments . MLL-ENL was transfected either alone or together with HA-Dot1l , HA-AF4 , or HA-AF5 into HEK293T cells . MLL without any fusion partner and a MLL-ENL variant lacking the last 15 amino acids of ENL ( MLLENL1–544 ) served as controls . As shown before , this deletion prohibited interaction of ENL with Dot1l in two-hybrid tests and abolished ENL-mediated elongation activity in the RNA tethering experiments . Western blots proved all MLL fusion derivatives to be correctly expressed ( unpublished data ) . Precipitations were done employing an anti-MLL antibody [33] recognizing the MLL N-terminus retained in the fusion proteins ( Figure 3A ) . MLL-ENL coprecipitated with HA-AF4 , HA-AF5 , HA-Dot1l , and notably also with endogenous CDK9 . Because there is no direct interaction of ENL with CDK9 or CYCT ( see Figure 1 ) , MLL-ENL most likely had to be associated also with endogenous AF4/DOT1L to bring down CDK9 . In line with the two-hybrid and RNA tethering results , the MLL-ENL1–544 mutation eliminated coprecipitation with HA-Dot1l and CDK9 , but still allowed some residual interaction with AF4 . Interestingly , this was not true for AF5 as this protein could not be detected in MLL-ENL1–544 precipitates . No protein precipitated with the N-terminus of MLL; therefore , all interactions must have been mediated by the respective fusion partner . As a control , all immunoprecipitates were also checked for the presence of the respective MLL fusion by an MLL-specific Western blot . In a second series of immunoprecipitations , we concentrated on the interaction of MLL-AF4 and MLL-AF5 with endogenous proteins ( Figure 3B ) . MLL-AF4/5 fusions that occur “naturally” in leukemia join MLL to a C-terminal portion of AF4/5 . Therefore , these proteins do not contain the N-terminal cyclin interaction domain of AF4/5 , but they retain the ENL interaction motif . MLL-AF4 and MLL-AF5 fusions built analogous to the patient-derived proteins ( MLLAF4758–1210 , MLLAF5731–1163 ) were expressed in HEK293T cells . Shortened constructs deleting also the respective ENL interaction domains ( MLLAF41023–1210 , MLLAF5991–1163 ) served as controls . MLLAF4758–1210 and MLLAF5731–1163 both efficiently coprecipitated with endogenous ENL and CDK9 . This interaction was not mediated by the MLL portion of the fusion , because the control proteins lacking the ENL binding domain were not capable of bringing down ENL or CDK9 . In summary , these results provided proof that despite their considerable size , MLL fusions could be accommodated within the EAP core complex without disturbing the stabilizing protein interaction network . To confirm the incorporation of MLL fusions into EAP also in authentic leukemia cells , the immunoprecipitation experiments were repeated with SEM cells , a B-ALL line transformed by MLL-AF4 [34] . Lymphoid REH cells without 11q23 translocation served as control ( Figure 4A ) . Anti-MLL precipitates from SEM contained ENL and CDK9 , corroborating the association of MLL fusion proteins with EAP . Because the MLL-AF4 protein from SEM cells does not encompass the CYCT interaction motif of AF4 , the coprecipitation of CDK9 and MLL-AF4 strongly suggests an indirect bridging of these proteins by ENL and DOT1L . An association with the N-terminal MLL moiety or a nonspecific binding to the immunoprecipitation reagents seemed unlikely , as precipitates from REH cells done under identical conditions were devoid of ENL and CDK9 . The recruitment of EAP by MLL-AF4 should be accompanied by a higher concentration of EAP components on target chromatin . To test this prediction , the distribution of ENL across the human HOXA9 locus , a known MLL target gene , was determined by ChIP in SEM and REH control cells ( Figure 4B ) . Indeed , a significantly higher amount of ENL could be detected across the transcribed region of the HOXA9 gene in SEM versus REH . This correlated well with an approximately 20-fold increased concentration of HOXA9 RNA in SEM cells compared to REH controls . To study the consequences of MLL fusion-mediated EAP recruitment for target chromatin , we first determined the binding sites of an MLL-ENL fusion across the HoxA locus by chromatin immunoprecipitation ( ChIP ) and hybridization to genomic arrays ( ChIP-chip ) . For this purpose , MLL-ENL-transformed cell lines were generated from primary murine hematopoietic cells by transduction with a flag-tagged version of MLL-ENL . These fMLL-ENL cells were used as starting material for a flag-specific ChIP to avoid cross detection of endogenous wild-type Mll or Enl . Precipitates were amplified by ligation-mediated PCR and hybridized to commercial promoter arrays that tile 2 . 5 kb of genomic sequence upstream and downstream of the start of all known reading frames . In addition , the expression level of every single Hoxa gene was determined by quantitative reverse-transcriptase PCR ( qRT-PCR ) ( Figure 5 ) . With the exception of Hoxa2 and Hoxa13 , all Hoxa genes could be detected in fMLL-ENL cells with expression levels in the order Hoxa6/11>Hoxa5/7/9/10>Hoxa1/3≫Hoxa4 . A close correlation was observed between fMLL-ENL bound upstream of the individual Hoxa genes and the presence of the respective transcript , suggesting an involvement of the fusion protein in control of Hoxa transcription . To get further insight into the molecular mechanism of gene regulation by MLL-ENL , we analyzed the genomic region upstream of Hoxa9 , including the newly identified gene for microRNA196b [35] , by a time-resolved ChIP . Primers were designed binding upstream of Mirna196b and at the 5′ as well as the 3′ ends of the first exon of Hoxa9 . ChIP was done with a cell line transformed by a conditional version of MLL-ENL [8] . In these cells , MLL-ENL is fused to a mutated estrogen receptor ligand binding domain . As a consequence , the oncogene is only active in the presence of the inductor tamoxifen ( TAM ) . Removal of TAM leads to a loss of MLL-ENL binding within 72 h , down-regulation of Hox gene expression , cellular differentiation , and growth arrest [8] , [14] . Approximately 2 wk after withdrawal of TAM , the cultures consisted mainly of mature granulocytes and macrophages ( Figure 6A ) . The kinetics of Hoxa9 transcript levels , H3K79 dimethylation , RNA Pol II occupancy , and the presence of inhibitory H3K9/H3K27 methylation after MLL-ENL shut-down was determined by ChIP and qRT-PCR ( Figure 6B ) . In the presence of MLL-ENL ( time point 0 days ) , activating H3K79 dimethylation of Mirna196b was 50-fold higher and repressive H3K9 dimethylation was 2 . 6-fold lower compared to a heterochromatic , nontranscribed satellite locus on the X chromosome . Loss of MLL-ENL function was followed by a reduction of Hoxa9 transcripts to approximately 20% within 3 d , and a further drop below detection threshold was observed at day 10 . Most strikingly , the decrease in Hoxa9 transcripts was exactly replicated by H3K79 dimethylation , but not by RNA Pol II occupancy . Whereas H3K79 dimethylation was removed within 3 d , RNA Pol II did not exit the locus till day 10 after TAM withdrawal . This observation strongly suggests that Pol II became unproductive in the absence of active MLL-ENL . Inhibitory H3K9 and H3K27 methylation could only be detected at the Mirna196b locus after prolonged differentiation for 14 d . The transcriptional landscape of the Hoxa locus is complex , and it is not known where the Hoxa9 transcript is initiated and where it terminates . Nevertheless , we used the currently available information ( Hoxa9 Ensembl mRNA: ENSMUST00000048680 ) to design primer pairs at the utmost 5′ and 3′ ends of this putative core transcript . The ChIP kinetics was repeated with antibodies specific for the serine-2 and serine-5 phosphorylated isoforms of RNA Pol II . Furthermore , mRNA levels were quantified by RT-PCR as a function of time at the 5′ and 3′ ends of the transcript and in addition with primers spanning the Hoxa9 intron ( Figure S3 ) . Serine-2 phosphorylated ( elongating ) RNA Pol II decreased faster at the 3′ than at the 5′ end of the chromatin corresponding to the Hoxa9 core transcript . There was no significant difference in the mRNA decay kinetics if measured at the 5′ or 3′ terminus . However , the amount of spliced RNA as detected by intron-spanning primers leveled off more rapidly . This would be consistent with the known function of serine-2 phosphorylated RNA Pol II as “landing pad” for RNA processing enzymes . Finally , serine-5 phosphorylated ( initiating ) RNA Pol II stayed relatively constant with a tendency to exit first at the 5′ and later at the 3′ end . Whereas all these results are consistent with a function of pTEFb in MLL-ENL-mediated Hoxa9 activation , a more far-reaching interpretation will have to await the exact knowledge of all existing Hoxa9 transcripts . A second cell line model was used to confirm that MLL-ENL rather than cellular differentiation is responsible for the observed changes at the Hoxa9 locus . It had been shown previously that treatment with G-CSF induces differentiation even in the presence of constitutively active MLL-ENL [36] , [37] . MLL-ENL-transformed cells cultured in G-CSF will therefore allow separating the effect of MLL-ENL on the Hoxa9 locus from the influence of cellular differentiation . For this purpose , H3K79 dimethylation and Hoxa9 expression were determined in primary cells transduced by MLL-ENL subjected to G-CSF treatment . These data were compared to those measured in MLL-ENL-ER cells after MLL-ENL shutdown . Cells transformed by constitutive MLL-ENL stopped proliferation and induced gr-1 lineage marker expression after 7 d of G-CSF treatment to a level comparable to conditional MLL-ENL-ER cells 3 d after TAM withdrawal ( Figure 7A and unpublished data ) . Despite these clear signs of differentiation , Hoxa9 levels in G-CSF-cultured cells remained almost stable , and H3K79 dimethylation even increased slightly ( Figure 7B ) , proving that MLL-ENL is directly responsible for these effects and that this molecule is able to override differentiation stimuli . All results obtained so far indicated that MLL fusion proteins transform through recruitment of the EAP-associated enzymatic activities . Therefore , MLL cells might be sensitive to a pharmacologic inhibition of EAP . To test this prediction , the proliferation of six MLL cell lines and four controls of different etiology was recorded in the presence of increasing concentrations of flavopiridol and alsterpaullone , two substances with known CDK inhibitory activity [38] , [39] ( Figure 8 ) . The study was restricted to CDK inhibition as currently , there is no H3K79 methyltransferase inhibitor available . A murine cell line experimentally transformed by MLL-ENL and the corresponding parental primary cells were also included in the assay because patient lines might have accumulated unknown additional mutations that render the cells more resistant to EAP inhibition . Plotting proliferation against inhibitor concentrations clearly separated the cells into a sensitive and a more resistant class with a cutoff value for the two groups at 50% inhibitory concentrations ( IC50s ) of approximately 80 nM for flavopiridol and 1 µM for alsterpaullone . Although two MLL lines fell within the more resistant group ( RS4;11 and SEM for flavopiridol; HB11;19 and SEM for alsterpaullone ) , the majority of MLL fusion-transformed cells reacted significantly more sensitively than the controls . MLL-ENL-transformed primary cells had anIC50 of 50 nM for flavopiridol and 0 . 3 µM for alsterpaullone , whereas nontransduced primary bone marrow cells grown in liquid culture had significantly higher ID50 values of approximately 100 nM for flavopiridol and 1 µM for alsterpaullone ( Figure 8C ) . This confirmed that MLL-transformed cells are particularly sensitive to these substances .
In this report , we present evidence that the most frequently occurring MLL fusion proteins exploit molecular control mechanisms of transcriptional elongation to transform hematopoietic cells . MLL fusions become incorporated into an “elongation assisting protein” complex , recruit it to their respective target genes , and enforce ectopic transcription . This is accompanied by DOT1L-mediated H3K79 methylation and Pol II phosphorylation through the pTEFb kinase ( Figure 9 ) . This mechanism explains and reconciles seemingly contradictory observations that have been made previously with respect to MLL fusion proteins . It has been noted that particular MLL fusion partners are almost exclusively encountered in MLL with more lymphoid characteristics , whereas others occur preferentially in the myeloid subtype . For example , MLL-AF4-transformed cells are very often of lymphatic nature . In contrast , MLL-AF9 leukemia cells are myeloid , and MLL-ENL is found in ALL , AML , and also in T-cell acute leukemia [25] , [40] . These divergent phenotypes have been used as an argument against a common function for MLL fusion partners . However , the particular core structure of EAP that is stabilized by protein–protein contacts of conserved interaction domains , allows a high degree of flexibility . There are four members of the AF4 family ( AF5 , LAF4 , and FMR2 ) . ENL is closely related to AF9 and two CyclinT molecules ( CYCT1 and CYCT2 ) exist in the cell . Incorporation of different homologs into the same framework might create variations of EAP that provide for cell-type or target gene specificity . Although it is generally assumed that all MLL fusions occupy identical targets , the preexisting protein environments will vary at different loci . A co-recruited EAP complex incorporating AF9 might engage in protein interactions different from those established by an EAP variant containing ENL . As a consequence , the final level of target gene activation could be dependent on the composition of the EAP . The results presented here demonstrate how the makeup of EAP is determined by the nature of the MLL fusion partner . For example , all patient-derived MLL-ENL and MLL-AF9 fusions retain the conserved C-terminus of ENL/AF9 [41] , [42] that allows simultaneous recruitment of DOT1L and AF4 ( or any AF4 family member ) that both bridge to pTEFb . On the contrary , naturally occurring MLL-AF4/5 fusions have lost the direct pTEFb interaction domain in the N-terminus of AF4/5 [43] and need to rely on a more indirect way via ENL and DOT1L to bring in pTEFb . Structural variations in EAP and the mode of recruitment likely contribute to the observed differences in the MLL phenotypes . As suggested by coimmunoprecipitations and RNA tethering , all protein–protein interactions that stabilize EAP seem to be conserved also in the fusion context despite the addition of an 180-kDa MLL moiety . This is corroborated also by the fact that introduction of small peptides blocking the AF4–AF9 interface has been found to be specifically toxic for MLL-AF4 cells but much less so for leukemic blasts of different etiology [44] , [45] . MLL-AF4 requires AF9 ( or potentially ENL ) as a mediator to recruit DOT1L and pTEFb , and this pathway is blocked by binding site mimetics . Both the positive readout in RNA tethering assays and the ChIP results indicate that MLL fusion proteins affect transcription through stimulation of elongation . In this regard , it is interesting to note that the ELL protein , the first MLL fusion partner with a known biochemical function , also is an elongation factor [46] . Later , elongation was dismissed as biochemical basis for MLL-ELL-mediated transformation because motifs in ELL important for elongation activity could be deleted in MLL-ELL with no effect for the transforming function of the protein [47] . However , it was never thoroughly tested whether domains in ELL that are essential for transformation might recruit other elongation stimulating proteins . In this regard , it will be interesting to see whether protein interaction partners of ELL [48] , [49] will provide a link to elongation control . Strikingly , these ELL-associated factors ( EAF1 and EAF2 ) have a limited but significant homology to domains in AF4 [49] . Traces of ELL have been detected in ENL precipitates [22] , a possible lead that should be further explored . At present , it is hard to predict whether the more rare fusion partners will be connected to elongation control , too . This seems rather unlikely , because these proteins are mostly cytoplasmatic . However , it has been shown that the MLL fusion partner ABI1 , normally also in the cytoplasm , is imported into the nucleus as MLL-ABI1 fusion due to the strong nuclear localization signals of MLL . There , ABI1 can directly interact with ENL [27] , pointing to a mechanism for how cytoplasmatic fusion partners might also link to EAP and elongation control . After initial reports to the contrary [50] , it is well established that methylation of H3K79 by DOT1L is tightly associated with actively transcribed chromatin [15] . Until now , DOT1L had been implicated only in the transforming mechanism of MLL-AF10 where it could be demonstrated that interaction with DOT1L was essential for oncogenic activity of the MLL-AF10 fusion protein [17] . Here , we demonstrate a participation of DOT1L in a much wider range of MLL abnormalities encompassing the majority of all clinically observed cases . The incorporation of DOT1L in EAP also provides a molecular explanation for the genome-wide correlation of MLL-AF4 binding and a drastic increase of H3K79 methylation at the corresponding loci , a fact that has raised much interest recently [11] , [12] . In addition , we show that H3K79 methylation is highly dynamic and that it is correlated with target RNA abundance . It will be interesting to know how this methyl mark is removed after MLL-ENL inactivation because no H3K79-specific demethylase has been described so far . MLL fusion proteins are able to override normal differentiation stimuli as demonstrated by the continuing Hoxa9 target expression and the persistent H3K79 modification of the respective locus even in cells forced to differentiate . This characterizes MLL fusions as typical “class II” oncogenic effectors that block normal maturation of precursor cells [51] . An inactivation of the fusion protein itself by pharmacological means is difficult . An inhibition of the enzymatic activities in EAP by small molecules might be a more feasible treatment option . The experiments presented here clearly show that transformation by MLL-ENL sensitizes hematopoietic cells to the effects of CDK inhibitors . This sensitivity persists in several MLL patient cell lines even after prolonged culture in vitro . In this regard , it is interesting to note that a recent report by Cleary and colleagues ( Wang et al . [52] ) postulated an essential role of GSK3ß for MLL fusion-mediated leukemogenesis . This is paradoxical as GSK3ß normally acts as a tumor suppressor that inactivates the Wnt pathway [53] . Therefore , GSK3ß inhibition would be expected to exacerbate the transformed phenotype . However , GSK3ß shares a 30% homology with CDK9 ( unpublished data ) and pharmacological GSK3ß inhibitors often also target CDKs and vice versa [54] . With regard to the involvement of CDK9 in the biochemical mechanism of MLL fusion proteins , it seems likely that at least part of the GSK3ß effect might be attributed to a concomitant block of CDK9 activity . Regardless of the contribution of each pathway , our experiments show a promising new strategy to find rational treatments for this devastating disease .
The cDNAs coding for either full-length or ENL mutants ( also known as MLLT1 , accession number NM005934 ) , AF4 ( AFF1 , accession number NM_005935 ) , AF5 ( AFF4 , accession number NM_014423 ) , CYCT2 ( CCNT2 accession number NM_001241 ) , and Dot1l ( accession number NM_199322 ) were introduced into the vectors pGADT7 or pACT2 as GAL4-activation fusions and into pGBKT7 or pAS2-1 as bait fusions ( all vectors from Clontech ) . The human DOT1L cDNA is unavailable in the repositories . Therefore , the highly homolog mouse Dot1l was used . For RNA tethering assays , Rev fusions were constructed in pcDNA3 . 1-V5-HisTopoRev ( gift of M . Peterlin , San Francisco , California ) and tested with the pGL3-HIV1-LTR luciferase reporter plasmid . In addition , the cDNAs of ENL , AF4 , AF5 , CYCT2 , and Dot1l were inserted into the general expression vectors pcDNA3 . 0 ( Invitrogen ) and pMSCVneo ( Clontech , TaKaRa ) . MLL-AF5 was a kind gift of Eric So ( London , United Kingdom ) . MLL-ENL and derivatives have been published previously [55] . Antibodies used for ChIP were from AbCam . The monoclonal anti-MLL IgM ( HRX107 ) antibody [33] is available from Santa Cruz Biotechnology . The ENL antibody has been described [22] . All cell lines were obtained either from the DSMZ or from laboratory stocks . For the yeast two-hybrid experiments , the bait and target plasmids were cotransfected into the yeast strain AH109 ( Clontech ) according to the instructions of the manufacturer ( Yeast protocols handbook , Clontech ) . Interactions were assessed by activation of the his and ade reporter genes on synthetic media missing histidine and adenine , respectively . For a detailed description of the yeast two-hybrid procedure , see [56] . RNA tethering assays were done in 293T cells . Luciferase reporter pGL3-HIV1-LTR ( 0 . 1 µg ) was cotransfected with 0 . 9 µg of pcDNA3-based Rev fusion plasmid by standard lipofection . Twenty-four hours after transfection , cells were lysed and luciferase was determined by standard procedures . Standard luciferase assays in Figure S2 were done with GAL4 DNA binding domain proteins on a luciferase reporter driven by a SV40 minimal promoter as described in [31] . The acidic transactivation domain ( AD42 ) was derived from the mammalian two-hybrid vector pBAD42 ( Clontech ) , VP16 was taken from a laboratory stock . Expression of all GAL4 constructs has been shown in [31] . Luciferase values were normalized to protein content of the cell lysates . A GFP reporter was used to control for transfection efficiencies . The normalized luciferase values were multiplied by the percentage of GFP-positive cells . MLL-ENL and HA-tagged versions of Dot1l or AF4 were coexpressed in 293T cells from pcDNA3-based plasmids . To detect interaction of MLL-ENL with endogenous proteins MLL-ENL was introduced alone into 293T . Analogous experiments were done for MLL-AF5 and MLL-AF4 . In addition , native SEM cells containing MLL-AF4 transcribed from a natural 4;11 translocation were employed . REH cells that do not carry any 11q23 translocation were used as control . Cell extracts were prepared in 20 mM HEPES ( pH 7 . 5 ) , 10 mM KCl , 300 mM NaCl , 0 . 5 mM EDTA , 0 . 1% TritonX-100 , and 10% glycerol supplemented with protease and phosphatase inhibitors . Anti-MLL IgM antibody ( HRX107 ) was coupled to cyanogen bromide-activated sepharose ( SIGMA ) , and this affinity resin was used to bring down MLL and associated proteins . Precipitates were analyzed by Western blotting for the presence of HA-tagged proteins and endogenous ENL and CDK9 . For controls , the lysates were also probed with an anti-MLL antibody of a different isotype ( MLL-N , mouse IgG; AbCam ) . ChIP was performed according to standard procedures . In short , 10 ml ( approximately 5×10e6 cells ) of logarithmically growing culture was cross-linked for 10 min with 1% formaldehyde at room temperature for 10 min . Cross-linking was stopped by addition of 0 . 25 M glycine for a further 5 min . Cells were lysed in 50 mM Tris/HCl ( pH 8 . 0 ) , 10 mM EDTA , and 1% SDS , and chromatin was prepared by sonication . Precipitations were done with 2 µg of antibody and 30 µl of protein A/G agarose slurry ( Santa Cruz Biotechnology ) for 4 h . Precipitates were washed twice with buffer A ( 20 mM HEPES [pH 7 . 5] , 10 mM KCl , 0 . 5 mM EDTA , 0 . 1% TritonX-100 , and 10% glycerol ) followed by two washes each in buffer A supplemented with 500 mM LiCl and buffer A + 500 mM LiCl + 0 . 1% SDS . After the final wash in 10 mM Tris/HCl [pH 8 . 0] , 0 . 5 mM EDTA , the precipitates were eluted and cross-links were removed by incubation overnight in 50 mM Tris/HCl ( pH 8 . 0 ) , 10 mM EDTA , 1% SDS supplemented with 66 µg/ml RNAse A ( for input samples only ) and 0 . 5 µg of proteinase K . The treated supernatants were purified by a QIAquick Spin column ( Qiagen ) according to the instructions of the manufacturer . Precipitated DNA was quantified by qPCR with SYBR based premixes from Stratagene and compared to nonenriched DNA from input samples . Primer sequences used to amplify precipitated material are available on request . For ChIP-chip experiments displayed in Figure 5 , two cell lines were derived from primary mouse hematopoietic cells by transduction with flag-tagged MLL-ENL as described [57] . Chromatin IP was done as above with anti-flag agarose ( M2 flag agarose , Sigma ) and as a control with flag-agarose preblocked with flag peptide ( Sigma ) . Precipitates were amplified by ligation-mediated PCR exactly as described by the manufacturer ( NimbleGen-Roche ) and hybridized to NimbleGen promoter MM8 RefSeqArrays . This array design tiles genomic DNA 500 bp downstream and 2 kb upstream of all known RefSeq transcripts . Data analysis was done by NimbleGen , and results were visualized by SignalMap software . The software detects potential fMLL-ENL binding sites by searching in a 500-bp sliding window for four or more peaks with a log2 signal-to-noise ratio above 25% of a calculated maximum that equals the average of all peaks plus six standard deviations . A false-positive discovery rate ( FDR ) is calculated by a 20-fold randomization of the ratio data , and a probability of “randomness” is assigned to each peak . Peaks with a probability of <0 . 2 are indicative for binding . In total , two independent experiments were performed . Six- to 8-wk-old Balb/C mice were treated by intraperitoneal administration of 150 mg/kg 5-fluorouracil . Five days after injection , bone marrow enriched in precursor cells was harvested . Half of the cells were cultivated for 3 to 4 d in RPMI medium supplemented with 10 ng/ml IL-3 , IL-6 , and GM-CSF , as well as with 100 ng/ml SCF ( all recombinant cytokines were obtained from PeproTech ) and with increasing amounts of alsterpaullone or flavopiridol ( Sigma ) . Proliferation was assessed by a standard MTT assay . The second batch of cells was retrovirally infected with MLL-ENL as described in [57] , and after selection for transduced cells , the resulting MLL-ENL-transformed lines were checked for sensitivity towards alsterpaullone and flavopiridol under identical cytokine conditions as used for nontransformed cells .
|
The expression level of a gene needs to be precisely adjusted to ensure proper function . Adjustments can be imposed at different stages during the overall process of gene expression , including transcription initiation , transcript elongation , and transcript processing . If control of one of these mechanisms fails , aberrant gene expression can occur , which may have severe consequences such as cellular transformation and the development of cancer . Here , we show that a class of aberrant fusion proteins that are causal in mixed-lineage leukemia ( MLL ) hijacks a transcriptional elongation complex . We analyze the architecture of this transcriptional elongation complex and demonstrate that the complex is targeted by MLL fusion proteins to genes that should normally be silenced to allow maturation of hematopoietic cells . We show that this mistargeting causes constitutive expression of the respective genes , which likely leads to inhibition of blood cell differentiation at a precursor cell stage in which the cells are highly proliferative . Such abnormal precursor cells have been shown previously to be resistant to normal differentiation signals and to form the leukemia-initiating population . We further show here that cells carrying MLL fusion proteins are more sensitive to chemical inhibition of transcriptional elongation than leukemic cells of different etiology . Our results propose transcriptional elongation as a new oncogenic mechanism and point to a potential specific therapy for this hard-to-cure leukemia .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"oncology",
"hematology/acute",
"lymphoblastic",
"leukemia",
"hematology/pediatric",
"hematology",
"molecular",
"biology/transcription",
"elongation",
"oncology/hematological",
"malignancies",
"hematology/acute",
"myeloid",
"leukemia"
] |
2009
|
Misguided Transcriptional Elongation Causes Mixed Lineage Leukemia
|
The pancreatic islets of Langerhans are multicellular micro-organs integral to maintaining glucose homeostasis through secretion of the hormone insulin . β-cells within the islet exist as a highly coupled electrical network which coordinates electrical activity and insulin release at high glucose , but leads to global suppression at basal glucose . Despite its importance , how network dynamics generate this emergent binary on/off behavior remains to be elucidated . Previous work has suggested that a small threshold of quiescent cells is able to suppress the entire network . By modeling the islet as a Boolean network , we predicted a phase-transition between globally active and inactive states would emerge near this threshold number of cells , indicative of critical behavior . This was tested using islets with an inducible-expression mutation which renders defined numbers of cells electrically inactive , together with pharmacological modulation of electrical activity . This was combined with real-time imaging of intracellular free-calcium activity [Ca2+]i and measurement of physiological parameters in mice . As the number of inexcitable cells was increased beyond ∼15% , a phase-transition in islet activity occurred , switching from globally active wild-type behavior to global quiescence . This phase-transition was also seen in insulin secretion and blood glucose , indicating physiological impact . This behavior was reproduced in a multicellular dynamical model suggesting critical behavior in the islet may obey general properties of coupled heterogeneous networks . This study represents the first detailed explanation for how the islet facilitates inhibitory activity in spite of a heterogeneous cell population , as well as the role this plays in diabetes and its reversal . We further explain how islets utilize this critical behavior to leverage cellular heterogeneity and coordinate a robust insulin response with high dynamic range . These findings also give new insight into emergent multicellular dynamics in general which are applicable to many coupled physiological systems , specifically where inhibitory dynamics result from coupled networks .
Most biological systems exist as dynamic multicellular structures where distinct functionalities are generated through cellular interactions . While important for proper function , the complexity in network architecture , cellular dynamics , as well as the presence of heterogeneity , noise and biological variability make the overall function of multicellular structures difficult to understand . Approaches to understanding coupled dynamical systems have handled this complexity by explaining system structure and function individually [1] , [2] . These two aspects are both of central importance when it comes to understanding the way living systems are organized and how their anatomy supports their function . Therefore , by employing network theory to inform or predict the architectural aspects of dynamical system models , we can better understand how structural properties can impact functional behaviors . One living system exhibiting complex multicellular dynamics , yet with a scale tractable for study with these approaches , is the islet of Langerhans where dysfunction generally leads to diabetes . As such the islet provides a physiologically relevant system in which we can examine properties of multicellular dynamical systems and discover behavior that is broadly applicable . The islets of Langerhans are multicellular micro-organs located in the pancreas which maintain glucose homeostasis through the secretion of hormones such as insulin . Glucose-stimulated insulin secretion ( GSIS ) from β-cells within the islet is driven by glucose-dependent electrical activity . The metabolism of glucose and increased ATP/ADP ratio inhibits ATP-sensitive K+ ( KATP ) channels , causing membrane depolarization . Activation of voltage-dependent Ca2+ channels elevates intracellular free-calcium activity ( [Ca2+]i ) to trigger insulin granule exocytosis [3] , [4] . Defects at several points in this signaling pathway , including the KATP channel , can cause or enhance the risk of developing diabetes [5]–[8] . Despite the importance of this pathway , it is important to recognize β-cells do not act autonomously . Rather , like many tissues , there are extensive cell-cell interactions within the islet that govern overall function . For example , isolated β-cells exhibit heterogeneous sensitivities to glucose with a low overall dynamic range of GSIS [9]–[11] , yet β-cells within the islet robustly release insulin . Connexin36 ( Cx36 ) gap junctions mediate the electrical coupling between β-cells [12]–[14] which coordinates oscillations in electrical activity and insulin release across the islet , enhancing the pulsatile release of insulin and glucose homeostasis [13]–[15] . In the absence of coupling many cells in the islet also show spontaneous elevations in [Ca2+]i; likely as a result of heterogeneities in glucose sensitivity [10] , [16] . Therefore , another equally important role gap junctions play is to coordinate a suppression of spontaneous electrical activity at lower glucose levels [17] . Given that basal regulation is integral to glucose homeostasis , electrical coupling and the coordinated electrical dynamics are a critical factor in the regulation of islet function and in diabetes . Multicellular electrical dynamics in the islet have been described as functional networks where synchronized changes in [Ca2+]i indicate functional connectivity between cells [14] , [18] , [19] . Such network analysis has been applied to examine the dependence of [Ca2+]i dynamics on the level of coupling and its regulation , and has indicated that β-cell connectivity is non-homogeneous with a small subset of connections dominating synchronized behavior . As part of this analysis , the network of functional connectivity can be approximated by a Boolean network which quantitatively describes changes in multicellular behavior , including changes in coupling strength , network size or network shape [14] , [20]–[22] . These studies have generally focused on the synchronization of [Ca2+]i oscillations , and such synchronized oscillatory/pulsatile behavior has been similarly examined in other physiological multicellular systems [23]–[25] . However , few studies have theoretically examined the suppressive effect of electrical coupling in the islet and its ability to shape the glucose-regulation of electrical activity . This is particularly warranted given a recent study that showed how severe diabetes caused by expression of mutant KATP channels could be prevented through a modulation in gap junction coupling [26] . Therefore , details for how the network structure and composition facilitate a highly sensitive and robust response from a heterogeneous cell population remain to be determined . In this study we examine how electrical coupling within β-cell networks in the islet provide resilience against heterogeneous cell populations to generate robust network responses . We first develop quantitative predictions derived from a Boolean approximation of the β-cell network , where the dependence of [Ca2+]i on variations in the constituent cellular excitability and coupling is described . We then test these predictions using two experimental systems involving transgenic mice that express mutant KATP channels with increased or decreased ATP-sensitivity [27] , [28] . This creates defined populations of cells within the islet which are ‘excitable’ or ‘inexcitable’ , and can be further used to examine how our theoretical predictions and experimental data extend to physiological regulation of glucose homeostasis . We next link the static Boolean network model predictions and experimental findings with a dynamic multicellular model of the islet which incorporates recent understanding of β-cell electrophysiology [29] , [30] . We finally extend these experimental and theoretical measurements to a general case with a continuum of heterogeneous cellular behavior . A consistent feature in this study is the emergence of critical behavior as a result of β-cell electrical coupling , where the islet exhibits a phase transition between globally active and inactive states as cellular excitability approaches a critical threshold value . We discuss how the robust functionality that emerges at the multicellular level is not only relevant to the islet of Langerhans and its dysfunction in diabetes , but also to the function of other multicellular biological systems .
Based on prior approximations of heterogeneity in cellular excitability and coupling , Boolean networks of connectivity were simulated to predict how general multicellular electrical activity depends on the relative excitability of the constituent cell population and the coupling between individual cells [14] , [20] . Nodes within a cubic lattice had a probability Pexc of being active , and adjacent nodes were functionally coupled with a ‘coupling probability’ p ( figure 1A ) , and resultant clusters of coupled nodes were identified . ‘Inexcitable’ β-cells can suppress activity in excitable β-cells via electrical coupling [17] , with <30% inexcitable cells necessary for this suppression . To simulate this , a logic rule was used for each cluster of coupled nodes within a given lattice , where greater than a threshold percentage of inexcitable cells ( Sp ) can suppress activity in all other cells in its coupled network . Simulations of the resultant average network activity were run with varying values of Pexc , Sp , and p to represent differing cellular excitabilities and electrical coupling ( figure 1B , SI figure S1 ) . An increase in electrical activity is predicted as Pexc is increased; however the functional form is highly dependent on p ( figure 1C ) . In the absence of coupling ( p = 0 ) , a trivial linear response is obtained where Pexc represents the level of electrical activity . With increasing p , the activity becomes increasingly non-linear as a function of Pexc . For higher values of p ( 0 . 3 to 1 ) a sharp transition between active and quiescent behavior is observed , representing a phase transition with emerging critical behavior . These higher values of p lead to network-spanning coupling ( figure S1 , S2 ) , and as such the ‘rule’ governing suppression acts over the whole network . For low values of p ( 0 to 0 . 2 ) , the network is composed of coupled ‘clusters’ ( figure S2 ) , and the simulation is close to linear without a strong transition . This level of p corresponds to insufficient coupling to span the network , which is similar to the critical coupling probability ( ∼0 . 25 ) in percolation theory [31] . As such for p>0 . 25 there are 3 specific regions of emergent network behavior: a small ( ∼10% ) decrease for Pexc>0 . 85 ( ‘pre-critical state’ ) ; a rapid ∼75% drop at Pexc = 0 . 85 ( ‘critical state’ ) , then a small linear decrease for Pexc<0 . 85 ( ‘post-critical state’ ) . The critical state emerges when Pexc approaches 1-Sp; with the sharpness of the transition as well as behavior in the pre- and post-critical states being dependent on p . Overall , this transition can be understood by considering the well-defined threshold for activity ( Sp ) and the network-spanning connectivity that occurs above the critical coupling probability ( p>0 . 25 ) . For values of Pexc< ( 1-Sp ) there is a gradual decrease in network activity with increasing p ( figure 1D ) , representing the suppressive effect of coupling . For Pexc> ( 1-Sp ) , the network activity remains high , although a slight drop occurs for low levels of coupling . Therefore in a general Boolean network , electrical coupling is predicted to lead to critical behavior , where a phase transition in the network activity occurs as a function of constituent cellular activity . To test the Boolean network model predictions , we measured intracellular free-calcium activity ( [Ca2+]i ) in islets which had defined levels of cellular excitability . Islets were isolated from mice with inducible , β-cell specific expression of a mutant ATP-insensitive KATP channel subunit ( Kir6 . 2[ΔN30 , K185Q] ) under CreER-recombinase control [27] . Expression of these over-active KATP channels render β-cells functionally inexcitable , causing an absence of insulin release , marked hyperglycemia and diabetes [27] . Tamoxifen induction of CreER controls Kir6 . 2[ΔN30 , K185Q] expression levels which can be monitored via GFP co-expression , leading to both controllable and quantifiable cellular excitabilities ( SI figure S3 ) . At 20 mM glucose islets show [Ca2+]i which decreased with increasing expression of GFP and therefore Kir6 . 2[ΔN30 , K185Q] , similar to model predictions . This showed critical behavior with 3 specific regions ( figure 2A ) : For low GFP expression <15% ( few Kir6 . 2[ΔN30 , K185Q] expressing cells ) , [Ca2+]i was active over the entire islet with similar behavior to wild-type islets lacking GFP ( GFP = 0 ) and Kir6 . 2[ΔN30 , K185Q] ( figure 2A , BI–II , ‘pre-critical’ behavior ) . Oscillations were almost fully synchronous in each case ( not shown ) . For GFP expression at 10–20% there was a sharp drop-off in islet [Ca2+]i , where small changes in GFP resulted in highly disproportionate changes in [Ca2+]i ( figure 2A , BIII , ‘critical’ behavior ) . Activity was focused to clustered areas of synchronization ( not shown ) . For high GFP expression >25% ( high number of Kir6 . 2[ΔN30 , K185Q] expressing cells ) , islets showed sporadic [Ca2+]i restricted to increasingly smaller clusters ( Figure 2DIV , ‘post-critical behavior’ ) . Although islets with <20% GFP has similar overall activity compared to wild-type islets ( 0% GFP ) , there was a marked reduction in the plateau fraction in the <20% GFP group ( 15–40% ) compared to the wild-type group ( 50–70% ) ; indicating that even small numbers of inactive cells impacts global behavior . In comparison , islets with high levels of GFP ( >25% ) had a low plateau fraction ( 15–20% ) in those cells that were active . GFP+ cells showed similar activity to GFP− cells albeit with a small increase in activity , likely due to a few inactive non-β-cells included in the GFP− analysis ( SI figure S4 ) . Comparison of experimental data to the Boolean model can be seen for a number of values of p ( figure 2C ) and Sp ( figure 2D ) . Varying p ( gap junction coupling ) matches the sharpness of the transition , whereas varying Sp ( number of inexcitable cells required to suppress activity ) matches the position of the transition . A p = 0 . 3 ( 95% CI: 0 . 280–0 . 311 ) and Sp = 0 . 135–0 . 15 best fits the experimental [Ca2+]i data ( figure 2A , E ) . The distribution for fitted p was relatively broad but for Sp was well defined ( figure 2E ) . These values of p are similar to those found in prior studies examining the synchronization of [Ca2+]i oscillations , which indicated a limited level of functional coupling in the islet ( p = 0 . 31–0 . 36 ) [20] . These values of Sp are also consistent with experimental studies that suggest between 1 and 30% of inactive cells can suppress activity in other cells [17] , [20] . Therefore , introducing inexcitable cells into the islet experimentally generates critical behavior which quantitatively agrees with a Boolean network model and predicts the importance of electrical coupling in regulating multicellular excitability . β-cell [Ca2+]i drives insulin release to regulate glucose homeostasis . Given that the behavior in [Ca2+]i following varied expression of over-active KATP channels , we next tested whether this also occurred in downstream physiological parameters . Averaged over each mouse , similar [Ca2+]i was observed in wild-type islets lacking GFP and islets with low-level GFP ( <20% , ‘pre-critical’ ) , while both were significantly greater than [Ca2+]i in islets with high GFP expression ( >20% , ‘post-critical’ ) ( figure 3A ) . Plasma insulin also showed a similar transition , with pre-critical ( GFP<20% ) plasma insulin being significantly greater than post-critical ( GFP>20% ) plasma insulin ( figure 3B ) . However mice lacking GFP did show significantly greater insulin than mice with low-level GFP , correlating with the reduced plateau fraction observed . Insulin reduces glucose levels , and as expected pre-critical mice ( GFP<20% ) had normal glucose levels ( figure 3C ) , while post-critical mice ( GFP>20% ) demonstrated elevated glucose levels . Glucose-stimulated insulin secretion from isolated islets showed similar behavior to that of plasma insulin ( figures 3D–F ) , where again islets lacking GFP showed significantly greater GSIS than islets with low level GFP . Therefore insulin dynamics and blood glucose levels follow similar behavior as the driving [Ca2+]i following varied expression of Kir6 . 2[ΔN30 , K185Q] , demonstrating a physiological link in the critical behavior in [Ca2+]i activity as a function of Pexc , . The Boolean model accurately predicts the impact of variable cellular excitabilities ( Pexc ) on [Ca2+]i suppression at elevated glucose through expression of over-active KATP channels ( Kir6 . 2[ΔN30 , K185Q] ) . However , the Boolean model also predicts how [Ca2+]i suppression varies as a function of gap junction coupling ( p ) ( Figure 1D ) . To test this , we measured [Ca2+]i in islets from mice with β-cell specific mosaic expression of an inactive KATP channel subunit ( Kir6 . 2[AAA] ) . This was combined with a knockout of Cx36 , yielding 100% ( Cx36+/+ ) , 50% ( Cx36+/− ) or 0% ( Cx36−/− ) gap junction coupling , as well application of the gap junction inhibitor 18-α-glycyrrhetinic acid [10] . Expression of inactive KATP channels render β-cells constitutively ( glucose-independent ) active , yet islets which have a majority ( but not all ) of their cells expressing inactive KATP channels show glucose-dependent electrical activity similar to wild-type islets [17] . GFP co-expression indicates ∼70% of β-cells express inactive KATP channels ( SI figure S3 ) such that Pexc = 0 . 7 . With increasing gap junction coupling [Ca2+]i progressively decreased until residual activity was observed at full coupling , similar to that in the post-critical state upon Kir6 . 2[ΔN30 , K185Q] expression . There was strong agreement between experimental measurements and the Boolean Network model , with a p at normal ( Cx36+/+ , 100% ) gap junction coupling of 0 . 38 ( 95% CI: 0 . 372–0 . 394 ) and a suppression threshold Sp = 0 . 15 giving the best fit ( figure 4A ) . This is similar to p , Sp derived in the first experimental system ( figure 2 ) . Varying Sp affects the gap junction dependence in [Ca2+]i , with little effect between 0 . 05–0 . 2 , but strong divergence above 0 . 2 ( figure 4B ) . Therefore the Boolean network model can accurately predict behavior in a different experimental model with defined levels of cellular excitability ( Pexc ) and gap junction coupling ( p ) . The Boolean network model accurately describes how [Ca2+]i critically depends on cellular excitability and coupling . Nevertheless it is a static framework of a dynamical system and does not take into account limit-cycle behavior . To investigate whether similar behavior exists in a coupled dynamical oscillator model of the islet , we generated a multi-cellular version of a recent β-cell model which includes a comprehensive description of β-cell electrophysiology [30] . Our model also included a more realistic quasi-spherical architecture , heterogeneity in gap junction coupling [14] , [32] , and heterogeneity in endogenous cellular activity [11] , [14] . To model KATP-overactivity resulting from Kir6 . 2[ΔN30 , K185Q] expression , a defined fraction of cells with reduced ATP-inhibition of KATP activity was introduced to render them inexcitable . As with the Boolean network model and experimentally measured [Ca2+]i , a clear phase transition was observed at 20 mM glucose in the coupled oscillator model with ∼15% KATP-overactivity ( figure 5A ) . Again critical behavior manifested in three regimes . Simulated islets without KATP over-activity showed [Ca2+]i dynamics closely matching previously published models ( figure 5BI ) [30] . Simulated islets with low KATP-overactivity ( <15% ) showed a linear decrease in activity with a reduced plateau fraction as experimentally observed ( figure 5A , BII , ‘pre-critical’ behavior ) , while maintaining near-full synchronization . Simulated islets with KATP-overactivity at 10–30% again showed a sharp drop-off in [Ca2+]i , with small changes in KATP-overactivity leading to highly disproportionate changes in [Ca2+]i ( figure 5A , BIII ) . Simulated islets with high KATP-overactivity ( >30% ) showed only sporadic low level [Ca2+]i ( Figure 5A , BIV , ‘post-critical’ behavior ) . A physiological mean gap junction conductance of 120 pS [14] , [32] was found to best describe experimental data ( figure 5A , C ) . The sharpness and position of the phase transition was highly dependent on the mean coupling conductance , with increasing conductance leading to a sharper transition occurring at lower KATP-overactivity ( figure 5C ) . The islet is commonly modeled as a cubic lattice or other regular geometry [14] , [21] , [33] . A spherical islet-like structure which has a heterogeneous number of cell-cell connections ( mean , SD = 5 . 3 , 1 . 7 ) generated a less-sharp transition compared to a regular cubic geometry , and this better matched experimental data ( SI figure S5 ) . Similarly , a heterogeneous level of coupling conductance generated a less-sharp transition ( SI figure S6A ) . This indicates the importance of coupling heterogeneity , in terms of connection geometry , connection number and connection strength . The endogenous heterogeneity in cellular activity did not significantly impact the phase-transition indicating the dominating effect of Kir6 . 2[ΔN30 , K185Q] expression ( SI figure S6B ) . A similar phase-transition was also observed for simulations run at 11 mM glucose ( not shown ) . Therefore critical behavior also emerges in a dynamic coupled β-cell oscillator model with quantitative agreement with experimental measurements and a static Boolean network model . We have examined how the coupling between heterogeneous cells leads to critical behavior by introducing defined mutant populations of inexcitable cells ( Kir6 . 2[ΔN30 , K185Q] ) or excitable cells ( Kir6 . 2[AAA] ) . However , endogenous β-cells are themselves highly heterogeneous under physiological ranges of glucose , showing a continuum of excitabilities rather than being constitutively excitable/inexcitable [9]–[11] . To examine how gap junction coupling leads to critical behavior in the presence of endogenous heterogeneity , we applied a ‘ramp’ of increasing diazoxide concentrations to uniformly promote KATP channel opening . At 11 mM glucose , [Ca2+]i in wild-type islets at 0 µM and 50 µM diazoxide was similar , but at 100 µM there was a rapid ∼60% drop ( figure 6A ) , where only a few remaining cells were active ( figure 6B ) . Similar low-level [Ca2+]i was observed at 250 µM . In islets from mice lacking Cx36 gap junction coupling , similar [Ca2+]i was observed to wild-type islets at 0 µM diazoxide , albeit with no synchronization . Upon increasing diazoxide , a more gradual decrease in [Ca2+]i was observed , with less [Ca2+]i observed at 50 µM diaozixde but more [Ca2+]i remained at 100 µM diazoxide ( figure 6A , B ) . These data were also well described using the coupled dynamic oscillator model . In the presence of endogenous heterogeneity at 11 mM glucose , a uniform reduction in ATP-sensitive KATP inhibition led to a clear phase transition in islet [Ca2+]i in the presence of normal coupling ( 120 pS ) ( figure 6C , D ) . However , in the absence of coupling a more gradual change occurred in good agreement with experimental measurements ( figure 6C , D ) ; where [Ca2+]i was elevated in the absence of coupling over a certain range of uniform KATP inhibition . As such , <50 µM diazoxide lies in the ‘pre-critical’ regime , >100 µM diazoxide lies in the ‘post-critical’ regime , and the transition lies at 50–100 µM . In experiments with mutant KATP subunit expression , cells were considered ‘inexcitable’ if they showed GFP and Kir6 . 2[ΔN30 , K185Q] expression . In this case of endogenous heterogeneity , for a given concentration of diazoxide , we can consider a cell is ‘inexcitable’ if it is quiescent in the absence of electrical coupling . By plotting activity in the presence of coupling ( representing the resultant activity ) against activity in the absence of coupling ( representing intrinsic cellular excitability ) similar phase-transitions are apparent; with quantitative agreement between experimental data , dynamic coupled oscillator model and static network model ( figure 7A–C ) . The phase transition in the dynamic coupled oscillator model was dependent on how heterogeneity was generated , where heterogeneity in multiple factors rather than any one factor was required for agreement with experimental data ( SI figure S7 ) . Therefore critical behavior can occur more generally from the coupling between heterogeneous cellular populations within the islet , as exemplified here experimentally and theoretically .
In line with previous work describing coupled electrical dynamics , we showed that the structure and function of the islet cellular network can be described through principles of network theory [20] , [21] . Both the Boolean network and dynamic oscillator models predict the emergent behavior upon coupling between a heterogeneous cell population . The islet rapidly transitions between globally coordinated active and inactive states upon disproportionally small changes in the excitability of the constituent cells as they approach a critical ‘threshold’ excitability . This occurs under both conditions of β-cell heterogeneity we examined: the imposed bimodal β-cell populations achieved through expression of Kir6 . 2[ΔN30 , K185Q] or Kir6 . 2[AAA] mutations; and endogenous β-cell heterogeneity with diazoxide activation of KATP . The Boolean model reveals that there is an imbalance in the ability of excitable and inexcitable cells to respectively propagate stimulation or suppression . A low Sp in the model indicates a preference for excitable cells to be suppressed by inexcitable cells . This describes how gain-of-function Kir6 . 2[ΔN30 , K185Q] expressing cells ( which are glucose-unresponsive ) suppress activity in coupled normal cells at high glucose , and how loss-of-function Kir6 . 2[AAA] expressing cells are suppressed by normal cells at low glucose ( figure 2 , 4 ) . The role of p ( gap junction coupling ) determines the spatial extent over which suppression occurs . As shown in figure S1 and S2 , a low p results in coupled behavior restricted to a few cells and therefore inactive cells are unlikely to couple to many active cells and mediate suppression . When p exceeds the critical coupling probability ( ∼0 . 25 ) then coupling spans the whole network and inactive cells can couple to and suppress most active cells in the network . The sharp transition that emerges upon p>0 . 25 can be understood by considering that the threshold for activity ( Sp ) is well defined with a sharp cutoff for the Pexc which determines whether the cluster is active or inactive . The agreement with experimental data indicates that there is little variability between cells in this threshold for suppression , as also supported by the distributions of fitted Sp ( figure 2E ) . While the coupled dynamic oscillator model also predicts and describes the phase transitions present , the Boolean model describes the essential features that govern multicellular regulation of islet excitability . Results suggest that the islet may fundamentally behave in a binary fashion in terms of gap junction coupling and KATP-regulated excitability . Given the proportion of cells that intrinsically ( i . e . in the absence of coupling ) show activity at a given glucose stimulation and the level of coupling , the overall response of the islet can be approximated through this reductionist model . Of course dynamical features are missing from the Boolean model which is only described by the coupled dynamic oscillator model: including the altered oscillatory characteristics in the pre-critical state . The low p ( 0 . 30–0 . 38 ) required for the Boolean network model to quantitatively describe experimental data points to incomplete coupling present; and this can explain the residual activity in the post-critical state ( figure 1C ) . Recent studies of coordinated [Ca2+]i oscillations and waves in the islet have indicated a ‘backbone’ of a few strong connections dominate coupled [Ca2+]i dynamics , which is equivalent to a similarly low p [19] , [20] . The ability of the coupled dynamic oscillator model to also describe the transition between globally active and inactive states , suggests that the dynamics of the islet may behave according to general principles of coupled dynamical systems . Further work is needed to examine this critical behavior in more detail , including power law scaling and its dependence on network parameters and cellular properties . The phase-transition behavior can also be explained through a mean-field theory analogy ( SI figure S8 ) . Cells expressing the mutant Kir6 . 2[ΔN30 , K185Q] are intrinsically inexcitable ( figure S8A ) . In the ‘pre-critical’ regime the number of these cells is below a critical threshold and insufficient to suppress glucose-stimulated activity via coupling; therefore all cells are recruited to elevate [Ca2+]i . When the number of these inactive cells approaches the critical threshold ( Sp = 0 . 15 for Kir6 . 2[ΔN30 , K185Q] expression ) critical behavior emerges and coupling mediates suppression of other active cells . In normal islets treated with diazoxide , endogenous β-cell heterogeneity leads to variable intrinsic excitabilities and we expect diazoxide renders cells less glucose sensitive to be inexcitable ( figure S8B ) . In the absence of coupling these are observed to be inactive ( figure 6 ) . Low concentrations of diazoxide ( <50 µM ) render only a few cells inexcitable , which is below the critical threshold ( Sp∼0 . 5 ) and insufficient to suppress [Ca2+]i . At higher concentrations of diazoxide ( >100 µM ) more cells are rendered inexcitable , and when this number exceeds the critical threshold , coupling mediates suppression of other normally excitable cells . We predict that observed glucose-dependent activity and the coupling dependence can also be explained in this way ( see below ) . The Sp for endogenous heterogeneity is higher than that for an imposed biomodal distributions ( i . e . diazoxide treatment versus Kir6 . 2[ΔN30 , K185Q] or Kir6 . 2[AAA] expression ) suggesting a more even balance between the ability of excitable and inexcitable cells to respectively propagate stimulation or suppression in wild-type islets ( figure 6 , 7 ) . This balance may arise from the different distribution of heterogeneity present , but a phase-transition still emerges in the presence of coupling indicating a more general regulation of multicellular excitability . Therefore through limited coupling of heterogeneous populations of cells , critical behavior emerges in the islet dynamical system where large changes in activity result from small changes in the constituent cellular excitabilities . Gap junctions impact islet behavior in two main ways . At high glucose ( KATP channel-closure ) , gap junctions coordinate oscillatory dynamics of membrane depolarization and [Ca2+]i to generate a robust pulsatile insulin secretion [13] , [14] , [32] . A number of recent studies have examined this aspect , including multicellular modelling and quantitative analyses [14] , [18] , [19] , [21] . Equally important however is that at lower glucose ( KATP channel-opening ) , gap junctions mediate a suppression of membrane depolarization , [Ca2+]i , and insulin secretion [10] , [13] , [17] . The mechanisms involved in mediating suppression are not well characterized , and several experimental perturbations have yielded unexpected results or have not been well described theoretically [17] , [34] , [35] . Here , we were able to quantitatively describe suppressive behavior resulting from coupling , which yields a more complete understanding for how the islet functions under conditions of KATP channel opening . At 6–7 mM glucose , the islet sharply transition between global quiescence and globally synchronized [Ca2+]i oscillations . In the absence of coupling , the progressive elevation in the number of cells showing [Ca2+]i elevations is gradual [10] . This follows the same behavior as variable Kir6 . 2[ΔN30 , K185Q] expression and diazoxide concentration ( figures 2 , 6 ) . At <6 mM glucose , global suppression is equivalent to >15% Kir6 . 2[ΔN30 , K185Q] expression or >100 µM diazoxide; whereas at >7 mM glucose global activity is equivalent to <15% Kir6 . 2[ΔN30 , K185Q] expression or <50 µM diazoxide . The 6–7 mM glucose transition is therefore equivalent to behavior at ∼15% Kir6 . 2[ΔN30 , K185Q] expression or 50–100 µM diazoxide . As such , we propose results from the Boolean model , as illustrated by the mean-field theory analogy , have greater implications by describing physiological glucose-dependent islet electrical activity ( figure S8C ) . Coupling heterogeneity and islet architecture lead to variability in the number and strength of connections , impacting the phase transition . These factors may therefore play a role in shaping the physiological regulation of glucose-stimulated [Ca2+]i and insulin secretion ( Figures S5 , S6 ) . At 11 mM glucose , heterogeneity leads to a small population of β-cells ( <10% ) remaining inactive in the absence of coupling [10] . In the presence of coupling there is global activity with a lower plateau fraction compared to higher glucose levels ( e . g . 20 mM ) [36] . This matches the behavior at 5–10% Kir6 . 2[ΔN30 , K185Q] expression or 50 µM diazoxide in the respective absence and presence of coupling . Therefore an alternative view for how oscillatory dynamics are shaped at an islet-wide level is that less-active cells within the β-cell network have a modulatory effect on overall oscillation waveform , rather than oscillations being shaped by purely intrinsic properties of the β-cells . Importantly , the reduced plateau fraction of [Ca2+]i bursts at 5–10% Kir6 . 2[ΔN30 , K185Q] expression correlates with a significant decrease in insulin secretion ( figure 3 ) . A decrease in burst duration has previously been suggested to reduce insulin release [37] , as supported by these results . Thus subtle alterations in the balance of constituent cell excitabilities have a strong physiological effect on islet function . Our results also have implications for neonatal diabetes mellitus ( NDM ) , where the majority of cases result from mutations to Kir6 . 2 or SUR1 KATP channel subunits [8] , [38] . Kir6 . 2[ΔN30 , K185Q] expression models this disease [27] . Our results show that NDM mutations gives rise to a disproportionate suppression in [Ca2+]i and insulin release , thereby causing diabetes due to the critical behavior that emerges from coupling and network dynamics . This also explains how an absence of coupling elevates [Ca2+]i and insulin release ( figures 1 , 6 ) to prevent the progression of diabetes . This was experimentally demonstrated in a recent study [26] , and the rescue of diabetes can only be understood mechanistically at the multicellular level . Clearly in human diabetes , mutations are not expressed mosaically . However , the diazoxide results which depend on a continuum of heterogeneity ( figure 6 ) demonstrate that critical behavior exacerbates NDM upon uniform KATP channels overactivity . Other monogenic diabetes causing mutations that affect β-cell excitability , such as Glucokinase [6] , may also have similar effects on islet excitability and lend themselves to analysis by the Boolean model and coupled oscillator model . Mutations causing NDM are functionally equivalent to >15% Kir6 . 2[ΔN30 , K185Q] or >100 µM diazoxide , effectively residing in a post-critical state suppressing global [Ca2+]i . There exists a spectrum of KATP channel mutations linked to diabetes , where weaker mutations to Kir6 . 2 and SUR1 elevate the risk of type2 diabetes [39]–[41] . These mutations likely have a more subtle effect on islet excitability and as a consequence we predict that islets residing in the pre-critical state ( <15% Kir6 . 2[ΔN30 , K185Q] or <50 µM diazoxide ) would still be susceptible to diabetes following metabolic stress . Further , while gap junction reduction recovers insulin release and glucose control in the post-critical regime ( i . e . NDM ) , we predict a gap junction increase would be beneficial in the pre-threshold regime ( i . e . type2 diabetes ) . Converse to this , results from Kir6 . 2[AAA] islets show how critical behavior provides the islet with a resilience to over-excitable β-cells . Given the ∼85% threshold of excitable cells required to elevate [Ca2+]i , with ∼70% over-excitable Kir6 . 2[AAA]-expressing cells , many of the ∼30% normal β-cells would also need to be active ( e . g . >50% at ∼5 . 5 mM glucose [10] ) . This explains only the minor shift in glucose-stimulated [Ca2+]i that occurs following KATP inactivity and highlights the role electrical coupling plays in protecting islets against hyper-excitability [17] . Therefore we describe the emergence of critical behavior linking multiple levels including molecular and cellular behavior , multicellular behavior , in-vivo physiology , disease and treatment . This study also has implications for general understanding of physiological systems composed of coupled dynamic units . Previous theoretical studies have shown how the introduction of non-oscillatory elements above a critical level in a generalized coupled oscillator system can lead to cessation of global oscillations with phase transitions [42] , [43] . Our study experimentally and theoretically demonstrates this in a disease relevant system . Further , prior studies theoretically demonstrated that the fraction of excitable elements ( i . e . Pexc ) and coupling strength ( i . e . p ) exist in a phase plane where increased coupling decreases the number of inactive elements required for suppression [43] , [44] . We demonstrated this experimentally and theoretically , with Kir6 . 2[ΔN30 , K185Q] and diazoxide-induced suppression . While strong coupling promotes robust synchronization , it will increase suppression from non-oscillatory inexcitable units . Given a small population of inactive β-cells exists as a result of cellular heterogeneity , these generalized theoretical studies imply inappropriate elevations in coupling would be deleterious , by reducing glucose-stimulated [Ca2+]i . As such the level of coupling is likely at an optimal level to balance global synchronization and suppression . The strong link between dynamical β-cell networks and generalized coupled oscillators implies similar behavior can be expected in other physiological systems . In the heart , electrical activity is initiated by pacemaker cells and propagates to excite contractile myocytes . In culture , non-excitable fibroblasts proportionally reduced cardiomyocyte wave propagation bursts frequency with Cx43 dependence [45] . No activity was reported for >30% fibroblast penetrance and modulation of action potential frequency occurred at <30% fibroblast penetrance ( implying Sp = 0 . 3 ) . While the mechanisms of coupling-dependent suppression are very different compared to our study , global responses are similar implying similar governing principles . Similarly , pacemaker cells exhibit dominance over myocytes at an optimal gap junction conductance [46]; where high coupling leads to arrhythmias and low coupling leads to poor synchronization [47] . Neurons also display intrinsic oscillatory behavior , and the effects of coupling and presence of inhibitory and excitatory neurons on synchronization and phase modulation is an active area of research [48]–[53] . Critical dynamics have been described theoretically to emerge from excitatory and inhibitory units in neuronal networks [54] , and a computational study which introduced ‘contrarian elements’ into neural coupled oscillator networks found that a similar threshold of 15% suppressed global dynamics [49] . We also anticipate that critical behavior resulting from coupling of heterogeneous units may be considered a general regulatory mechanism . Many systems respond to a stimulus by transitioning between inactive and active states ( e . g . contractile , hormone-secretory ) . Our study implies that constituent cellular units need not themselves have a uniform or robust response to generate a robust multicellular response . Rather , a robust response can emerge from coupling a heterogeneous collection of cells; where coupling and architecture need only have sufficient strength and connection number on average . This makes the overall system robust against noise and variability , and loosens the requirement for tight regulatory mechanisms within the constituent cells . Similarly , given a constant stimulus , a robust transition between globally active and inactive states could be achieved by remodeling connectivity with a small number of inhibitory units . For example , down-regulating connections from ∼20% to ∼5% inhibitory cells would transition the system from inactive to active requiring minimal system remodeling . This also suggests how inappropriate changes in coupling or constituent cells may lead to global non-responsiveness and disease . We speculate these principles may apply to other neuroendocrine cell systems such as GH-cells or the adrenal medulla , where functional remodeling elevates hormone secretion upon physiological stimuli [18] , [55] . Indeed many of these principles have been linked with GnRH neuron function during development [56] , [57] . Therefore we suggest that new and robust functionalities can be generated at the multicellular level from the coupling of non-robust constituent cell function , requiring minimal system resources compared to the requirements were cells to act autonomously . Any living system cannot avoid deterioration through mutation or other pathological insult . This study experimentally and theoretically demonstrates that if the fraction of inactive elements exceeds a coupling-dependent threshold , the global activity of the system can be abolished . In the case of the islet this explains how inactive cells can suppress the activity of other cells , thereby preventing the secretion of insulin . In the case of KATP mutations , this quantifies the threshold of inexcitable cells required for pathogenic symptoms and explains how coupling can eliminate the emergence of diabetes or exacerbate it . Overall , this gives a new understanding for how emergent properties of the islet as a β-cell network are generated; as well as for understanding islet dysfunction in diabetes and novel ways to overcome dysfunction . More broadly , this generates insight into emergent behavior of multicellular systems in general .
All experiments were performed in compliance with the relevant laws and institutional guidelines , and were approved by the University of Colorado Institutional Biosafety Committee ( IBC ) and Institutional Animal Care and Use Committee ( IACUC ) . The generation of Rosa26-Kir6 . 2[ΔN30 , K185Q] ( ‘gain-of-function’ KATP subunit with GFP co-expression ) , Pdx-CreER ( β-cell specific inducible Cre ) , Kir6 . 2[AAA] ( ‘loss-of-function’ KATP subunit with GFP tag ) , and Cx36−/− ( Connexin36 global knockout ) have been described previously [27] , [28] , [58] , [59] . Expression of variable Kir6 . 2[ΔN30 , K185Q] was achieved in β-cells by crossing Rosa26-Kir6 . 2[ΔN30 , K185Q] and Pdx-CreER mice , and inducing Kir6 . 2[ΔN30 , K185Q] expression in 8–16 week old mice by 1–5 daily doses of tamoxifen ( 50 mg/g body-weight ) . Littermates lacking Rosa26-Kir6 . 2[ΔN30 , K185Q] and/or Pdx-CreER were used as controls . Blood glucose was measured daily and averaged over day 27–29 post tamoxifen induction using a glucometer ( Ascensia Contour , Bayer ) . Plasma insulin was measured at day 29 from blood samples centrifuged for 15 minutes at 13 , 900RCF , and assayed using mouse ultrasensitive insulin ELISA ( Alpco ) . Islets were isolated by collagenase injection into the pancreas through the pancreatic duct; the pancreas was harvested and digested , and islets were handpicked [23] . Islets were maintained in RPMI medium ( Invitrogen ) supplemented with 10% FBS , 11 mM glucose , 100 U/ml penicillin , 100 µg/ml streptomycin , at 37°C under humidified 5% CO2 for 24–48 hours prior to study . For insulin secretion measurements , islets ( 5/column , duplicates ) were pre-incubated in Krebs-Ringer buffer ( 128 . 8 mM NaCl , 5 mM NaHCO3 , 5 . 8 mM KCl , 1 . 2 mM KH2PO4 , 2 . 5mM CaCl2 , 1 . 2 mM MgSO4 , 10 mM HEPES , 0 . 1% BSA , pH 7 . 4 ) plus 2 mM glucose; then incubated for 60 minutes in Krebs-Ringer buffer plus 20 mM glucose . After incubation , the medium was sampled and insulin concentration assayed using mouse ultrasensitive insulin ELISA . To estimate insulin content , islets were lysed in 1% TritonX-100 and frozen at −20°C overnight . To measure [Ca2+]i dynamics , isolated islets were loaded with 4 µM FuraRed-AM ( Invitrogen ) in imaging medium ( 125 mM NaCl , 5 . 7 mM KCl , 2 . 5 mM CaCl2 , 1 . 2 mM MgCl2 , 10 mM Hepes , 2 mM glucose , and 0 . 1% BSA , pH 7 . 4 ) at room temperature for 90–120 minutes and held in polymdimethylsiloxane ( PDMS ) microfluidic devices [17] . FuraRed fluorescence was imaged on a spinning-disk confocal microscope ( Marianas , 3I ) with a 40× 1 . 3NA Plan-NEOFluar oil-immersion objective ( Zeiss ) maintained at 37°C . Images were acquired at 1 frame/sec using a 488 nm diode laser for excitation and a 580–655 nm long-pass filter for emission . Time-courses were acquired 10 minutes after change in glucose concentration , diazoixide or 18-α-glycyrrhetinic acid application . Time-courses are displayed as normalized to the average fluorescence . All images were analyzed using custom MATLAB ( Mathworks ) routines or using Slidebook ( 3I ) . To calculate islet activity , images were smoothed using a 5×5 average filter . The variance of pixel time-courses was first calculated for a quiescent reference cell; manually selected from an area which displayed no fluctuations in intensity over time compared with image noise . A pixel was considered ‘active’ if its time-course showed a variance >2 standard deviations above the variance of the quiescent reference cell [10] , [20] . Photobleaching was accounted for through a linear fit , and time-courses were rejected if excessive motion artifacts occurred . The area of active cells in terms of pixels , was determined for each condition and expressed as a fraction of total islet pixel area as defined by mean FuraRed fluorescence . GFP+ regions were defined as having a mean pixel fluorescence intensity above that measured in GFP− wild-type cells . The area of active cells in GFP+ regions was expressed as a fraction of the total GFP+ area . Information describing activity is represented as a false-color HSV image where Hue is set to 1 ( red ) , Saturation is set to 1 for active cells and 0 ( no color ) for inactive cells , and Value ( intensity ) is set to the average FuraRed fluorescence . Data are presented as mean±SEM . For comparison of two means , Student's t-test was utilized . For comparison of multiple means , ANOVA was utilized along with Tukey's HSD test . Bond percolation is a sub-model of percolation theory [31] , [60] which can be used to simulate the islet [14] , [20] , [21] . For a lattice of nodes ( cells ) in a given geometry , adjacent nodes are connected with a ‘coupling probability’ p , or not connected with a probability ( 1-p ) . Connected nodes are considered ‘functionally coupled’ , where activity is synchronized at high glucose and suppression mediated at low glucose . We implemented simulations of bond percolation lattices ( figure S1 ) as previously described [20] . Briefly , cubic lattices with alternating node and bond sites ( length L = 11 ) were generated . Probabilities were assigned to each bond site with a uniform distribution ( 0 to 1 ) . Neighboring nodes were coupled if the bond site probability was less than or equal to the coupling probability p . Clusters of coupled nodes were identified and potential bond sites removed to establish a matrix of identified coupled nodes ( figure S1A ) . Coupling-mediated suppression is based on the principle that a threshold fraction of non-responsive ( ‘inexcitable’ ) cells can suppress all other cells to which they are coupled [17] , [20] . Within a cluster of coupled nodes , if the fraction of inexcitable cells is greater than a threshold fraction ( Sp ) , then all cells within the cluster are inactive . Experimental studies indicate this threshold is <30% [17] and has been modelled to be ∼15% for MIN6 aggregates [20] . The probability Pexc defines the fraction of cells within the islet that are intrinsically excitable; where in the absence of coupling they would be active . Within each cluster of coupled nodes a binomial distribution was used to estimate the probability of there being a threshold number of inactive cells to lead to suppression . Given a threshold number of inactive cells required for suppression ( k ) , a total number of cells in a cluster ( n ) , and the fraction of inactive cells ( q ) ; the probability that a coupled cluster is active ( Pr ) is: ( 1 ) where ( 2 ) The cumulative distribution for k to n , P ( X≤k ) , represents the probability of suppression ( sufficient inexcitable cells ) in a coupled cluster . To obtain the resultant % activity , P ( X≤k ) was averaged over all clusters within the islet , weighted by the number of cells n in each cluster . ‘k’ normalized to ‘n’ gives the fraction of inexcitable cells required for suppression ( Sp ) . ‘1-q’ gives the fraction of excitable cells ( Pexc ) . 500 simulations were run for each Pexc = [0 1] and p = [0 1] , at 0 . 01 increments for given values of Sp . The p and Sp parameters that generated the best fit for the simulation mean to experimental data were determined by a chi-squared minimization . To determine the probability distribution for p and Sp , 4000 simulations were run and each simulation was separately fitted for p and Sp by chi-squared minimization . The islet model was based on the Cha-Noma β-cell model [29] , [30] , itself based on the Fridlyand β-cell model [61] , [62] , and adapted to include cell-cell coupling . We also included further aspects of cell-cell coupling and altered KATP channel function . A list of parameters used in the model is included in SI Table S1 . The membrane potential ( Vi ) of each β-cell i is related to the total transmembrane current as: ( 3 ) Where the kinetics of each current is described in [29] , [30] . To simulate gap junction coupling and a multicellular islet , multiple ‘cells’ were simulated with a coupling current between each neighboring cell . The membrane potential for each cell i was modified to account for coupling to j neighboring cells: ( 4 ) Heterogeneity in coupling was introduced by randomly assigning the gap junction conductance between cells i and j , according to an experimentally measured distribution [unpublished data] , with SD/mean = 70% . To more accurately model β-cell coupling architecture , random cell lattices were created using a position- and availability-based sphere-packing algorithm ( mean , SD number of cell-cell connections = 5 . 3 , 1 . 7 ) [63] ( figure S3 ) . IK ( ATP ) was described in [29] , [30] as: ( 5 ) Where is the open channel conductance and represents the mean open probability which is given by: ( 6 ) Endogenous heterogeneity was modelled by randomizing all parameters indicated in Table S1 between cells about a mean value according to a Gaussian distribution with SD/mean = 10% . To generate heterogeneity in electrical responses equivalent to experimental measurements , the open channel conductance was randomized between cells about a mean value according to a Gaussian distribution with SD/mean = 25% . This heterogeneity achieves variability in activity that matches experimental measurements in islets lacking Cx36 [10] , [14] . To model Kir6 . 2[ΔN30 , K185Q] expression , the open probability was modified in a proportion ( Pexc ) of simulated cells: ( 7 ) where γ is a constant representing the fraction of ATP-insensitive current ( ) , and was set to 0 . 5 . To model diazoxide treatment , the fraction of ATP-insensitive current was increased in all cells uniformly according to: ( 8 ) such that α = 1 represents an untreated islet , and α = 0 . 5 is equivalent to 100% expression of Kir6 . 2[ΔN30 , K185Q] . All simulations were initially written and verified in MATLAB , then rewritten in C++ and simulated on the University of Colorado JANUS supercomputer . The model was solved using a constant time-step Euler integration scheme ( Boost C++ Libraries ) with 100 µs step-time and 100 ms sampling-time . Rendering simulations was performed with Mathematica 9 . 0 ( Wolfram Research ) .
|
As science has successfully broken down the elements of many biological systems , the network dynamics of large-scale cellular interactions has emerged as a new frontier . One way to understand how dynamical elements within large networks behave collectively is via mathematical modeling . Diabetes , which is of increasing international concern , is commonly caused by a deterioration of these complex dynamics in a highly coupled micro-organ called the islet of Langerhans . Therefore , if we are to understand diabetes and how to treat it , we must understand how coupling affects ensemble dynamics . While the role of network connectivity in islet excitation under stimulatory conditions has been well studied , how connectivity also suppresses activity under fasting conditions remains to be elucidated . Here we use two network models of islet connectivity to investigate this process . Using genetically altered islets and pharmacological treatments , we show how suppression of islet activity is solely dependent on a threshold number of inactive cells . We found that the islet exhibits critical behavior in the threshold region , rapidly transitioning from global activity to inactivity . We therefore propose how the islet and multicellular systems in general can generate a robust stimulated response from a heterogeneous cell population .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biotechnology",
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"membrane",
"potential",
"electrophysiology",
"endocrine",
"physiology",
"diabetes",
"mellitus",
"biological",
"systems",
"engineering",
"bioengineering",
"computer",
"and",
"information",
"sciences",
"biophysics",
"metabolic",
"disorders",
"physics",
"computer",
"modeling",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"insulin",
"secretion",
"biophysical",
"simulations"
] |
2014
|
Phase Transitions in the Multi-cellular Regulatory Behavior of Pancreatic Islet Excitability
|
Detailed modeling and simulation of biochemical systems is complicated by the problem of combinatorial complexity , an explosion in the number of species and reactions due to myriad protein-protein interactions and post-translational modifications . Rule-based modeling overcomes this problem by representing molecules as structured objects and encoding their interactions as pattern-based rules . This greatly simplifies the process of model specification , avoiding the tedious and error prone task of manually enumerating all species and reactions that can potentially exist in a system . From a simulation perspective , rule-based models can be expanded algorithmically into fully-enumerated reaction networks and simulated using a variety of network-based simulation methods , such as ordinary differential equations or Gillespie's algorithm , provided that the network is not exceedingly large . Alternatively , rule-based models can be simulated directly using particle-based kinetic Monte Carlo methods . This “network-free” approach produces exact stochastic trajectories with a computational cost that is independent of network size . However , memory and run time costs increase with the number of particles , limiting the size of system that can be feasibly simulated . Here , we present a hybrid particle/population simulation method that combines the best attributes of both the network-based and network-free approaches . The method takes as input a rule-based model and a user-specified subset of species to treat as population variables rather than as particles . The model is then transformed by a process of “partial network expansion” into a dynamically equivalent form that can be simulated using a population-adapted network-free simulator . The transformation method has been implemented within the open-source rule-based modeling platform BioNetGen , and resulting hybrid models can be simulated using the particle-based simulator NFsim . Performance tests show that significant memory savings can be achieved using the new approach and a monetary cost analysis provides a practical measure of its utility .
Cell signaling encompasses the collection of cellular processes that sample the extracellular environment , process and transmit that information to the interior of the cell , and regulate cellular responses . In a typical scenario , molecules outside of the cell bind to cognate receptors on the cell membrane , resulting in conformational changes or clustering of receptors . A complex series of protein binding and biochemical events then occurs , ultimately leading to the activation or deactivation of proteins that regulate gene expression or other cellular processes [1] . A typical signaling protein possesses multiple interaction sites with activities that can be modified by direct chemical modification or by the effects of modification or interaction at other sites . This complexity at the protein level leads to a combinatorial explosion in the number of possible species and reactions at the level of signaling networks [2] . Combinatorial complexity poses a major barrier to the development of detailed , mechanistic models of biochemical systems . Traditional modeling approaches that require manual enumeration of all potential species and reactions in a network are infeasible or impractical [2]–[4] . This has motivated the development of rule-based modeling languages , such as the BioNetGen language ( BNGL ) [5] , [6] , Kappa [7] , [8] , and others [9]–[12] , that provide a rich yet concise description of signaling proteins and their interactions [13] . The combinatorial explosion problem is avoided by representing interacting molecules as structured objects and using pattern-based rules to encode their interactions . In the graph-based formalisms of BNGL and Kappa , molecules are represented as graphs and biochemical interactions by graph-rewriting rules . Rules are local in the sense that only the properties of the reactants that are transformed , or are required for the transformation to take place , affect their ability to react . As such , each rule defines a class of reactions that share a common set of transformations ( e . g . , the formation of a bond between molecules ) and requirements for those transformations to take place ( e . g . , that one or more components have a particular covalent modification ) . The number of reactions encoded by a rule varies depending on the specifics of the model; a rule-based encoding is considered compact if it contains rules that encode large numbers of reactions . Overviews of rule-based modeling with BNGL can be found in Sec . S3 . 1 of Text S1 and Refs . [6] , [14] . A description of the graph-theoretic formalism underlying BNGL is provided in Sec . S4 . 1 of Text S1 , building on a previous graph-theoretical treatment [15] . An important characteristic of rule-based models is that they can encode both finite and infinite reaction networks . If the network is finite and “not too large” ( ≲10000 reactions [16] ) it can be generated from the rule-based model algorithmically by a process known as “network generation” [5] , [6] , [14] , [15] , [17] . Network generation begins by applying the rules of a rule-based model to a set of initial “seed” species , which define the initial state of the model system , to generate new species and reactions . The new species are then matched against the existing species to determine whether or not they are already present in the network [18] . Any species that are not already present are added to the network and an additional round of rule application is performed . This iterative process continues until an iteration is encountered in which no new species are generated . The resulting system of reactions can then be simulated using a variety of network-based deterministic and stochastic simulation methods . For example , network-based simulation methods currently implemented within BioNetGen include SUNDIALS CVODE [19] for ordinary differential equation ( ODE ) -based simulations , Gillespie's stochastic simulation algorithm ( SSA; direct method with dynamic propensity sorting ) [20] , [21] , and the accelerated-stochastic “partitioned-leaping algorithm” [22] . The rule-based methodology also provides a way to simulate models with prohibitively large or infinite numbers of species and reactions . This “network-free” approach involves representing molecular complexes as particles and applying rule transformations to those particles at runtime using a kinetic Monte Carlo update scheme [23] , [24] . At each simulation step , reactant patterns are matched to the molecular complexes within the system to calculate rule propensities . The rule to next fire is then selected probabilistically as in the SSA [20] and the particle ( s ) to participate in the transformation is ( are ) selected randomly from the set of matches . When the rule fires , transformations are applied to the reactant complexes to create the products . Since the reactants and products are determined at runtime there is no need to enumerate all species and reactions a priori as in network-based methods . This procedure is a particle-based variant of Gillespie's algorithm [23] , [24] and a generalization of the “n-fold way” of Bortz et al . [25] , which was originally developed to accelerate the simulation of Ising spin systems . An efficient , open-source implementation that is compatible with BNGL models is NFsim , the “network-free simulator” [16] . Other network-free simulation tools for rule-based models include RuleMonkey [26] , DYNSTOC [27] , SRsim [28] , and KaSim [24] . A recent paper [29] compares the rejection-based sampling technique [23] used in NFsim with the rejection-free approach employed in RuleMonkey . For models of multivalent ligand-receptor binding , rejection-based sampling was shown to be more efficient in the vicinity of the solution-gel phase boundary , while rejection-free sampling was more efficient for simulating the dynamics within the gel phase . Since only the current set of molecular complexes and the transformations that can be applied to them are tracked , network-free methods can efficiently simulate systems that are intractable to network-based methods [16] , [23] , [24] , [29] . However , the explicit representation of every molecule in the system is a major shortcoming of the approach . As such , network-free methods can require large amounts of computational memory for systems that contain large numbers of particles , a potential barrier to simulating systems such as the regulatory networks of a whole cell [30] , [31] . A typical eukaryotic cell , for example , contains on the order of protein-coding genes , mRNA molecules , and protein molecules [32] , [33] , along with much larger numbers of metabolites , lipids , and other small molecules . Simulating a cell at this level of detail using a network-free approach would be impractical . There is a need , therefore , for new approaches that can reduce the memory requirements of network-free simulation methods . A common measure of the computational cost of an algorithm is its computational complexity . In basic terms , computational complexity measures how the computational cost increases as an algorithm is applied to increasingly larger data sets [34] . For the simulation methods considered in this paper , two types of computational complexity are important: ( i ) space complexity , the number of memory units consumed during the execution of an algorithm; ( ii ) time complexity , the number of computational steps required to complete an algorithm . Network-based exact-stochastic simulation methods , like Gillespie's SSA [20] , [35] , [36] , treat species as lumped variables with a population counter . Therefore , their space complexity is constant in the number of particles in the system . However , representing the reaction network has a space complexity that is linear ( or worse if a reaction dependency graph is used [37] , [38] ) in the number of reactions . Network-based SSA methods are thus space efficient for systems with large numbers of particles , but less so for systems with large numbers of reactions . The time complexity of SSA methods is more difficult to quantify . It depends on model-specific factors such as the number of reactions in the network and the values of rate constants and species concentrations , as well as methodological factors such as how the next reaction to fire in the system is selected [20] , [21] , [37]–[41] and how reaction propensities are updated after each reaction firing [37] , [38] . However , for our purposes , what matters is that the time cost per event ( reaction firing ) for these methods is constant in the number of particles in the system and increases with the number of reactions in the network . Network-free methods , in contrast , represent each particle individually . Thus , their space complexity is linear in the number of particles . This is the primary shortcoming of these methods , as it limits the size of system that can be feasibly simulated . However , since reactions are not enumerated , their space complexity is linear in the number of rules , rather than the number of reactions . This is a key advantage for models where very large reaction networks are encoded by a small number of rules . Network-free methods also have an advantage over network-based methods in that their time complexity per event also scales with the number of rules , rather than the number of reactions . Since the number of rules in a rule-based model is typically far less than the number of reactions , this can be a substantial improvement . For example , NFsim has been demonstrated to significantly outperform network-based SSA methods for a family of receptor signaling models with large reaction networks [16] . We also note that for many models network-free methods have a time cost per event that is constant in the number of particles . However , for systems in which large aggregates form ( e . g . , models with polymerization dynamics [42] , [43] ) the cost can be significantly higher , scaling with the number of particles [16] , [24] . Nevertheless , network-free methods are still usually the best option in these cases because these types of models tend to encode very large reaction networks [16] . In Table 1 , we summarize the space and time complexities for different network-based SSA variants and for the network-free algorithm . Of most relevance to the current work are the entries that show: ( i ) the space complexity of network-based methods is constant in the number of particles and linear ( or worse ) in the reaction network size; ( ii ) the space complexity of network-free methods is linear in the number of particles and independent of the reaction network size , depending instead on the number of rules; ( iii ) the time complexity of network-based methods depends on the number of reactions in the network while for network-free methods it depends on the number of rules . Network-based methods are thus the best choice for systems with large numbers of particles and a small to moderate reaction network , and network-free methods are the best choice for systems with a large reaction network and small to moderate numbers of particles . However , neither method is optimal for systems that contain both a large number of particles and a large reaction network . The key idea pursued in this work is that memory consumption can be reduced in network-free simulators if simple species and small molecular complexes that exist in the system in large numbers are treated as population variables with counters rather than as particles . However , retaining the ability to address combinatorial complexity requires retaining the particle representation for species and complexes that are comprised of many molecules and/or have a large number of internal states . Here , we present an approach , termed the hybrid particle/population ( HPP ) simulation method , that accomplishes this . Given a user-defined set of species to treat as population variables , the HPP method partially expands the network around these population species and then simulates the partially-expanded model using a population-adapted particle-based method . By treating complex species as structured particles , HPP capitalizes on the reduced time complexity with respect to network size characteristic of the network-free approach . However , for the subset of species treated as population variables , we take advantage of the constant memory requirements of the network-based methodology . It is important to emphasize that in the HPP approach it is the system that is represented in a hybrid manner , as a collection of particles and population variables . The underlying simulator remains the same particle-based variant of Gillespie's algorithm that is used in existing network-free simulators [23] , [24] , but with small modifications to support population variables . This distinguishes HPP from other types of hybrid methods that combine different simulation methodologies , e . g . , ODE/SSA integrators [44]–[53] . While numerous rule-based modeling frameworks have been developed , little has been done with regard to hybrid particle/population simulation . Kappa [7] , [8] has the concept of “tokens , ” which are structureless population-type species . Modelers can write hybrid models in terms of both structured “agents” and structureless tokens and simulate them using KaSim 3 , the most recent version of the Kappa-compatible network-free simulator ( https://github . com/jkrivine/KaSim ) . However , there is no facility for transforming a model written exclusively in terms of agents into a hybrid form , as in our HPP method . Bittig et al . [10] have developed a spatial rule-based language called ML-Space that builds upon the multi-level language ML-Rules [9] . “Entities” that are assigned optional attributes such as shape , volume , and position in continuous space are automatically treated as particles diffusing via Brownian motion , while those without these attributes are treated as population variables reacting and diffusing within a discretized space ( subvolumes ) . For non-spatial models , the population-based network-free algorithm ( PNFA ) of Liu et al . [54] employs a similar philosophy: all multi-state ( structured ) species are automatically treated as particles , while single-state species are treated as population variables . Both ML-Space and PNFA lack a general representation of intermolecular bonding , which makes it difficult to account for combinatorial complexity associated with aggregation processes [2] , [29] . Falkenberg et al . [55] have proposed a hybrid deterministic/stochastic method that specifically addresses the problem of aggregation . Their approach first calculates occupancy probabilities as a function of time for all binding-site types by treating them as population variables and numerically integrating an associated set of deterministic ODEs describing the binding/unbinding kinetics . An ensemble of system states is then obtained by randomly distributing bonds , based on these probabilities , among a finite number of discrete molecules . The method assumes that inter- and intra-molecular bond formations occur with equal rates . Thus , although efficient for problems with high symmetry , its applicability to more general cases may be limited . Other approaches aimed at improving the efficiency of rule-based simulations include “on-the-fly” network generation [17] , [56] , [57] , where the reaction network is gradually built up by adding reactions only when new species appear in the system . The approach has only been developed within the context of discrete-stochastic simulation and has been shown to be significantly less efficient than network-free approaches when applied to combinatorially-complex models [23] , [58] . An alternative approach to reducing computational cost is exact model reduction ( EMR ) [59]–[64] . EMR aims to reduce the state space of a rule-based model while preserving the exact system dynamics with respect to observable quantities . These methods can achieve dramatic reductions in model complexity when applied within the context of ODEs , so long as the model does not contain significant cooperative or allosteric interactions [62] , [64] . EMR for stochastic simulations , however , has so far been less successful ( see http://infoscience . epfl . ch/record/142570/files/stochastic_fragments . pdf ) .
We have tested the performance of the HPP method by applying it to four example models , summarized in Table 2 and discussed in further detail below . All of the models are biologically relevant and are either taken directly from the literature or are based on models taken from the literature . Complete BNGL encodings , HPP configuration files ( containing actions for loading models , defining population maps , and executing simulations ) , and partially-expanded versions of all example models are provided as Texts S5 , S6 , S7 , S8 , S9 , S10 , S11 , S12 , S13 , S14 , S15 , S16 , S17 of the supporting information . HPP was evaluated for peak memory use , CPU run time , and accuracy as compared to particle-based NFsim simulations . For models where network generation is possible ( and EGFR ) , comparisons were also made to SSA simulations ( as implemented within BioNetGen [6] ) . All simulations were run on a Intel Xeon E5520 @ 2 . 27 GHz ( 8 cores , 16 threads , x86_64 instruction set ) with of RAM running the GNU/Linux operating system . To ensure that each process had access to of the compute cycles of a thread , no more than simulations were run simultaneously . All HPP and NFsim simulations reported in this work were run using NFsim version 1 . 11 , which is available for download at http://emonet . biology . yale . edu/nfsim . All simulations ( SSA included ) were invoked through BioNetGen version 2 . 2 . 4 , which implements the hybrid model generator and is distributed with NFsim 1 . 11 . Instructions for running simulations with BioNetGen ( ODE , SSA , and HPP ) can be found in Secs . S3 . 2 and S3 . 3 of Text S1 and Refs . [6] , [14] . NFsim and BioNetGen source code are available at http://code . google . com/p/nfsim and http://code . google . com/p/bionetgen , respectively . Additional documentation for BioNetGen can be found at http://bionetgen . org .
In this section , we first present an approach , termed “partial network expansion , ” for transforming a rule-based model into a dynamically-equivalent , partially-expanded form . We then describe a simple modification to the network-free simulation protocol that permits simulation of the transformed model as a collection of both particles and population variables . We refer to the combination of these methods as the hybrid particle/population ( HPP ) simulation method . The basic workflow is shown in Fig . 1 . The HPP approach is analogous to the coupled procedure of network generation and simulation described above , where a rule-based model is first transformed into a fully-expanded reaction network and then simulated as a collection of population variables ( i . e . , species ) using a network-based simulator . The obvious differences are that in HPP the network is only partially expanded and the system can only be simulated stochastically using a population-adapted network-free simulator . The partial network expansion algorithm has been implemented within the open-source rule-based modeling package BioNetGen [5] , [6] , [14] and resulting hybrid models can be simulated using version 1 . 11 ( or later ) of the network-free simulator NFsim [16] , which has been modified to handle population-type species . For convenience , we adhere in this paper to the BNGL syntax , which is summarized in Sec . S3 . 1 of Text S1 of the supporting material . However , the HPP method is generally applicable to any rule-based modeling language for which there exists a network-free simulator capable of handling a mixed particle/population system representation , e . g . , KaSim 3 . x for Kappa language models ( see https://github . com/jkrivine/KaSim ) .
We have presented a hybrid particle/population simulation approach for rule-based models of biological systems . The HPP approach is applied in two stages ( Fig . 1 ) : ( i ) transformation of a rule-based model into a dynamically-equivalent hybrid form by partially expanding the network around a selected set of population species; ( ii ) simulation of the transformed model using a population-adapted network-free simulator . The method is formally exact for an infinite population lumping rate constant , but can produce statistically exact results in practice provided that a sufficiently large value is used ( Figs . 5–8 , panels C and D ) . As currently implemented , the primary advantage of the HPP method is in reducing memory usage during simulation ( Figs . 5–8 , panels A ) . Importantly , this is accomplished with little to no impact on simulation run time ( Figs . 5–8 , panels B ) . We have shown that peak memory use for HPP scales linearly with particle number ( with a slope that is smaller than for NFsim; Figs . 5–8 , panels A ) and confirmed that when network generation is possible SSA memory use is approximately independent of particle number ( Figs . 7A and 8A ) . At the system volumes that we have considered here , HPP memory use is significantly less than for SSA . However , the linear scaling of HPP and the constant scaling of SSA indicate that with further increases in the system volume there will invariably come a point where HPP memory use exceeds that of SSA . This is because species that are rare at small volumes , and hence chosen to be treated as particles , become plentiful at large volumes . Intuitively , a partially-expanded network should never require more memory than a fully-enumerated network . However , as currently implemented , there is no way to strictly enforce this restriction because HPP requires that population species be chosen prior to PNE . In Fig . 9 , we have shown how a systematic approach to choosing population species can optimize memory usage for a given system volume . However , this approach requires running an NFsim pre-simulation , which may not be feasible for systems with extremely large numbers of particles ( e . g . , whole cells ) . Thus , we propose to develop a more general version of HPP that dynamically tracks the populations of species during the course of a simulation and automatically selects those to treat as population variables based on some criteria , e . g . , that their population exceeds a certain threshold . In this automated version of HPP ( aHPP ) , PNE would be performed every time a new species is lumped . If all species in the system become lumped then the network will naturally become fully enumerated . Hence , the memory load will never exceed that of the fully-expanded network . In Fig . 10 , we provide a qualitative sketch of how we expect the memory usage of this hypothetical aHPP method to scale with system volume ( particle number ) . Included for comparison are scalings for HPP , NFsim , and SSA . For models with finite networks ( such as and EGFR ) , aHPP memory use should plateau once the entire reaction network has been generated . For models with infinite networks ( such as TLBR and Actin ) , we expect aHPP memory use at large volumes to scale somewhere between constant and linear ( no worse than HPP ) depending on the model . A detailed analysis of the space complexity of a hypothetical , “optimal” aHPP method is provided in Sec . S2 of supplementary Text S1 . In order to frame our results within a real-world context , we have estimated the cost of simulation based on hourly rates of on-demand instances on the Amazon Elastic Compute Cloud ( EC2 ) . In Fig . 11 , we show the hourly cost ( per “effective compute unit” ) of simulation as a function of required memory per simulation ( details of the calculation can be found in Sec . S1 of Text S1 ) . Also included in the plot are values for HPP ( ) , NFsim ( ) , and SSA ( ) simulations of the EGFR model at cell fraction ( Fig . 8A ) . Our calculations show that below of required memory High-CPU instances are the most cost effective . Above this threshold High-Memory instances are the better option . The HPP simulation falls below this cutoff while both NFsim and SSA lie above . There is a quantifiable benefit , therefore , to reducing memory usage in this case; HPP simulations on the EC2 would be and times less expensive , respectively , than NFsim and SSA ( HPP is slightly faster than NFsim and significantly faster than SSA; Fig . 8B ) . Thus , the reduction in memory usage offered by HPP is not simply of academic interest but can impact , in a tangible way , the cost of doing computational research . Finally , even greater benefits are possible if , in addition to reducing memory usage , the speed of HPP simulations can be increased . leaping [36] , [77]–[79] is an approach for accelerating stochastic simulations of chemically reactive systems . With a few exceptions ( e . g . , Ref . [80] ) , leaping has been applied primarily to fully-enumerated reaction networks . We believe that the HPP method provides a unique setting for the application of because , unlike in pure particle-based methods , there exists a partial network of reactions that act on population species . Thus , a network-based leaping method can be applied exclusively to the population component of a system while retaining the network-free approach in the particle component . We have recently implemented a leaping variant in BioNetGen , known as the partitioned-leaping algorithm [22] , and are actively working on integrating it with the HPP .
|
Rule-based modeling is a modeling paradigm that addresses the problem of combinatorial complexity in biochemical systems . The key idea is to specify only those components of a biological macromolecule that are directly involved in a biochemical transformation . Until recently , this “pattern-based” approach greatly simplified the process of model building but did nothing to improve the performance of model simulation . This changed with the introduction of “network-free” simulation methods , which operate directly on the compressed rule set of a rule-based model rather than on a fully-enumerated set of reactions and species . However , these methods represent every molecule in a system as a particle , limiting their use to systems containing less than a few million molecules . Here , we describe an extension to the network-free approach that treats rare , complex species as particles and plentiful , simple species as population variables , while retaining the exact dynamics of the model system . By making more efficient use of computational resources for species that do not require the level of detail of a particle representation , this hybrid particle/population approach can simulate systems much larger than is possible using network-free methods and is an important step towards realizing the practical simulation of detailed , mechanistic models of whole cells .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"systems",
"biology",
"biochemistry",
"biochemical",
"simulations",
"computer",
"and",
"information",
"sciences",
"computer",
"modeling",
"biology",
"and",
"life",
"sciences",
"computational",
"biology"
] |
2014
|
Exact Hybrid Particle/Population Simulation of Rule-Based Models of Biochemical Systems
|
Immunizing human volunteers by mosquito bite with radiation-attenuated Plasmodium falciparum sporozoites ( RAS ) results in high-level protection against infection . Only two volunteers have been similarly immunized with P . vivax ( Pv ) RAS , and both were protected . A phase 2 controlled clinical trial was conducted to assess the safety and protective efficacy of PvRAS immunization . A randomized , single-blinded trial was conducted . Duffy positive ( Fy+; Pv susceptible ) individuals were enrolled: 14 received bites from irradiated ( 150 ± 10 cGy ) Pv-infected Anopheles mosquitoes ( RAS ) and 7 from non-irradiated non-infected mosquitoes ( Ctl ) . An additional group of seven Fy- ( Pv refractory ) volunteers was immunized with bites from non-irradiated Pv-infected mosquitoes . A total of seven immunizations were carried out at mean intervals of nine weeks . Eight weeks after last immunization , a controlled human malaria infection ( CHMI ) with non-irradiated Pv-infected mosquitoes was performed . Nineteen volunteers completed seven immunizations ( 12 RAS , 2 Ctl , and 5 Fy- ) and received a CHMI . Five of 12 ( 42% ) RAS volunteers were protected ( receiving a median of 434 infective bites ) compared with 0/2 Ctl . None of the Fy- volunteers developed infection by the seventh immunization or after CHMI . All non-protected volunteers developed symptoms 8–13 days after CHMI with a mean pre-patent period of 12 . 8 days . No serious adverse events related to the immunizations were observed . Specific IgG1 anti-PvCS response was associated with protection . Immunization with PvRAS was safe , immunogenic , and induced sterile immunity in 42% of the Fy+ volunteers . Moreover , Fy- volunteers were refractory to Pv malaria . Identifier: NCT01082341 .
Although there has been a decrease in malaria incidence globally during the past 15 years ( ~37% ) [1] , this infection remains a major public health problem with 214 million cases and 438 , 000 deaths estimated in 2015 [1] . Plasmodium falciparum ( Pf ) causes the greatest malaria burden particularly in Africa , and is the focus of most attention , including the search for a vaccine . Recently , a vaccine based on the Pf circumsporozoite ( CS ) protein ( RTS , S ) received a positive decision by the European Medicines Agency ( EMA ) for potential use in African children to reduce episodes of clinical malaria , based on the results of phase 3 studies , while the World Health Organization ( WHO ) recommended feasibility and pilot effectiveness implementations [2] . Protection afforded by RTS , S is limited to reduction of clinical disease in infants and young children; the vaccine is not intended for older children or adults , for use in Europe or the USA , or to block infection or prevent transmission . Plasmodium vivax ( Pv ) is the second most abundant malaria parasite , posing a serious threat in Asia , Oceania , and Latin America and also requires a specific and effective vaccine . Progress in developing Pv vaccines lags far behind that for Pf . Acquisition of clinical immunity to malaria is a slow process and sterile immunity is never achieved under natural conditions , although it can be reproducibly induced by immunization via mosquito bite with radiation-attenuated sporozoites ( RAS ) , the parasite stage transmitted by mosquitoes to humans [3–5] . This approach induces immune responses that block the sporozoite ( SPZ ) invasion of hepatocytes and subsequent schizogonic development in the liver , thereby preventing the pathogenic asexual blood stage infection that causes malaria disease . Such responses also prevent the development of gametocytes ( sexual blood stages ) ; thus , RAS immunization could serve as a vaccine to interrupt malaria transmission . Pre-erythrocytic stage vaccines such as RAS , therefore represent an ideal approach for vaccine development [6] as has been reported previously for Pf [7] . In the 1970s , sterile immunity against malaria was first demonstrated in humans vaccinated using RAS [3 , 4 , 8] . Since then , multiple studies have confirmed the high reproducibility of this vaccination model [9 , 10] . Significant efforts are now being invested and good progress has been achieved in developing a parenterally injectable vaccine based on cryopreserved PfRAS [7] . Several PfRAS phase 1 and 2 trials have been conducted by Sanaria Inc . and collaborators , using a PfSPZ vaccine , a GMP product consisting of aseptic , purified , radiation-attenuated , cryopreserved PfSPZ . This vaccine has shown high-level efficacy in naïve adults [7] . Additionally , several parasite antigens found to be active in RAS immunization and possibly associated with protection have been the subject of intense research on the development of subunit vaccines ( reviewed in [11] ) . Despite the epidemiological importance of Pv , the PvRAS model has not been reproduced since the early 1970s , when two volunteers were immunized by receiving >1000 mosquito infectious bites; both were protected from infectious Pv spz challenge [12] . This lag is partly explained by the lack of Pv in vitro culture methods , promoting the development of alternative , more complex infection methods that rely on obtaining fresh , gametocytemic blood from Pv-infected donors . Anopheles mosquito colonies have been established [13] and methods to routinely infect mosquitoes using blood from acutely ill Pv malaria patients have now been standardized [14] , resulting in safe , reliable and reproducible infection of human volunteers through mosquito bites [15–17] . The purpose of the study described here was first to establish a solid proof-of-principle that humans could be protected by immunization via the bites of PvRAS-infected mosquitoes and second , to obtain sera and cells to study the mechanisms of protective immunity and identify the antigenic targets of immune responses . A phase 2 trial was conducted in healthy adult Colombian volunteers without previous exposure to malaria .
This trial was conducted according to ICH E-6 Guidelines for Good Clinical Practices [18] . Institutional Review Boards of the Malaria Vaccine and Drug Development Center ( MVDC , CECIV ) , and Centro Médico Imbanaco ( CEICMI ) , Cali , approved the protocol . Written informed consent ( IC ) was obtained from all volunteers , with a separate IC for HIV screening . The clinical trial was registered on clinicaltrials . gov , registry number NCT01082341 . The protocol for this trial and supporting CONSORT checklist are available as supporting information ( S1 Checklist and S1 Protocol ) . A phase 2 controlled randomized , single-blinded clinical trial was conducted at the MVDC , Cali , Colombia . A total of 89 malaria-naïve volunteers ( 18–45 years old ) were assessed for eligibility ( Fig 1 ) . Two approaches to immunization were used in this study . First , Duffy-positive ( Fy+ ) individuals were assigned to RAS or mock-immunized control groups using a single-blinded design ( volunteers but not investigators blinded ) to assess the safety , tolerability , immunogenicity and protective efficacy of PvRAS immunization . Second , taking advantage of the fact that Fy- erythrocytes are refractory to Pv invasion , a third group of Fy- volunteers was immunized with bites from infected non-irradiated mosquitoes to assess the impact of exposure to PvSPZ developing fully in the liver ( as opposed to arresting early in liver stage development , as in the case of RAS ) . Immunization was performed by direct exposure to bites of irradiated ( Fy+ volunteers ) or non-irradiated ( Fy- volunteers ) Pv-infected mosquitoes , and mock immunization by exposure to the bites of non-irradiated , non-infected mosquitoes . After the immunization schedule , volunteers were subjected to a Pv controlled human malaria infection ( CHMI ) , carried out by exposing volunteers to the bites of non-irradiated , Pv-infected mosquitoes . Clinical outcome , parasitemia as measured by thick blood smear microscopy ( TBS ) , and clinical laboratory and immunological parameters were assessed . Antimalarial treatment was provided to all volunteers becoming TBS-positive or completing the study to day 60 post-CHMI . Volunteers were informed about the risks of participation and were provided sufficient opportunity to read the IC forms . Before signing the IC , volunteers had to pass an oral or written exam concerning the trial and its risks as described elsewhere [16] . In addition , all participants were informed about their right to voluntarily withdraw from the study at any time for any reason . Exclusion criteria included pregnancy , abnormal clinical hematology , and chemistry test results , glucose-6-phosphate dehydrogenase deficiency ( G6PDd ) , and infectious diseases ( syphilis , HIV , Chagas disease , HTLV 1–2 , hepatitis B and hepatitis C; S1 Table , S2 Table , S1 text ) . Anopheles albimanus mosquitoes reared at the MVDC insectary in Cali were infected with blood from Pv-infected patients ( 18–45 years old ) recruited at outpatient clinics in malaria-endemic areas of Colombia . TBS was performed on all volunteers seeking care for malaria diagnosis as required by the National Malaria Control Program . Only volunteers who tested positive by this method were invited to participate in the study and were informed of the research aims , potential risks , and benefits . After signing the IC and before the antimalarial treatment , whole blood ( 35 mL ) was collected by venipuncture . All samples were confirmed to be Pv malaria mono-infections by quantitative PCR ( qPCR ) and negative for other infectious agents ( syphilis , HIV , Chagas disease , HTLV 1–2 , hepatitis B and hepatitis C; S2 Table ) . Mosquitoes were membrane-fed with infected blood as described previously [19] . Batches with >50% mosquitoes harboring spz in their salivary glands were used for immunization and CHMI . For both procedures , individual screen-meshed boxes containing infected mosquitoes were used . Mosquitoes were allowed to feed on the volunteer for a 5–10 minute period as previously standardized [14] . After biting , all mosquitoes were dissected and microscopically examined to confirm the presence of blood meal and spz in the salivary glands . CHMI of all volunteers was carried out on the same day by exposing volunteers to bites of 2–4 mosquitoes infected with the same parasite isolate [15–17] . Infected bites were calculated as the number of fed mosquitoes times the percentage infected . Sporozoite attenuation was performed by exposure of Pv-infected mosquitoes to 150 ± 10 cGy of gamma radiation using a Varian Clinac IX Series 927 linear accelerator at the radiotherapy unit of Hospital Universitario del Valle in Cali as previously described [20] . The primary objective of the study was the immunization and CHMI of all volunteers using mosquitoes as described above . Fy+ volunteers were assigned to either RAS ( n = 14 ) or Ctl ( n = 7 ) groups , and Fy- volunteers to the Fy- group ( n = 7 ) . A total of seven immunizations were carried out using for each immunization a mean of 65 infectious mosquito bites . Two weeks after the last immunization , all volunteers were treated orally with curative doses of chloroquine ( 600 mg on day one and 450 mg on days two and three ) and primaquine ( 30 mg daily for 14 days ) to eliminate any subpatent malaria infections that may have developed during the immunization period , so that incident infections from CHMI could be accurately determined . Plasma levels of chloroquine and primaquine were measured by high-performance liquid chromatography ( HPLC; [21] ) two weeks prior to CHMI , to ensure drug clearance . Eight weeks after the last immunization , and one month after completing antimalarial treatment , all volunteers received CHMI using 2–4 Pv-infected mosquito bites . Physical examination , clinical laboratory , and immunological tests were performed after every immunization and CHMI ( Fig 2 ) . Adverse events ( AE ) were recorded , graded and classified according to FDA recommendations [22] . Whole blood was collected by venipuncture of the arm at inclusion ( baseline ) , ten days after each immunization , before CHMI , and six months post-CHMI for clinical laboratory and immunological tests . After each immunization , volunteers were followed-up on days 1 , 2 and 10 in person for a physical examination and by phone on days 7 and 14 . Likewise , after CHMI , volunteers were followed up every day by phone until day 5 and malaria infection monitored daily in an outpatient clinic from day 6–28 post-CHMI; thereafter twice a week from day 29–60 post-CHMI for volunteers who did not develop fever or patent infection within 28 days post-CHMI as determined by TBS microscopy and qPCR . Additionally , volunteers were encouraged to visit the Centre for medical consultation at any moment if they developed any symptom or had any concern . Treatment was initiated immediately after parasitemia was confirmed by TBS and the volunteers followed-up until three consecutive TBS resulted negative . Afterward , volunteers had TBS assessed on days 7 , 14 , 21 and 45 post-treatment to confirm cure and absence of relapse [15–17] . Serum and plasma were stored at -20°C until use . Peripheral blood mononuclear cells ( PBMC ) obtained by Ficoll density gradient centrifugation were stored in liquid nitrogen until use . Vaccine efficacy was assessed by prevention of patent parasitemia . Infection was diagnosed by TBS examination by two independent experienced microscopists , and parasitemia determined by counting the number of asexual Pv parasites per 400 white blood cells ( WBC ) , assuming normal WBC counts ( 8 , 000 cells/μL ) . Samples were considered negative after observation of 200 microscopic fields and qPCR was performed subsequently for retrospective analyses . Clinical laboratory tests were periodically performed during immunizations and as required by clinical judgment after the CHMI to ascertain health status ( same methods as recruitment screening tests , S1 Table ) . A secondary outcome was the evaluation of humoral immune responses . Specific antimalarial antibodies ( Ab ) were determined by enzyme-linked immuno-sorbent assay ( ELISA ) . The presence of IgG to PvCS ( NRC and N peptides ) and to merozoite surface protein-1 ( PvMSP-1 ) was assessed in sera diluted 1:200 as previously described [17] . PvCS corresponded to a chimeric synthetic polypeptide composed of the amino ( N ) flank , the VK210 and VK247 natural repeat variants ( R ) , and the carboxyl ( C ) flanking sequences of the protein [23]; PvMSP-1 corresponded to a recombinant fragment from the N region of the protein , namely r200L [24] . IgG isotypes against PvCS-NRC peptide were detected using mouse monoclonal Abs to specific to human IgG1 , IgG2 , IgG3 and IgG4 ( Sigma-Aldrich ) , followed by HRP-conjugated anti-mouse . In all cases , the optical density ( OD ) was measured using a BioTek ELISA Reader ( BioTek , Winooski , VT ) . Cut-off values were calculated as three SD above the mean OD value of negative control sera . Results were expressed as reactivity index ( RI ) , defined as optical density ( OD ) values of test sample divided by the cut-off value . Immunofluorescence tests ( IFAT ) were used to assess the Ab reactivity with PvSPZ . To determine the frequency of T cells responding to P . vivax antigens , IFN-γ production was quantified using an ELISpot assay . Briefly , the assay was performed in multiscreen 96-well plates ( MAHAS 4510 , Millipore ) coated with anti-human IFN-γ capture antibody ( 1-D1K; Mabtech AB ) . Fresh PBMC collected 12 days previous to the CHMI were plated into duplicate wells at 4 x 105 cells in complete RPMI-1640 medium ( cRPMI; Gibco , Invitrogen ) supplemented with 10% FBS . The PBMC were stimulated for 40 h at 37°C with 10 μg/mL of PvSPZ lysate , PvCS-NRC or PvTRAP ( thrombospondin-related adhesive protein ) . cRPMI medium-only and PHA controls were used in all assays . Biotinylated anti-IFN-γ antibody ( 7-B6-1; Mabtech AB ) was added followed by alkaline phosphatase-streptavidin conjugate ( Mabtech AB ) . Spots were visualized by adding BCIP/NBT ( Sigma-Aldrich ) , scanned and counted using the AID ELISpot reader ( AID Autoimmun Diagnostika GmbH , Germany ) to determine the number of spots/well . Results were expressed as spots per 106 PBMC , normalized by the antigen-stimulated spots less cRPMI medium . Data were collected and managed using REDCap ( Nashville , TN , USA ) electronic data capture tools , analyzed using SPSS version 16 . 0 software ( SPSS Inc . , Chicago , IL , USA ) , and plotted using GraphPad Prism version 6 . 0 ( GraphPad Software , San Diego , California , USA ) . We estimated a sample size of 21 Fy+ individuals ( 2:1 , RAS to Ctl ) at a 5% significance level and 80% power to assess the protective efficacy of immunization . Nominal variables were analyzed using descriptive statistics . Mann-Whitney U or the Kruskal-Wallis tests were used as needed . Fisher's exact test was used to compare proportions . Spearman’s rank correlation ( rs ) was used to assess the correlation between numeric variables . Incubation and pre-patent periods were determined by TBS and qPCR and visualized using Kaplan–Meier estimator . A p value < 0 . 05 was considered statistically significant .
A total of 28 of the screened volunteers were enrolled and began the immunization schedule between Sept 26 , 2013 , and Feb 15 , 2014 . However , only 19 completed the schedule ( Fig 1 and Fig 2 ) . Mean age at enrollment was 30 , 29 and 25 years , and the male/female ratio was 5:9 , 5:2 , 0:7 for the RAS , Ctl , and Fy- groups , respectively ( Table 1 ) . A total of seven immunizations were carried out at mean intervals of nine weeks ( range 3–25 weeks ) in volunteers who then continued to complete the CHMI . The RAS and Fy- groups received a median of 434 ( range 362–497 ) and 476 ( range 358–487 ) total infective bites over the seven immunizations , respectively , whereas the Ctl group received 954 ( range 945–963 ) non-infective ( placebo ) bites during the immunization protocol . The total number of infective bites , non-infective bites , fed mosquitoes , and spz in salivary glands per volunteer were determined by post-feeding salivary gland dissection and microscopy examination ( S3 Table ) . No volunteer developed clinical malaria or parasitemia by TBS during the immunization phase , although low levels of parasite DNA were detected in peripheral blood by qPCR from day 8–16 after immunizations in the Fy- group , which declined after every subsequent immunization ( Fig 3A ) . At the time of the CHMI , all volunteers had cleared both primaquine and chloroquine in plasma , although two volunteers in the RAS group had low detectable levels of chloroquine two weeks prior to the CHMI . Notably , both volunteers developed malaria infection . Seven to nine days after the first immunization , 1/14 and 5/7 volunteers of the RAS and Fy- groups , respectively , developed fever , chills , headache and profuse sweating consistent with malaria , which lasted 1–2 days . All five symptomatic Fy- volunteers had negative TBS but positive qPCR that resolved spontaneously , whereas the RAS volunteer was negative by TBS and qPCR . Headache and local reaction in the immunization site were the most common AE during initial immunizations with decreasing frequency throughout the immunizations ( S4 Table ) . After CHMI headache , chills , fever , and malaise were common AE ( Fig 4 ) . In the RAS and Ctl groups , a mean of 11 and 16 AE per individual were reported after CHMI , respectively . In contrast , in the Fy- group a mean of two AE was reported . No serious AE related to immunizations were observed , although one female developed severe elevation of hepatic transaminases after CHMI ( >10 times upper limit of normal [xULN] ) and lactic dehydrogenase ( 2 . 5xULN ) with abdominal pain , nausea , and vomiting during Pv malaria mono-infection . This patient was observed in the emergency room and completely recovered without sequelae . No alternative etiologies for the elevated transaminases were identified ( volunteer was negative for hepatitis C , hepatitis B , HIV and hepatitis A , and she was not consuming any medications ) . At day 60 post-CHMI ( last day of follow-up ) , the total protective efficacy in the RAS group was calculated at 42% ( 5/12 RAS , 0/2 Ctl ) as determined by TBS and confirmed by qPCR; all protected subjects were women ( Table 1 ) . All malaria-positive volunteers presented with low parasitemia , with median density values lower by TBS than by qPCR ( TBS: 140 parasites/μL; IQR 95–210 , and qPCR: 220 parasites/μL; IQR 29 . 2–361 ) . Mean incubation period was 9 . 9 days ( range 8–13 ) ; mean prepatent period was 12 . 8 days ( range 12–13 ) by TBS; and 9 . 0 days ( range 8–11 ) by qPCR . No significant differences were found between the Ctl and positive RAS subjects in prepatent period or density of parasitemia by TBS or qPCR ( Fig 3B ) . However , survival analysis showed a significantly greater incubation period in RAS than in Ctl volunteers ( S1 Fig ) . These results were compared with the parasite dynamics of a previous CHMI trial carried out in naïve volunteers using the same infection protocol . Those volunteers who did not develop malaria were followed up until day 60 post-CHMI after which antimalarial treatment was administered . Volunteer 001 of the RAS group developed malaria-related symptoms at day 58 post-CHMI , but parasitemia was only detected on day 66 by TBS . This prepatent period was considered as partial protection induced by vaccination . Seroconversion using the PvCS-NRC peptide was observed in all 12 RAS volunteers , mostly after the second immunization ( 10/12 ) and in all Fy- volunteers between the second and fifth immunizations . In both groups , IgG reactivity was low ( RI < 6 ) ; all Ctl volunteers remained seronegative during the immunization phase ( Fig 5A and 5B ) . A positive correlation between the RI for PvCS-NRC and number of infective bites was observed for the Fy- group but not for the RAS group ( Fig 5C ) . No significant association between total anti PvCS-NRC RI and protection was found ( Fig 5D ) ; however , the specific IgG1 response was significantly higher in protected individuals ( Fig 6A and 6B ) . All Fy- volunteers and one in RAS group developed anti-PvMSP-1 IgG response after seven immunizations . In contrast , all Ctl volunteers remained negative for all antigens tested ( S2 Fig ) . After immunization , 11/12 of RAS and 4/5 of Fy- volunteers had IFAT Abs to Pv spz , respectively , but no association with protection was found ( S5 Table ) . Moreover , all RAS and Fy- sera recognized PvCS by Western blot ( S3 Fig ) . After seven immunizations and before to CHMI , PMBCs of the RAS group were able to produce IFN-γ after stimulation with the tested antigens Pv spz lysate , PvCS-NRC , and PvTRAP ( Fig 7A ) at significantly higher levels than the other two groups ( p <0 . 05 for all antigens ) . In the Fy- group , PvCS-NRC and PvTRAP induced IFN-γ production but was not significantly higher than the observed in the Ctl volunteers ( Fig 7B ) . No significant differences were observed between protected and not protected volunteers in the RAS group ( Fig 7C ) .
This trial has allowed the establishment of the PvRAS immunization model with protection in an unprecedented number of volunteers . To our knowledge , only two volunteers had been previously reported to be protected from CHMI by PvRAS immunization [12] . As is true for PfRAS , immunization by mosquito bite with the PvRAS is safe , immunogenic and able to induce sterile protection . A series of clinical trials conducted with PfRAS indicated high protective efficacy ( ~90% ) and protection lasting up to 42 weeks with a dose-dependent efficacy based on ten immunization sessions and a total of ~1000 RAS mosquito bites [3–5 , 8 , 9 , 12] . This study could not reproduce those conditions due to the difficulty of obtaining PvRAS , which include the need of regular P . vivax infected donors from malaria endemic areas , willing to participate and complying with all inclusion criteria . In addition , not all P . vivax samples are successfully infective to An . albimanus mosquitoes due to numerous biological factors [25] . Nevertheless , seven immunization sessions provided a median of 434 PvRAS bites for an efficacy of 42% . This is similar to what has been found with Pf immunization , where the protective efficacy against CHMI in volunteers receiving < 1000 infectious bites was 40% [9] . Despite the high number of volunteers that withdrew from the Ctl group , the two remaining showed a trend in the parasite dynamics similar to that observed in a total of 29 naïve Fy+ volunteers infected with 2–4 An . albimanus mosquito bites in three previous CHMI trials [15–17] . In one of these trials [17] , even semi-immune volunteers from endemic areas previously exposed to natural malaria infection developed similar parasite patency , indicating the relevance of the sterile protection induced here by PvRAS immunization . Therefore , given that both controls turned positive following CHMI , we are confident that PvRAS immunization induced sterile protection as described in RAS group . A summary of parasitological data for naïve volunteers participating in previous CHMI carried out in our Centre , which demonstrate the consistency of this procedure , is shown ( Table 2 ) . Since no detectable levels of chloroquine were observed in most of RAS volunteers , and the two volunteers with low detectable levels had undetectable levels immediately previous to CHMI and developed malaria infection , we concluded that lack of parasitemia was not dependent of chloroquine antiplasmodial activity . Significant progress has been achieved regarding the development of a practical approach for Pf immunization based on whole SPZ [7] . Intravenous administration of aseptic , purified , cryopreserved , radiation-attenuated PvSPZ [7] has shown the highest efficacy , protecting up to 100% of study subjects . Based on the results of our study , we can anticipate that a similar Pv product would be equally protective , and could potentially be combined with Pfspz to induce potent immunity to the two major species of human malaria . The use of the whole SPZ approach may therefore be an effective route to solving the malaria problem , given that subunit vaccines appear to have a longer development trajectory . The RTS , S vaccine , the most advanced Pf malaria subunit vaccine , has been assessed as meeting EMA standards , setting the stage for potential licensure in African countries , and has subsequently been recommended by WHO for testing in pilot implementations in Africa [2] . Several other vaccine candidates are also under development [26] . RTS , S is to be licensed for reducing the incidence of clinical malaria but not preventing malaria infection , and is insufficiently potent for use in elimination campaigns . Progress in subunit vaccines for Pv has been especially limited due to the lack of an in vitro culture method and the scarce funding . In order to fill this gap , we approached the development of PvRAS by accessing to fresh gametocytemic blood from patients to assess the PvRAS model’s feasibility and reproducibility under controlled conditions , and to generate immune reagents to determine correlates of protection . Additionally , this study took the novel approach of immunizing Fy- volunteers by repeated exposure to viable PvSPZ . Because most Fy- individuals are refractory to blood infection by Pv , this allowed evaluation of immune responses elicited specifically against liver-stage parasites . Although there have been reports from Madagascar and Cameroon-endemic areas that some of these subjects may develop the blood cycle when infected by Pv [27 , 28] , this did not happen in our study with different natural parasite isolates . To our knowledge , this was the first time that Fy- volunteers have been used as a model for a better understanding of the immune responses to Pv liver stages . The presence of symptoms in Fy- following only the first immunization , and the diminishing qPCR positivity as immunizations continued , indicate that the Fy- volunteers developed sterile immunity to Pv infection based on immunity targeting the pre-erythrocytic stages . Reagents generated in this study allow the use of both classic and high throughput methods to analyze the immune response to PvRAS , and comparison of responses to the early liver stages in the PvRAS and Fy- groups . Sera and cells are currently being studied using high throughput systems in an attempt to determine correlates of immune protection . Interestingly , all protected volunteers were women , whereas all men developed malaria despite receiving similar parasite doses . No covariates , such as numbers of immunizing bites , were identified to explain this finding . This is consistent with other studies where women mounted a more vigorous immune response than men ( reviewed in [29] ) , although this was not evident here at least for the parameters evaluated , which may or may not serve as correlates of protection . We achieved the doses necessary to protect almost all challenged women ( 5/7 ) but not men ( 0/5 ) . Immunization with both PvRAS and viable Pv spz induced a measurable although weak ELISA antibody response to PvCS , and there was no association between total IgG Ab levels to PvCS and protection . Nonetheless , the protected volunteers had a greater IgG1 response against PvCS-NRC peptide , which is in agreement with studies describing associations between higher levels of IgG1 and IgG3 Abs and protection against severe Pf malaria episodes [30] as well as predominant markers for exposure to Pv malaria [31] . However , the borderline p value ( p = 0 . 048 ) for the association between PvCS-NRC peptide ELISA titer and protection was possibly due to the relatively small number of individuals tested . All Fy- individuals and only one PvRAS developed Ab levels against PvMSP-1 protein after seven immunizations . This appears to be in agreement with the fact that Pv is able to completely develop the liver cycle and release merozoites into circulation in Fy- volunteers as demonstrated by qPCR , whereas RAS appears to arrest development in early phases of the liver cycle [32] . It is also consistent with the fact that Fy- volunteers developed fever and other malaria symptoms during the first immunizations and the fact that parasite DNA was detected up to the fourth immunization . The decrease in anti-CS Ab levels after the third and fourth immunizations , when there was a pause of several months in immunizations , indicates that these Abs are short-lived , although memory cell responses were present as demonstrated by the rapid boosting of specific Abs after the fifth immunization . This trial confirms the reproducibility of the RAS vaccination model in Pv malaria . Despite the lack of correlation between protection and the tested immune responses , high throughput analyses of cells and sera , i . e . , transcriptomics and anti-parasite Ab microarray profiles , may offer a better understanding of the parasite targets involved and the immune effector mechanisms associated with protection .
|
Despite the advances in Plasmodium falciparum ( Pf ) vaccine development , progress in developing P . vivax ( Pv ) vaccines lags far behind . Immunization via mosquito bites with Pf radiation-attenuated sporozoites ( RAS ) has been the gold standard model for induction of sterile protection against malaria infection and has allowed the study of the complex mechanisms of immunity . The first trials using PfRAS were performed in the late 1960’s , and thereafter greatly contributed to the development of vaccines against Pf . However , PvRAS immunization in humans has only been carried out in two volunteers since 1974 . To our knowledge , this is the first clinical trial using significant numbers of volunteers for PvRAS immunization . Our findings confirm that immunization with PvRAS is safe , immunogenic and induces sterile immunity in 42% of the volunteers . It demonstrates that it is possible to induce sterile protection with PvRAS as seen with PfRAS and confirms that immunity against the PvCS protein ( IgG1 levels ) correlates with protection . Research findings and reagents generated in this study are expected to yield insights on key immune determinants of sterile protection against Pv , which may guide the development of a cost-effective vaccine against this parasite species .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"animals",
"parasitology",
"vaccines",
"preventive",
"medicine",
"parasitemia",
"bacterial",
"diseases",
"protozoans",
"vaccination",
"and",
"immunization",
"insect",
"vectors",
"immunologic",
"techniques",
"quantitative",
"parasitology",
"research",
"and",
"analysis",
"methods",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"malarial",
"parasites",
"tuberculosis",
"epidemiology",
"immunoassays",
"disease",
"vectors",
"insects",
"arthropoda",
"mosquitoes",
"biology",
"and",
"life",
"sciences",
"malaria",
"organisms"
] |
2016
|
Protective Efficacy of Plasmodium vivax Radiation-Attenuated Sporozoites in Colombian Volunteers: A Randomized Controlled Trial
|
Ideal decision-makers should constantly assess all sources of information about opportunities and threats , and be able to redetermine their choices promptly in the face of change . However , perpetual monitoring and reassessment impose inordinate sensing and computational costs , making them impractical for animals and machines alike . The obvious alternative of committing for extended periods of time to limited sensory strategies associated with particular courses of action can be dangerous and wasteful . Here , we explore the intermediate possibility of making provisional temporal commitments whilst admitting interruption based on limited broader observation . We simulate foraging under threat of predation to elucidate the benefits of such a scheme . We relate our results to diseases of distractibility and roving attention , and consider mechanistic substrates such as noradrenergic neuromodulation .
The full model of foraging under predation risk that we consider below includes a number of detailed components . Therefore , in order to illustrate and provide intuition for the workings of interruption , we start with a simple , stripped-down model . Consider an animal ( e . g . , a rat ) foraging in a habitat ( Fig 1A ) in which a predator ( e . g . , a hawk ) arrives with probability 0 . 01 per unit time . The animal has to choose between three actions: ( i ) to continue to feed; ( ii ) to stop feeding to assess whether a predator is present; or ( iii ) to escape . Opting to feed has the benefit of gaining resources ( worth 1 unit of reward per unit time ) , but has the drawback of providing only poor information ( + ) about whether the predator is present ( e . g . , because it restricts the animal’s field of view ) . Furthermore , if the predator is present , then the animal gets caught with probability 0 . 1 per unit time , incurring a cost of −100 units of reward , and causing the task to terminate . By contrast , assess provides the animal with improved information ( ++ ) about the presence of the predator , but has reward 0 per unit time . The exact details of how the quality of information is varied shall be described below , but the important point here is that feed and assess have distinct informational consequences . Finally , choosing to escape leads to a location that is safe , but does not permit foraging , thus gaining 0 reward; it also causes the task to terminate . The animal’s assumed task is to maximize the expected sum of its undiscounted future rewards ( we consider the more conventional case of long run discounted rewards when we amplify this simple example ) . The animal here has to negotiate two fundamental tradeoffs . One is between the rewards that can be directly obtained by choosing to feed and the improved information that can be obtained by choosing to assess . The second tradeoff is between future possible rewards that may be obtained by remaining in the habitat , and the sacrifice of those rewards in favour of safety by choosing to escape . The final detail of the example is that each time the animal makes a choice about what to do , it selects both an action a ( i . e . , feed , assess , or escape ) and a duration τ with which to perform it . Critically , each time the animal makes a decision it may incur a decision cost , cd ≥ 0 , which summarizes the various computational demands associated with such a decision; as discussed above , these may include planning and/or switching costs . If this decision cost is non-negligible , the animal additionally faces a tradeoff between reducing decision costs by making a prolonged temporal commitment to an action—i . e . , reducing decision frequency by choosing a long duration τ—and being able to respond quickly if it is likely that the predator has arrived ( Fig 1B ) . Whether or not the animal is able to interrupt its ongoing behaviour is a critical determinant of its optimal policy . Here , a policy is a mapping between the degree to which the animal believes that the predator has arrived/is present , β ( P ) , and an action-duration pair ( a , τ ) . Fig 1C–1F display optimal policies , comprising optimal actions a* ( left ) and associated optimal durations τ* ( right ) , for four different cases . The first ( Fig 1C ) is the optimal policy when there is no decision cost , cd = 0 . In terms of actions , the animal chooses to feed when it is unlikely that the predator has arrived ( β ( P ) low ) , to escape when this is more likely than not ( β ( P ) > 0 . 5 ) , and to assess otherwise . More importantly for our purposes is the observation that optimal durations τ* are uniformly chosen to be as short as possible . Since making a decision has no cost , doing so at the highest possible frequency is the best way to ensure that the animal is best able to respond if it thinks the predator may have arrived . Next , we consider the case where there is a small decision cost , cd = 0 . 01 , and the animal is unable to interrupt its own activity ( Fig 1D ) . That is , if the animal chooses to engage in an action for τ seconds , it will be fully committed to performing the action for that duration . Here , there is no change in choice of actions , and very little change in duration except a slight increase in τ* for feed when the predator is very unlikely to have arrived . The low durations here , even though this entails a high frequency of decisions , mean that the animal is willing to pay these decision costs in order to remain responsive to possible threat . In Fig 1E , the decision cost remains the same ( cd = 0 . 01 ) , but the animal is now able to interrupt its behaviour . Interrupts also occur as a function of β ( P ) , but ( a ) as noted , the information available to change β ( P ) is of lower quality during feed than assess; and ( b ) we assume that interruption itself is free ( although there is then a standard decision cost associated with the necessary re-planning ) . Again , we see no change in which actions are chosen , but now both feed and assess are chosen to have the longest possible duration ( in this example , there is no advantage to the animal of choosing to spend longer on escape ) . Since the animal can always interrupt itself , making such provisional commitments means that it can minimize decision costs by interrupting and making a new decision only when strictly necessary . Finally , if the decision cost is increased further , changes in the pattern of optimal actions are observed . Fig 1F shows the optimal policy for cd = 1 , showing an increased propensity of the animal to escape . The alternatives are increasingly disfavoured by the animal not because of any change in risk of predation—this is unchanged—but because of the increased expense of making on-going decisions if it remains in the habitat . The preceding discussion suggests that the ability to interrupt behaviour should be advantageous: it allows an animal to minimize decision costs by making provisional commitments to temporally-extended actions , while maintaining its ability to respond to changes in the environment . Indeed , an advantage in terms of rate of rewards is observed ( Fig 1G ) . When there is no decision cost ( cd = 0 ) , the ratio of reward rates ( non-interruptible/interruptible ) is 1 , reflecting the fact that an animal would perform equally well whether or not it is equipped with an interrupt . However , as cd increases , this ratio decreases , reflecting a reward advantage when the option to interrupt is available . Beyond a critical value , the ratio reverts to 1 , with the animal deciding to escape immediately . This advantage in terms of reward rate arises from the reduction in the frequency of decisions when the animal is able to interrupt ( Fig 1H ) . This reduced decision rate makes it worthwhile for the animal to spend more time ( safely ) foraging in the environment ( Fig 1I ) , thereby increasing its haul of rewards . The advantage disappears when planning is so expensive that the animal escapes immediately . However , the cost at which this occurs is lower when interruption is not possible ( cf . Fig 1H and 1I ) .
States are determined by the values of three binary variables: 1 ) location ∈ {refuge , patch} , where refuge affords safety but not food , while patch affords foraging but also possible predation; 2 ) habitat quality ∈ {good ( G ) , bad ( B ) } , which determines the current utility of feeding ( see below ) ; and 3 ) predator ∈ {present ( P ) , absent ( A ) } , which describes whether there is currently a predator in the vicinity or not . We assume that the animal always knows its location , but that the state of predation and habitat quality are hidden variables whose values need to be inferred based on evidence . As before , choice involves selection of both an action a and a duration τ for its performance , and we refer to an action-duration pair ( a , τ ) as an activity . Possible actions are location-specific . In the refuge , the set is A refuge = { rest , assess , transit } , where rest is a recuperative behaviour; assess specifically aims to increase certainty about whether a predator is currently in the environment from a position of relative safety ( e . g . , sniffing near the entrance of the refuge ) ; and transit simply means moving to the patch location to forage . In the patch location , the set of available actions is A patch = { feed , assess , freeze , escape } , where feed refers to the ingestion of food; assess , as before , is aimed at detecting whether a predator is present , though here from a position of possible danger; freeze aims at avoiding predation by decreasing the probability of detection; and escape also aims at evading predation but through returning the animal back to the refuge . The patch quality and predator presence are assumed to obey simple semi-Markov dynamics which are independent of the animal’s actions ( we capture the consequences of feeding and predation in the rewards; see Discussion ) . In particular , we assume that ( a ) the predator transitions from being absent ( A ) to present ( P ) with transition rate γAP , and from present to absent with transition rate γPA; and similarly , ( b ) the habitat quality transitions between being good ( G ) and bad ( B ) according to transition rates ( γGB , γBG ) . In terms of predation risk , a crucial factor is the probability that the animal is detected and subsequently caught if a predator is present , given that the animal is currently engaged in a particular action . One of the more substantial simplifications we make is to assume that being detected inevitably leads to getting caught , but incurs only a fixed , finite , negative reward , rather than having a more extreme sanction . We formulate predation risk directly in terms of a rate of detection , assuming that this is directly mirrored in the predation rate . For rest and assess in the refuge location , we assume that this rate is 0; in the patch location , we assume that the detection rate is a function of the current action , written in abbreviated form as δa , where a denotes the current action . We assume that detection rate is lowest when the animal chooses freeze , and more generally assume the ordering δescape ≥ δtransit ≥ δfeed > δassess > δfreeze . Note that we assume that the risk of being caught is also lower if the animal chooses assess , since assess and freeze may reasonably be seen as lying on a continuum which trades off the degree of immobility with the amount of information garnered through risk assessment [30] . Habitats which are good ( G ) or bad ( B ) are defined in terms of the animal’s encounter rate with either rich or poor patches while feeding . A habitat which is good has a relatively high encounter rate ρ r G with rich patches and low encounter rate ρ p G with poor patches . Conversely , a habitat which is bad yields a low encounter rate ρ r B with rich patches , and a high encounter rate ρ p B with poor patches . Each action is associated with a reward rate r . These are assumed to be negative ( indicating an energetic cost ) except for feed ( net positive ) and rest ( zero ) . For net-negative reward actions , we assume the general ordering rescape < rtransit ≤ rassess ≤ rfreeze , so that escape is assumed to be most costly , while we are generally agnostic about the relative energetic costs of the other actions . For the feed action , the reward rate depends on whether the currently-encountered patch is rich ( rate r feed r ) or poor ( rate r feed p ) . A large negative reward rpred ≪ 0 is associated with being detected/caught by a predator ( note that this is not a rate ) , which may be considered a cost of injury ( we return to the issue of modelling predation costs in the Discussion ) . As in the simple example above , we assume that making a decision incurs a constant decision cost , cd ≥ 0 , which summarizes the computational , and presumably metabolic , costs associated with deliberation about which activity to pursue . The animal’s location is assumed to be directly observed , while the values of the predator and habitat variables are assumed to be only partially-observable . The animal is therefore required to make inferences about the latter which will depend on both prior knowledge of environment dynamics and observations . We assume that certain observations are more probable when a predator is present rather than absent , provided the animal takes appropriate measures to detect them . Two distinct types of cue are assumed . Firstly , we assume that an indirect , or ‘passive’ , cue oi is available regardless of the activity in which the animal is engaged ( e . g . , hearing a rustle in the bushes ) . This is emitted at a rate λ o i + when a predator is present , and at a rate λ o i - when a predator is absent . We write ¬oi for a non-observation of oi when it could potentially have been observed . Secondly , we assume that a direct , or ‘active’ , cue od is additionally available , but only if the animal is engaged in the assess activity ( e . g . , detecting a visual pattern at a particular location in the foliage ) . This is emitted at a rate λ o d + when a predator is present , and at a rate λ o d - when a predator is absent . Again , ¬od represents the non-observation of od when the latter would have been possible . Therefore , the animal may enter into different ‘information states’ regarding the predator variable depending on its choice of action , with assess providing the most reliable evidence . We assume that neither type of cue is available when the animal is engaged in rest . An absence of information is different from information about absence , as in ¬oi or ¬od . Information about habitat quality is assumed to be only available when the animal opts to feed . The relevant cue here is the current reward rate r ∈ { r feed r , r feed p } experienced while feeding . This will to some degree be informative about ( by depending on ) habitat quality—rich patches are more commonly encountered in a good habitat—but will not completely disambiguate the quality of the current habitat , since both types of habitat contain rich and poor patches . The animal is assumed to discount future rewards according to an exponential function with rate α ∈ [0 , 1] ( i . e . , a unit reward received after a delay of τ seconds is treated as having present value e−ατ , so that a larger value of α leads to more rapid discounting ) . The effect on behaviour of varying the discount rate is not a primary focus of the current work , and it is set to α = 0 . 1 throughout . In addition to its current location , the animal’s belief about whether a predator is present ( β t D ( P ) ) and whether the habitat is good ( β t Q ( G ) ) are jointly a sufficient basis on which to choose its actions . These collectively form the animal’s ‘belief state’—allowing us to solve the induced belief state semi-Markov decision process . How the animal’s beliefs change over time depends on both prior knowledge of the environment’s dynamics and any pertinent observations made . Since the predator and habitat quality variables evolve independently , we can consider belief updates for these separately . In the case where there is no observation ( such as when the animal engages in rest within the refuge ) , belief updates only depend on the environment dynamics . The two components of the belief state change from time t to t + τ according to β t + τ D ( P ) = γ A P γ A P + γ P A + ( β t D ( P ) - γ A P γ A P + γ P A ) e - ( γ A P + γ P A ) τ , ( 1 ) β t + τ Q ( G ) = γ B G γ B G + γ G B + ( β t Q ( G ) - γ B G γ B G + γ G B ) e - ( γ B G + γ G B ) τ . ( 2 ) For convenience , we consider an approximation to this for τ = Δt ≪ 1 β t + Δ t D ( P ) ≈ β t D ( P ) e - γ P A Δ t + ( 1 - β t D ( P ) ) ( 1 - e - γ A P Δ t ) , ( 3 ) β t + Δ t Q ( G ) ≈ β t Q ( G ) e - γ G B Δ t + ( 1 - β t Q ( G ) ) ( 1 - e - γ B G Δ t ) . ( 4 ) When additional information is provided by observations , this needs to be combined with prior expectations according to Bayes rule . If the animal is engaged in an activity for which only indirect observations oi provide information about the state of predation , then from Bayes rule , the updated belief β ˜ t + Δ t D ( P ) having received an instance of oi in the interval ( t + Δt ) is β ˜ t + Δ t D ( P ) ∝ P ( o i | present ) β t + Δ t D ( P ) , ( 5 ) where P ( o i | present ) ≈ λ o i + Δ t , and β t + Δ t D ( P ) is given by Eq ( 3 ) . For an omission , ¬oi , the likelihood P ( ¬ o i | present ) ≈ λ o i - Δ t is used instead . If the animal has access to both indirect and direct observations ( i . e . , when engaging in assess ) , belief updates follow a similar pattern , e . g . , β ˜ t + Δ t D ( P ) ∝ P ( o d | present ) P ( o i | present ) β t + Δ t D ( P ) , ( 6 ) where P ( o d | present ) ≈ λ o d + Δ t , and so forth for the possible combinations of values for oi and od . Relevant information about the quality of the habitat is only available when the animal selects feed , and encounters either rich or poor patches . Thus , the updated belief β ˜ t + Δ t Q ( G ) when encountering a rich patch in the interval ( t + Δt ) is β ˜ t + Δ t Q ( G ) ∝ P ( rich | good ) β t + Δ t Q ( G ) , ( 7 ) where P ( rich | good ) ≈ ρ r G Δ t , and β t + Δ t Q ( G ) is given by Eq ( 4 ) . If the encounter is with a poor patch , the likelihood P ( poor | good ) ≈ ρ p G Δ t is used instead . Fig 2B and 2C illustrate how beliefs βQ ( G ) and βD ( P ) evolve over time for the various different cases . When available , interruption is defined as the ability to stop an activity prematurely and make a new decision . That is , given an initial commitment at time t to an activity ( a , τ ) , interruption is the capacity to stop that activity at any intermediate time in the interval ( t , t + τ ) , and make a new decision ( at a cost of cd ) . Interruption can therefore be thought of as an additional , ‘internal’ action; how decisions about this action should be made , as well as issues surrounding its potential cost and mechanisms , are the principal concerns of the following sections .
To examine the basic behaviour of the model , we start by setting the decision cost to zero , cd = 0 . Fig 3A displays the optimal actions a* as a function of the animal’s location , refuge ( left ) or patch ( right ) , and belief state {βD ( P ) , βQ ( G ) } . In the refuge , the animal opts to transit to the patch only if a predator is unlikely , chooses to rest if a predator is likely , and selects assess when more uncertain . Choice is modulated by the probability that the habitat quality is good: the animal is slightly more likely to tolerate a higher probability that a predator is present in order to transit , while if the habitat quality is probably bad , the animal is increasingly likely to rest , in fact even when the probability of a predator is lower than 0 . 5 . Note that even though the utility of rest was assumed to be 0 , there are conditions under which the animal chooses it anyway . This occurs when the cost associated with assess is deemed too high to be worth paying ( i . e . , when it is highly probable that a predator is present , or when the habitat quality is likely to be bad ) . The advantage of assess is that it more quickly moves the animal to a state of increased certainty about whether a predator is present or not: if the predator is likely present , it is better to conserve energy by selecting rest; if the predator is likely absent , then the sooner the animal chooses to transit , the better . When in the patch location , the animal selects feed when βD ( P ) is low , assess when there is greater uncertainty , and chooses a defensive action ∈ {freeze , escape} when a predator is more likely than not to be present . Again , these tendencies are slightly modulated by the belief βQ ( G ) . The decision between freeze and escape is controlled by a number of factors in the model . Firstly , there is a difference in the cost of performing these actions , as we assumed that escape is more costly than freeze . Secondly , there is a difference in detectability while performing these actions—it is assumed that the detection rate is higher for escape than for freeze . Finally , there is the fact—which is the reason why escape is selected at all—that a successful escape will get the animal back to safety , while freeze leaves the animal in the patch location . Unsurprisingly , if escape is rendered ineffectual , in the sense that its performance also leaves the animal in the patch , then freeze is always preferred ( Fig 3B ) , consistent with changes in defensive pattern observed in rats and mice when flight is not possible [30 , 31] . As in the simple example considered above , when there is no decision cost it is always optimal for the animal to choose τ to be as short as possible ( see ‘Supporting information’ , S1 Fig ) . This is because there is no cost to doing so , while , in the absence of an interrupt , there is a potential cost of committing to longer durations ( viz . , not being able to change course of action if observations indicate a change in the environment ) . τ becomes relevant when we consider a nonzero decision cost below . In Fig 3C–3E , we summarize aspects of behaviour when the optimal policy is repeatedly exposed to an environment where a predator is either absent ( white bars ) or present ( black bars ) . Unsurprisingly , when a predator is present , the animal spends most of its time in the refuge ( Fig 3C ) engaged in either rest or assess activity ( Fig 3D ) , whereas it spends most of its time feeding in the patch when there is no predator . The adaptiveness of these behaviours is evident , and qualitatively similar reconfigurations of activity patterns in response to predator presence/absence are observed in laboratory-based ethological studies ( e . g . , [30] ) . Fig 3E makes the further point that even if the shortest duration τ is always selected , this doesn’t mean that ‘bouts’ of behaviour , defined as continuous periods of performing a single action , will always be of minimal duration . Instead , an external observer would sometimes measure longer behavioural bouts , particularly in the case of rest and feed activities . As the decision cost rises from zero ( cd > 0 ) , optimal behaviour changes . Fig 4A shows the optimal policy for decision cost cd = 0 . 01 , now displaying both optimal actions a* ( left panels ) in each location and corresponding optimal durations τ* ( right panels ) . Optimal actions a* are essentially identical to the cd = 0 case above ( cf . Fig 3A ) , and optimal durations τ* are generally selected to be as short as possible . The exception engendered by this rather minimal decision cost comes in the case of rest , where longer durations are observed , particularly when the habitat is likely to be bad . These reflect the dynamics of the environment: if habitat quality is currently likely to be bad , and conditions change relatively slowly , then conditions are likely to remain bad in the near future . Thus , rather than making another ( costly ) decision to rest at that future time , the animal can safely commit to rest for a longer period . By contrast , all other actions are associated with short durations . As in the previous case , this is sensible considering the consequences of doing otherwise . For example , when in the refuge , an animal that commits to assess for an extended duration may thereby forego time that could be better spent foraging or resting; in the patch , the same commitment risks foregoing the opportunity to feed or respond defensively ( freeze/escape ) . If decisions are made even more expensive , e . g . cd = 0 . 2 , further changes in policy are observed ( Fig 4B ) . In the refuge , rest becomes the most prominent action , with a duration that is uniformly chosen to be the longest possible . This minimizes decision costs when the environment is determined to be unfavourable . Note that under the assumed environment dynamics , beliefs βD ( P ) and βQ ( G ) will both move towards 0 . 5 once rest is initiated ( recall that no observation is available during this activity ) , and yet the belief state ( 0 . 5 , 0 . 5 ) yields the selection of rest under this policy . In other words , once this policy initiates rest , it will never do anything else , since the potential benefits of doing anything else are outweighed by the costs ( i . e . , any other option must have an average reward < 0 , since rest has net reward 0 ) . Of course , in reality , various factors would militate against this , including the stochasticity of action choice and progressive starvation ( see Discussion ) . In the patch , it is notable that freeze has all but disappeared from the behavioural repertoire , replaced by escape . This is because its benefits—lower detectability and the avoidance of unnecessary excursions back to the refuge— are now outweighed by the burden of greater future decision costs incurred in the patch . Some subtle increases in τ* are discernible for feed at low levels of βD ( P ) ; but overall , variation in τ* remains limited . What happens if we now allow the animal to interrupt activities prior to their completion ? The animal’s optimal policy only recommends the activity ( a* , τ* ) in initial belief state {βD ( P ) , βQ ( G ) } which is best in expectation over possible future belief trajectories . It is therefore perfectly possible that , having chosen optimally with respect to the initial belief state , experience sends the animal on a particular trajectory where it reaches a belief state in which an alternative activity would be preferable . Described at the level of a meta-decision , interruption should occur exactly when the benefit of interrupting an activity is greater than that of continuing it ( cf . [15] ) . Fig 4C shows the optimal interruptible policy for the same cd = 0 . 01 case as before . The clear difference is that for actions assess , feed , and freeze , it is optimal to set the duration to be as long as possible: τ* = τmax . Equipped with the ( free ) option of interrupting itself at any time , the animal can minimize its decision costs by provisionally committing to long activity durations , only interrupting and making another decision when it really needs to . The exceptions are transit and escape . For transit ( i . e . , moving from refuge to patch ) , choosing a longer duration never makes any sense—it would only increase the associated energetic cost and predation risk—and there is no advantage to interrupting this activity in the model . The same reasoning applies to escape . Since beliefs evolve in a predictable manner for rest , there is never a reason to interrupt this activity—the duration is calibrated in the same manner as in the non-interruptible case . Fig 4D displays the optimal interruptible policy for the more expensive , cd = 0 . 2 , case . Rest begins to occupy greater regions of belief space when in the refuge . This is similar to the trend in the non-interruptible case ( cf . Fig 4B ) , albeit to a lesser extent . The interruptible policy does not get ‘stuck’ permanently selecting rest , but will rather select assess when uncertainty is greatest ( i . e . , at ( 0 . 5 , 0 . 5 ) ) , and at many other points in this region . Note also that in contrast to the uniform choice of the longest duration for rest in the non-interruptible case , there is still some gradation in choice of τ in the interruptible policy ( Fig 4D , upper right panel ) : when it is strongly believed that habitat quality is currently good , it is better to choose shorter durations of rest to be able to take advantage of predator-free foraging conditions in the near future ( cf . Fig 4A and 4C ) . Equipped with the capacity to interrupt itself , an animal should perform at least as well as when lacking this capacity ( cf . [15] , Theorem 2 ) . At worst , decisions could be taken at maximum frequency , and the same decision costs incurred as in the non-interruptible case . We expect the interruptible case to do better than this , however: interruption should allow the animal to commit to extended activity flexibly , and so decrease the cost of unnecessary decisions—just as in the simplified example we first considered . As in that example , we compared performance in a simulated experiment . Here , behaviour is measured over a 6-minute period in which a predator is initially absent ( 2 min ) , then present ( 2 min ) , then absent again ( 2 min ) ; habitat quality is allowed to fluctuate randomly . We ran this experiment 1000 times for different settings of the decision cost cd , ensuring that conditions were exactly matched between policies . We measured the resulting rewards averaged over both episodes and experiments . Fig 5A ( inverted triangles; ‘exact’ ) plots the reward rate achieved by the non-interruptible policy as a fraction of that achieved by the interruptible policy ( cf . Fig 1G ) . As seen before , when cd = 0 , this fraction is 1 , reflecting the fact that the optimal strategy here is to make decisions as often as possible , since there is no penalty to doing so . However , as decisions become increasingly expensive ( i . e . , cd becomes larger ) , the fractional utility rate decreases , reflecting the fact that the interruptible policy is achieving higher rewards . Beyond a certain level of decision cost ( here , cd > 0 . 4 ) , the fraction reverts to 1 . Again , the reward advantage comes from the reduction in the frequency of decisions that is possible when the animal has the capacity to interrupt ( Fig 5B ) . Reversion of the fractional reward rate to 1 in this case corresponds to the point at which simply spending all of its time engaged in rest is the optimal course of action for the animal , regardless of the capacity to interrupt . If we explicitly compare the average time during the experiment spent at rest for non-interruptible and interruptible policies , both are eventually driven to exclusive choice of this action as decisions become more expensive . However , this occurs for lower values of cd when interruption is unavailable ( Fig 5C ) . Fig 5D compares the behaviour and beliefs of the non-interruptible ( upper ) and interruptible ( lower ) policies for a particular run of the experiment at an intermediate decision cost , cd = 0 . 1 . This clearly illustrates the fact that the non-interruptible policy makes decisions at a much higher frequency ( vertical dashed lines ) , but also highlights differences in choices . Most notably , in both cases the animal returns to the patch after the predator has been removed from the environment , but when faced with observations that may indicate a potential threat , their behaviour differs . In the non-interruptible case , the animal opts to escape to the refuge and engage in rest , since the costs of foraging are deemed too high when a predator is thought likely and habitat quality is low ( Fig 5D , upper ) . By contrast , in the interruptable case , the animal opts to assess whether there is a real threat , and continues to feed when it transpires that no predator is around—it is sufficiently flexible in its behaviour for it to be worthwhile to continue in the patch and intermix periods of both feeding and assessment ( Fig 5D , lower ) . Note also the shorter bouts of rest in the interruptible case , leading to a greater frequency of assess and so , at least in this case , a marginally earlier return to the patch . Interruption is evidently useful; however , we have not considered the costs of the calculations associated with this operation . Since interruption can be thought of as another layer of decision-making , we might just have increased the computational burden . One answer is to conceive of a cheap and ‘light-weight’ interruption process . The points at which interruption should be triggered effectively form a boundary on belief space . We might therefore consider that at the outset of an activity , one could ‘construct’ such a boundary , or threshold , and then have interruption occur whenever this threshold is breached . Belief-monitoring would still be required to recognize if and when the animal enters the termination set of beliefs . This still implies a decision—whether or not this threshold has been reached—but of a particularly simple kind . Fig 6A–6C ( symbols ) plot the optimal interruption thresholds θ , i . e . , levels of belief βD ( P ) at which activities should be interrupted , as a function of decision cost cd and three fixed levels of belief βQ ( G ) . The thresholds are plotted for feed and for assess ∈ {patch , refuge} , which are activities that show a robust increase in duration with nonzero decision costs ( cf . Fig 4C and 4D above ) ; since freeze tends to be disfavoured at higher decision costs ( cf . Fig 4D ) , data points are much more sparse , and so we do not consider it here . Note that feed only has an upper boundary on βD ( P ) , while assess has both upper and lower boundaries . Also , for some combinations of cd and βQ ( G ) , the set of beliefs {βD ( P ) } for which assess in the refuge location is optimal ( i . e . , the ‘initiation set’ ) is empty , and so interruption thresholds are omitted in these cases ( e . g . , for βQ ( G ) = 0 . 2 , cd > 0 . 2; Fig 6C ) . As one might expect from the example optimal policies seen previously ( cf . Fig 4C and 4D ) , the interruption threshold for feed increases for higher βQ ( G ) , i . e . , as the payoff for feeding increases ( Fig 6A ) . With assess in the patch location , the upper and lower thresholds move upwards as βQ ( G ) increases , respectively indicating a greater willingness to interrupt and initiate feeding ( lower boundary ) , and a greater reluctance to trigger defensive behaviour ( upper boundary ) ( Fig 6B ) . For assess in the refuge , the upper boundary moves upwards as βQ ( G ) increases , indicating greater reluctance to interrupt and initiate rest when habitat quality is likely to be good; the lower boundary shows less variation—the animal requires βD ( P ) to be very low to interrupt and initiate transit , regardless of βQ ( G ) ( Fig 6C ) . The thresholds θ for the first two cases are well approximated by linear functions of βQ ( G ) and cd , while this is less true of assess in the refuge location ( Fig 6A–6C , dashed lines ) . We can nevertheless ask how well , in comparison to the optimal policy , an animal will do when selecting interruption thresholds according to the linear function which most closely approximates the optimal thresholds . Fig 5A ( red asterisks ) shows that this simple , approximate way of setting thresholds leads to benefits that are extremely close to that of the exact case . It is also informative to examine how interruption thresholds change as a function of cue reliability . Fig 6D–6F show these as a function of decision cost cd and true positive rates for either indirect cues , λ o i + ( Fig 6D ) , or direct cues , λ o d + ( Fig 6E and 6F ) , while βQ ( G ) is kept constant . Reducing the true positive rate in either case means less reliable information about the predator . For feed , thresholds decrease with less reliable indirect cues , consistent with greater caution—the animal is increasingly prepared to interrupt feeding in order to assess and gain better information ( Fig 6D ) . For assess in the patch location , upper thresholds similarly decrease with less reliable direct cues , reflecting a greater willingness to trigger interruption and switch to defensive behaviour ( Fig 6E , upper ) . That lower thresholds actually increase with less reliable cues indicates a greater willingness to interrupt and initiate feed in spite of having less reliable predation cues; this might seem the exact opposite of more cautious behaviour . The reason is that gathering information via assess decreases in marginal value for unreliable cues , making feed an increasingly attractive alternative . Note that the relative reliabilities of direct and indirect cues are important in this trade off , since feed still provides some information via the latter class of cues . Finally , thresholds for assess in the refuge location follow a qualitatively similar pattern ( Fig 6F ) . With less reliable cues , the animal is quicker to interrupt its behaviour and initiate rest , thereby conserving its energy; with more reliable cues , the animal is more conservative in making this transition ( Fig 6F , upper thresholds ) . The difference in thresholds is less pronounced in the downwards direction , so that a relatively high degree of certainty regarding predator absence is required before initiating transit in all cases .
We used the simple example of foraging under predation risk to explore the possible advantages of being able to interrupt on-going behaviours . We showed that this allows animals to make provisional commitments to courses of behaviour rather than having either to check obsessively or just not to exploit available resources at all . This had measurable benefits in terms of efficiency and effectiveness . We also showed that it is not necessary to solve a complex decision problem to choose whether to interrupt , but rather that a simple , cheaply-parameterized , approximate , threshold-based policy can perform almost as well as the optimal policy across a variety of parameter settings . In our model , decision cost was summarized by a simple scalar value . However it may in reality comprise separate components , including an intrinsic ( e . g . , metabolic ) cost of performing the computation , and an opportunity cost , which summarizes what could have been obtained by employing the engaged resource ( time , computation ) otherwise [32–34] . The cost we considered , cd , is a version of the former , and could arise from steps of expansion and calculation in a decision tree used for planning . The opportunity cost of time arises from discounting—the fact that taking time to think postpones future rewards , making them less valuable ( an issue more extensively explored in the case of long-run average reward; cf . [35] ) . The opportunity cost of the use of other cognitive resources for deliberation , such as working memory , are starting to be considered and quantified [36–39] . Model-free planning [40 , 41] is likely to impose far smaller cognitive costs than the sort of model-based planning that we have so far been considering . However , at the very least , there will still be costs associated with task switching [42] . The activities ( a , τ ) that formed the objects of choice in the current work may be recognised as a simple form of option [15]—a policy for taking actions over an extended period of time—which has formed one basis ( amongst many , cf . [43] ) for exploring the issue of temporal abstraction in reinforcement learning . The benefit of being able to interrupt an option before it would otherwise terminate was highlighted in the initial options paper by Sutton and colleagues [15] ( the authors also cite previous work by Kaelbling [44] ) , but issues surrounding decision cost , plausible mechanisms for interruption , and partial observability were not considered there . More recent work by Precup and colleagues [45] , who introduce an ‘option-critic architecture’ in the context of discovering and learning options , explicitly considers the idea that a form of switching cost could encourage commitment to option execution . We showed that changing the informativeness of cues about the predator had consequences such as increasing the propensity to feed and altering interruption thresholds . The effects of such changes on the observed temporal structure of behaviour are subtle , since the speed with which the belief about the presence of the predator changes will also change . In the case we simulated , observations remained sufficiently informative that the latter effect had little impact , but it would be interesting to examine more systematically how thresholds and speeds interact in determining when interruption occurs . Since this would require some significant adjustments to the current model , we leave exploration of this subtlety for future work . The trade off between energetic gain and predation risk is a central topic of behavioural ecology [27 , 46] and has been the subject of extensive previous theoretical work , though principally at more ‘molar’ levels of analysis than our approach here [28 , 29 , 47 , 48] . While our model of foraging under predation risk was loosely inspired by ethological and ethoexperimental studies of rodent behaviour in such settings [30 , 49–51] , a more realistic model would extend this in a number of ways . A first extension concerns the model of predation . Most notably , in our detailed model , getting caught by the predator was associated with a large negative cost ( as from a severe injury ) rather than an outright termination of the decision process ( as from extermination ) . If the cost of injury is sufficient , then the difference becomes rather moot; however , a more realistic model involving procreation and death from natural and unnatural causes would be most interesting . Second , we made the assumption that when a predator is present , the animal has a constant probability of being detected and harmed—indeed , we made no distinction between detection and being caught , and have not otherwise separated out the ‘subcomponents’ of predation risk [27] . Third , we assumed that the rate at which a predator enters and leaves the environment is constant , whereas one might expect that the predator would be more likely to remain in the environment if it has detected prey . Finally , since the model is non-spatial , it cannot address important factors such as differences in time to reach safety from different locations , and associated variation in the distance an animal will tolerate from a simulated predator before initiating flight ( ‘flight-initiation distance’ in the light of predatory imminence [51–53] ) . Further unmodelled complexity arises through the behavioural sophistication that animals display both in assessing predation risk and in responding to the presence of a predator . These behaviours have been extensively studied in wild and laboratory rats [49 , 51] , including investigations of approach-avoidance conflict [54] . Predatory risk assessment alternates between cautious forays into a potentially dangerous area and rapid retreat to safety , if available . If escape or concealment is not possible , the animal instead alternates between freezing and scanning with the head and vibrissae . When actually confronted with a predator , a rat will variously respond by fleeing , freezing , or attacking , depending on the nature of the current environment—particularly whether a place of relative safety , or refuge , is available—and the intensity of the perceived threat , or ‘defensive distance’ [55] . Capturing the latter concept would require a richer spatial and perhaps temporal model . There were also marked simplifications concerning the benefits of foraging , both in terms of the agent and the environment . In terms of the former , we did not capture the possibility of running out of resources . Thus , for instance , it would have been possible for the agent to stay in rest in perpetuity ( as indeed would seem optimal for expensive decision-making and no interruption; Fig 4B ) . In reality , as threats to homeostatic integrity loom , we can expect animals to exhibit more risk-seeking behaviour [56] . This would emerge in the current formulation given a more realistic characterization of the utilities [57] . In terms of the environment , a key simplification is to ignore resource depletion by the agent , and the existence of multiple patches . Then , critical concerns in foraging theory such as the marginal value theorem [58] would be important , and the rate of prey encounter or capture might also contribute to interruption . While we primarily focused on the computational and algorithmic aspects of interruption , it is of clear interest to relate the current work to neural substrates . The present theory of interruption can be seen as a mesoscopic bridge between the macroscopic view of the neuromodulator norepinephrine ( NE ) suggested by the reversal experiments of Devauges and Sara [20] , in which it reports changes to the whole rules of the environment [25] , and the microscopic view of Bouret and Sara [21] , and Dayan and Yu’s [26] interpretation of [22] , in which it reports the current level of uncertainty about the ongoing belief state in a single perceptual inference problem , triggering interruption on reaching a pre-defined threshold , allowing switching to a better hypothesis [59] . This was proposed as part of the approximate strategy of provisionally committing to a single hypothesis , but keeping track of how this might be erroneous . At all levels , the common theme is the question of whether to interrupt a default state , whether that be a default belief about the current state of the world , as in [25 , 26] , or a default program of activity , as in the current case . Closer examination of recordings from noradrenergic neurons and associated circuitry during naturalistic behaviour ( including foraging tasks ) for evidence of the sort of multilevel dynamics predicted by these three accounts would therefore be merited . NE activity has also be associated with arousal [60]; the relationship between this concept and that of interrupts is the subject of on-going theoretical work . Its association with other functions such as learning [61 , 62] and exploration [23 , 63 , 64] are arguably further removed . Possible divisions of labour and interactions between cortical and subcortical systems in this context are also of interest [65] . One can speculate about the relationship between the putative function of NE as an interrupt and its association with stress-induced anxiety [66 , 67] . In environments with a high proportion of unpredictable events , or indeed an environment that is either frankly aversive or believed to contain possible sources of threat , the interrupt mechanism is likely to be frequently engaged , whether received stimuli reflect real threats or not . This state of high interruptibility , or distractibility—reminiscent in some respects of ‘hypervigilance’—is plausibly associated with higher costs , both in terms of time and energy , and would be manifest in our own model in a greater frequency of costly deliberations . In the limit of an extremely inconstant environment , interruptibility loses any net benefit; however , whether this happens before the costs are such that the animal will refuse to engage with the environment at all depends on the details of the cost structure . All these forms of interrupt are likely to be distinct from the motor interrupt that plays a critical role in tasks such as the stop signal reaction time task [68] , or the ‘hold your horses’ interrupt [69 , 70] that has been suggested to suppress a prepotent action temporarily to allow time for a correct choice to be made . The former may also be associated with the form of cognitive state change associated with the phasic NE signal [21 , 26] , but this would be a distinct consequence of the same underlying detection . Indeed , the neural substrates for these others forms of inhibition are notably different , implicating regions of the superior colliculus and basal ganglia , respectively . Although we focused on the benefits of interruption and its possible mechanisms in the context of foraging , similar considerations are expected to be more generally applicable . One area of particular interest is decisions about ‘internal’ rather than external behaviour ( i . e . , meta-cognition ) . Indeed , the interpretation of NE we mentioned above as a signal for interrupting a likely incorrect ongoing belief is an example of this [26] . The strategy of provisional commitment could be particularly beneficial when there are such large numbers of potential hypotheses that they cannot simultaneously be entertained . As another example , consider the problem of trying to decide whether to perform a particular action by considering its possible future consequences ( i . e . , by model-based planning; [40] ) . Ultimately , this will be intractable due to the myriad possibilities , and the challenge of determining the uncertainties and utilities of each . However , some degree of planning should be useful , until the fog of uncertainties about the future and the complexity of calculating it overwhelm the utility of attempting to do so . This question has also been of great interest in artificial intelligence [1 , 71 , 72] . One could try to determine how deep to plan before planning—planning to plan—but doing so optimally presents an even more formidable computational challenge . A more realistic option would be to commit to the planning process provisionally , while monitoring its progress ( e . g . , the extent to which one’s uncertainty about the value of the action decreases with planning depth ) , and to interrupt this process when further planning appears unjustifiable ( see , e . g . , [73] ) . In such cases of diminishing marginal returns , interruption according to a relatively simple threshold rule may apply , similar to the logic of the marginal value theorem in foraging theory [32 , 58] . Consideration of how long to run an algorithm is a central concern of work on anytime algorithms ( or ‘interruptible’ algorithms ) , i . e . , algorithms which are always guaranteed to return a valid solution but where the solution quality typically improves with time [6 , 74 , 75] . It is pressing to consider these insights in the light of what we know about the neural substrates of interruption .
|
Animals should ideally be able to monitor all relevant aspects of their environment constantly and be ever prepared to alter their course of action in the face of unexpected change . However , the impractically high costs of continual monitoring and deliberation mean that a more realistic strategy is required . Here , we explore a solution in which an animal makes provisional commitments to a temporally-extended action while maintaining the ability to interrupt this behaviour prematurely on the basis of more limited sensing . We demonstrate the benefits of such a scheme through the example of foraging under predation risk , and propose a simple mechanism for implementing interruption . We suggest possible relationships between these results and neural substrates , particularly norepinephrine , and also highlight potential relevance to diseases of distractibility .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"ecology",
"and",
"environmental",
"sciences",
"neurochemistry",
"chemical",
"compounds",
"decision",
"making",
"predator-prey",
"dynamics",
"population",
"dynamics",
"social",
"sciences",
"neuroscience",
"organic",
"compounds",
"hormones",
"cognitive",
"psychology",
"animal",
"behavior",
"cognition",
"amines",
"neurotransmitters",
"population",
"biology",
"catecholamines",
"animal",
"management",
"physical",
"laws",
"and",
"principles",
"zoology",
"animal",
"performance",
"norepinephrine",
"conservation",
"of",
"energy",
"foraging",
"behavior",
"chemistry",
"agriculture",
"community",
"ecology",
"physics",
"biochemistry",
"psychology",
"organic",
"chemistry",
"ecology",
"predation",
"trophic",
"interactions",
"biogenic",
"amines",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"cognitive",
"science"
] |
2018
|
Interrupting behaviour: Minimizing decision costs via temporal commitment and low-level interrupts
|
The role of stochasticity on gene expression is widely discussed . Both potential advantages and disadvantages have been revealed . In some systems , noise in gene expression has been quantified , in among others the lac operon of Escherichia coli . Whether stochastic gene expression in this system is detrimental or beneficial for the cells is , however , still unclear . We are interested in the effects of stochasticity from an evolutionary point of view . We study this question in the lac operon , taking a computational approach: using a detailed , quantitative , spatial model , we evolve through a mutation–selection process the shape of the promoter function and therewith the effective amount of stochasticity . We find that noise values for lactose , the natural inducer , are much lower than for artificial , nonmetabolizable inducers , because these artificial inducers experience a stronger positive feedback . In the evolved promoter functions , noise due to stochasticity in gene expression , when induced by lactose , only plays a very minor role in short-term physiological adaptation , because other sources of population heterogeneity dominate . Finally , promoter functions evolved in the stochastic model evolve to higher repressed transcription rates than those evolved in a deterministic version of the model . This causes these promoter functions to experience less stochasticity in gene expression . We show that a high repression rate and hence high stochasticity increases the delay in lactose uptake in a variable environment . We conclude that the lac operon evolved such that the impact of stochastic gene expression is minor in its natural environment , but happens to respond with much stronger stochasticity when confronted with artificial inducers . In this particular system , we have shown that stochasticity is detrimental . Moreover , we demonstrate that in silico evolution in a quantitative model , by mutating the parameters of interest , is a promising way to unravel the functional properties of biological systems .
Noise in gene expression , i . e . , the variation in gene expression in an isogenic population in a homogeneous environment , has drawn much attention in recent years . When two isogenic cells vary in gene expression , this can be due to variation in factors determining gene expression in these cells , such as transcription factors , the concentration of RNA polymerase , the cell cycle , etc . , which is called extrinsic noise . When , however , all extrinsic noise is absent , gene expression between these cells would still be different , because gene expression is inherently stochastic , due to the low numbers of molecules involved . The latter is called intrinsic noise . Indeed , it has been clearly shown that gene expression can be stochastic [1–3] . The implications of stochastic gene expression are , however , much less clear . Stochasticity has been proposed to be deleterious in some systems [4] , while being advantageous [5 , 6] in others . However , there is very little known about the consequences of stochasticity on particular systems . Maybe the best-known system for genetic regulation is the lac operon of E . coli . The lac operon codes for three genes , two of which have a function in lactose uptake and metabolism . It codes for a permease protein that transports lactose into the cell and β-galactosidase , which degrades lactose . Gene expression in the lac operon has been experimentally shown [2 , 7] to be stochastic . The lac operon is regulated via a positive feedback loop . The operon is induced by allolactose ( which is formed by β-galactosidase , from lactose ) . Induction of the operon again leads to higher permease and β-galactosidase concentrations and hence to higher allolactose concentrations . This positive feedback loop can cause bistability , which means that for certain extracellular inducer concentrations , two stable equilibria exist for the operon , induced and repressed . In a bistable system , stochastic gene expression can cause switching between these equilibria and hence give rise to a bimodal population . Such a bimodal population can be advantageous for the population [5] , a phenomenon called bet-hedging . Bistability for the lac operon has been demonstrated using thiomethyl β-D-galactoside ( TMG ) [8] . Recently , these experiments were repeated and bistability for TMG was confirmed [9] . In this paper , bistability for isopropyl β-D-thiogalactopyranoside ( IPTG ) and the natural inducer lactose was also tested . Although it is known that IPTG also enters the cell independently from the operon , it still behaved bistably . For lactose , no bistability was found . In [10] we showed that this difference in behavior is because artificial inducers are not degraded by β-galactosidase , while lactose is . Therefore , the positive feedback loop is much stronger for these artificial inducers . Furthermore , we showed that , using a deterministic evolutionary model of the lac operon in a fluctuating environment , cells adapt their promoter function such that the response with respect to lactose becomes continuous instead of bistable [10] . This can be explained by the increase in delay in lactose uptake that bistability unavoidably causes . These in silico evolved promoter functions , however , still behaved bistably with respect to artificial inducers . Indeed , the in silico evolved promoter function resembled the experimentally measured promoter function . Here we study how noise in gene expression influences the adaptation of the lac operon promoter function , both on a physiological and an evolutionary time scale . We use a computational approach to tackle this problem . We modified the previously developed deterministic model [10] for the evolution of the lac operon to include stochasticity in gene expression . This model is spatially explicit and consists of cells that grow on glucose and lactose , divide , and die . The intracellular model consists of detailed differential equations describing lac operon transcription , translation , and metabolism , with parameters taken from literature . The cells evolve the parameters which determine the lac operon promoter function and in this way adapt to the ( fluctuating ) environment . Importantly , the cells can in this way also adapt to the constraints that are imposed by the fixation of the other parameters . In our view , this is a good way to cope with the inevitable parameter uncertainty . See Materials and Methods for a more detailed description of the model . We added stochasticity in gene expression on the protein level . We assumed that protein production occurs in bursts , as is experimentally observed [7] . The amount of protein produced per burst ( i . e . , the burst size ) was shown to be geometrically distributed with a mean of five proteins [7] . This observation suffices to make our deterministic model stochastic , as is explained in the Materials and Methods section . Protein degradation is modeled binomially . When a cell divides , the number of proteins is divided between the two cells in a binomial way . In this way we added stochasticity without introducing any unknown parameter . By comparing the deterministic and stochastic models , we can directly observe the consequences of stochasticity on the lac operon , which is experimentally difficult , because the lac operon is inherently stochastic . We compared the amount of noise in gene expression in our model with experimentally observed values of noise for the lac operon [2] and found that noise in gene expression in our model is comparable to the experimentally observed noise . These noise values were measured using IPTG . IPTG is not degraded by β-galactosidase and therefore behaves very differently than the natural inducer , lactose . The positive feedback loop is much weaker for lactose than for artificial inducers such as IPTG . Accordingly , we find that noise values for lactose are much lower than for IPTG . Therefore , the effect of stochasticity on evolution of the lac operon might be lower than what would be expected from these experiments . In experiments where stochasticity in gene expression is measured , isogenic populations in well-mixed systems are considered in order to exclude all other sources of population heterogeneity . When we want to investigate the importance of stochasticity in gene expression in natural circumstances , we should , however , also take these other sources of population heterogeneity into account . Therefore , by using a spatially explicit model of cells that evolve their lac operon promoter function , spatial and genetic heterogeneity are automatically taken into account . We find that both genetic and spatial heterogeneity contribute more to population heterogeneity than stochasticity in gene expression . To explore the effect of stochasticity on evolution of the lac operon , we compared the results of the evolutionary simulations between the stochastic and the deterministic models [10] . We found that in the stochastic simulations , cells evolve a higher repressed transcription rate than in the deterministic model . Therefore , the promoter functions that evolved in the deterministic model experience more stochasticity when placed in the stochastic model than the promoter functions that evolved in the stochastic model . We show that this causes a reduction of fitness compared with the promoter functions evolved in the stochastic model , due to an increase in delay in lactose uptake . We conclude that in the stochastic model , the promoter functions evolve to minimize stochasticity in gene expression . Indeed , stochasticity , when growing on lactose , is relatively unimportant for the dynamics of these evolved promoter functions , except sometimes at high glucose , low lactose concentrations . The dynamics when growing on lactose can well be described using a deterministic model . When modeling the dynamics of induction with artificial inducers , stochasticity , however , is much more important , due to the stronger positive feedback , and should be incorporated .
First we studied how stochasticity in gene expression influences noise in gene expression . The promoter function of the cell in large part determines the amount of stochasticity a cell experiences , because stochasticity is higher at low protein levels . Therefore , promoter functions with low repressed transcription rates experience more noise than promoter functions with high repressed transcription rates . It has been shown that noise in gene expression can be split up into two orthogonal components , intrinsic and extrinsic noise , such that [11] . Here the total noise ηtot is defined as the standard deviation divided by the mean of the population . Extrinsic noise is all noise that would affect two identical , independent copies of one gene in a single cell in exactly the same way , and intrinsic noise is noise that causes differences in expression levels of identical copies in a single cell . The amount of intrinsic and extrinsic noise of lac-repressible promoters in different E . coli strains has been measured [2] . This was done by placing two genes , coding for two fluorescent proteins , which are controlled by identical promoters , in an E . coli strain . The intrinsic and extrinsic noise can now be simultaneously measured , by measuring how the protein levels fluctuate . We performed similar simulations to validate the stochastic model with these experiments . We initiated a population of 100 induced or repressed cells ( solid and dotted lines , respectively , in Figure 1 ) in a homogeneous environment with a certain extracellular inducer concentration . As in the experiments , spatial and genetic heterogeneity are absent . We waited for 41 hours , more than enough time for cells to go to equilibrium in the deterministic model . Then we calculated the noise in the protein number in the population . Noise levels were obtained in the same way as in the experiments . We kept track of the activity of two independent , identical lac-repressible genes that do not have a function in lactose uptake . Because we did not introduce any free parameter in the model when introducing stochasticity , noise levels only depend on the promoter function used . We used a promoter function that has the same repression rate as the one used in the experiments ( ≈110 , see Figure 1A ) . As a comparison , for the wild-type lac promoter function , the repression rate has been reported to be 170 [9] . Furthermore , we used a Hill-coefficient of 4 . 0 [12] . For the growth rate , we took 0 . 01/min in these simulations . The induced transcription rate of the operon in these simulations is equal to the maximal transcription rate we imposed during the evolutionary experiments . The experiments were done using IPTG , an artificial inducer of the lac operon . In contrast to lactose , IPTG is not degraded by β-galactosidase . This changes the dynamics , hence the noise , of the system considerably . Our model describes operon dynamics with lactose as inducer . Using parameters for IPTG , as used in [13] , and not allowing for degradation of IPTG by β-galactosidase , our model can also describe lac operon dynamics when induced by IPTG . It is known that IPTG also enters the cell in the absence of permease . This is not taken into account in our model describing lactose dynamics and hence we add a permease-independent influx term for IPTG ( see Materials and Methods ) . The amount of permease independent influx we chose is such that the maximum in the total noise is similar to that in the experiments . In Figure 1A the results for IPTG are shown , using a permease-independent influx , kIPTG , of 1 . 35/min . The intrinsic noise found in our simulations is almost identical to the experimentally observed intrinsic noise . Intrinsic noise levels in the model only depend on the number of proteins . We found that the data can almost perfectly be fitted by the function ηint = C*Pa , with a = −0 . 505 , and C = 2 . 35 with an R2-value of 0 . 99996 . At steady state , the theoretical prediction [14] is a = −0 . 5 and , where b equals the average burst size and ζ is the ratio of mRNA to protein lifetimes . This figure also confirms that the induced transcription rate we used has the right order of magnitude . It has been reported that the maximal in vitro transcription rate is approximately 0 . 18/min , but that the maximal in vivo transcription rate is approximately 1–10/min [15] . In contrast to [16] , who used the maximal in vitro transcription rate , we used a maximal transcription rate of 11/min . Qualitatively , the extrinsic noise corresponds with the experimentally observed noise . The maximum in the extrinsic noise is caused by the positive feedback loop in the lac operon . Fluctuations due to the intrinsic noise are amplified by the positive feedback , but only at intermediate inducer ( hence protein ) concentrations , where the promoter function is steepest . Note that the cell cycle and the intracellular inducer concentration are the only extrinsic noise sources we included in the model . The maximum in the extrinsic noise is located at lower protein concentrations in the data than in our model . This is because the extrinsic noise is high if the positive feedback is strong . Sufficient positive feedback can only be accomplished if the protein-dependent and protein-independent inducer influx are of the same order of magnitude . Therefore , when the protein-independent inducer influx is high , the maximum in the extrinsic noise will be located at high protein concentrations . We can reproduce the experimental data better if we assume a lower growth rate or a higher protein-dependent inducer efflux . Both these changes diminish the positive feedback and therefore we need a smaller protein-independent influx to fit the maximum amount of extrinsic noise . Therefore , the maximum in extrinsic noise will also be shifted to lower protein concentrations . When we use , for example , a protein-dependent inducer efflux of 300 mM/ ( mM permease min ) instead of 49 . 35 , the value reported by [13] , we indeed find that the maximum in the extrinsic noise shifts to lower protein numbers ( Figure 1B . ) We obtained these curves for a protein-independent inducer influx ( kIPTG ) of 0 . 1/min . For lactose as inducer , the picture is very different ( Figure 1C ) . The intrinsic noise remained unchanged , but extrinsic noise changed considerably . There is still a maximum in extrinsic noise , but this is barely visible . The reason is that the positive feedback loop is much weaker for lactose than for IPTG , due to the degradation of lactose by β-galactosidase . Indeed , the operon used is monostable for lactose . For lactose , we found that extrinsic noise is almost completely determined by the cell cycle for all protein numbers , instead of only for high protein numbers as we found for IPTG . We conclude that the results of the experiments can only be understood when we realize that IPTG was used as an inducer . When lactose as inducer is used , extrinsic noise levels as high as in the experiments can in our model only be observed when the repressed transcription rate is considerably lower . Finally , we tried to simulate noise in gene expression for TMG , another artificial inducer . As IPTG , TMG is not degraded by β-galactosidase , but in contrast to IPTG , there is no permease-independent influx . Therefore , the positive feedback loop is stronger for TMG . We simulate TMG by using the same parameters as for IPTG , except the permease influx rate is changed from 1 . 35/min to 0/min . We observe that the extrinsic noise increases drastically ( Figure 1D ) , while the intrinsic noise again remains unchanged . Indeed , the stronger positive feedback loop increases extrinsic noise . Experimentally , it has been observed that the wild-type lac operon behaves bistably for both IPTG and TMG [9] . For IPTG , however , bistability is expected to be much less severe [9] . We find that the experimentally observed noise can best be described by a promoter function that is just bistable for IPTG . This can be seen when we compare the amount of noise for initially induced and repressed cells . Although all individual cells are in equilibrium , the results are different when we start with an initially induced population or an initially repressed population . For TMG , we also find that the promoter function is much more bistable . To observe the effect of bistability better , we also show how noise levels depend on the extracellular inducer concentration , for both IPTG and TMG ( Figure 1E and 1F ) . The hysteresis-loop , indicating that for certain extracellular inducer concentrations the amount of proteins is dependent on the history of the cells , is clearly visible . Indeed , for TMG this loop is much larger than for IPTG . Furthermore , we see that when inducer concentrations are low , transitions from the induced to the repressed equilibrium are more likely , and vice versa . In the previous section , we showed that stochasticity in gene expression can cause significant amounts of noise when the operon is repressed . The amount of noise at intermediate inducer concentrations was , however , much larger for IPTG and TMG than for lactose . These experiments and simulations were done using cells with identical promoters in a well-mixed environment . In natural and evolving populations , different cells will have slightly different promoter functions , due to genetic variation , and the environment will not be well-mixed [17 , 18] . Both these factors can cause population heterogeneity , but are neglected in almost all experiments . This makes sense if we want to study whether stochasticity in gene expression occurs at all . Stochasticity in gene expression can only be proven when gene expression is different in two identical cells in an identical environment . For the importance of stochasticity in natural circumstances , however , other factors causing population heterogeneity need to be considered . We embedded the intracellular stochastic model of lac operon dynamics in an evolutionary , spatial context , as described in the Materials and Methods section and in more detail in [10] . We used this stochastic model and the deterministic model [10] to unravel the contributions of the various factors on population heterogeneity . To this end , we performed evolutionary simulations in both a spatial and a well-mixed environment , for both the stochastic model and the deterministic model . In these simulations , cells evolve their promoter function in an environment in which the glucose and lactose concentrations fluctuate . To understand the effect of genetic diversity , we also performed simulations using a genetically identical population . We used the last common ancestor ( the last individual cell that had the whole population at the end of a simulation as offspring ) of each evolutionary simulation for this . In this way , we can compare the effects when genetic variation , spatial variation , both , or none are incorporated , in both the deterministic model ( Figure 2A ) and the stochastic model ( Figure 2B ) , yielding eight different simulations . Of the four sources of population heterogeneity in our model , spatial , genetic , stochastic , and cell cycle related , we can exclude all , except for the cell cycle . When all these three noise sources were excluded , we still observed some population heterogeneity , but it was completely independent of protein number ( Figure 2A , blue dots ) . The population heterogeneity varied wildly over time , due to the partial synchronization of the cells . Sometimes all cells have just divided , and population heterogeneity is very low . When only half of the population has recently divided , population heterogeneity is maximal . This explains the extremely broad distribution of blue dots in Figure 2A . When genetic or spatial heterogeneity was present , very low values of population heterogeneity did not occur anymore ( Figure 2A , red , green , and black dots ) . When these sources of population heterogeneity were present , population heterogeneity became inversely correlated with protein number . It is clear that in our model , space has a larger effect on population heterogeneity than genetic variation ( see Figure 2A , red and green dots ) . When we compare population heterogeneity in the deterministic model ( Figure 2A ) with the population heterogeneity in the stochastic model ( Figure 2B ) , we find , as expected , the largest difference when spatial and genetic heterogeneity are both absent . Intrinsic noise ( Figure 2B , yellow line ) gives a lower boundary to the population heterogeneity in the stochastic model . The mean population heterogeneity is increased considerably by stochastic gene expression . When genetic or spatial heterogeneity was present in the stochastic model ( Figure 2B , red and green ) , we observed , however , that much of the difference in population heterogeneity between the deterministic model and the stochastic model disappeared . Population heterogeneity due to stochastic gene expression , therefore , apparently is small compared with population heterogeneity due to spatial and genetic variation . Intrinsic noise , as shown in the previous section , follows power law behavior with respect to protein number , with a coefficient of 0 . 5 , just like Poisson noise . Surprisingly , both in the deterministic model ( Figure 2A ) and the stochastic model ( Figure 2B ) we see that if spatial or genetic variation is present , the data can still reasonably be described using a power law . In Table 1 , we give the regression and correlation coefficient of a power law regression of the data . The regression coefficient indicates how strong population heterogeneity and protein number are correlated . The more sources causing population heterogeneity , the higher the regression coefficient is . Genetic heterogeneity correlates with protein number because the selection pressure on the induced operon is higher than on the repressed operon . The reason for this is that the cost for promoter activity is much more important for the induced operon than for the repressed operon . This appears to be a reasonable assumption that is likely to hold in the natural environment of E . coli . Population heterogeneity due to space is largest at intermediate extracellular inducer concentrations . At very high inducer concentrations , noise in the inducer concentrations does not cause noise in gene expression , because of the sigmoid shape of the promoter function . This also holds for very low extracellular inducer concentrations . For intermediate inducer concentration , gene expression is very much influenced by spatial heterogeneity . We observed , however , a monotonic relationship between protein number and noise due to spatial heterogeneity . This is because when the promoter activity of cells becomes low , cells stop lactose consumption , and therefore cells never experience very low lactose concentrations . This effect is likely to play a role in the natural environment , but in the natural environment , lactose degradation is probably not only due to E . coli , and we might expect noise due to spatial heterogeneity not to decrease monotonically with protein number . From all this we expect that only in a monomorphic population , living in a well-mixed environment , does stochasticity in gene expression play an important role . Whether stochasticity influences evolution , however , is therefore doubtful . To study whether stochasticity does or does not play an important role in evolution , in the next section we compare evolution of the lac operon in the stochastic and deterministic models . We do this in a well-mixed environment , because there stochasticity is expected to have the largest influence . We performed six independent evolutionary simulations in well-mixed environments , in both the deterministic model and the stochastic model . For all 12 last common ancestors , we plotted the promoter activity if no glucose is present ( Figure 3 ) . Most of the changes between the initial promoter function ( dotted line ) and the evolved promoter functions are consistent between all simulations . The induced transcription rate is increased , to approximately the same value in all 12 simulations . The steep part of the promoter function evolved to approximately 10-fold lower allolactose concentrations . Finally , the repressed transcription rate increased very significantly . As proven in [10] , the repressed transcription rate in large part determines whether a promoter function is bistable . The initial promoter function was chosen to be bistable . Of the 12 last common ancestors , only one was bistable . This is indeed the promoter function with the lowest repressed transcription rate . There appears to be a trend that promoter functions , which evolved in the stochastic simulations , have higher repressed transcription rates than those evolved in the deterministic simulations . To check this more precisely , we calculated the time average of the repressed transcription rate at zero glucose for all 12 simulations . There is considerable variation in this quantity over time during the whole evolutionary simulation . The first quarter of the evolutionary simulation was not taken into account , to give the cells time to adapt somewhat to the environment . We found that on average in the stochastic simulations the repressed transcription rate was 5 . 5-fold higher than in the deterministic simulations . In the stochastic simulations , we found an average repression rate of approximately 45 , while for the deterministic model this was approximately 250 . Previously , a repression rate of approximately 170 was experimentally found [9] . Whether this difference is significant , we checked by permuting the average repressed transcription rates over the different simulations and calculating the difference between the stochastic and deterministic simulations for all possible permutations . In less than 3% of the permutations , we found a larger average difference than observed , which gives a measure of the significance of this result . Next , we performed competition experiments between two promoter functions , one evolved in the stochastic model , the other in the deterministic model . We used the same environment as was used for the evolutionary simulations . For clarity we chose the promoter function with the highest repressed transcription rate and the promoter function with the lowest repressed transcription rate in Figure 3 . Both these promoter functions perform well in comparison with the five other evolved promoter functions . We placed these two promoter functions in the stochastic and deterministic model . When one population died out , we stopped the simulation and scored which promoter function had died out . No mutation was allowed during these competition experiments . Both for the stochastic model and for the deterministic model , we performed 100 of these competition experiments . In the deterministic model , we found that in 61 cases the promoter function with the low repressed transcription rate won , while in the stochastic model in 60 cases the promoter function with the high repressed transcription rate won . Using a two-tailed binomial test , this corresponds to a p-value of 0 . 057 and 0 . 035 , respectively . Combined , this gives a p-value of 0 . 0020 . This confirms that the promoter with the low repressed transcription rate performs better in the deterministic model , while the promoter with the high repressed transcription rate performs better in the stochastic model . Due to the small population sizes during the simulations ( on average approximately 200 ) , there is a lot of drift . This causes the most competitive promoter function to not always win the competition experiment . To understand this result , we studied the dynamics of these two promoter functions in both the deterministic model and the stochastic model . For both models , we again initiated a heterogeneous population and followed the dynamics of two individual cells , one with a high and one with a low repression rate . In both models , we used an identical environment , such that we can compare the dynamics . An example of the dynamics during a period of lactose influx is shown in Figure 4 . Figure 4A shows the dynamics of both promoter functions in the stochastic model , while Figure 4B shows the dynamics in the deterministic model , in the same environment . The promoter function with the high repressed transcription rate behaves almost identically , whether placed in the stochastic or the deterministic model . The only observable difference is caused by the cell cycle . For the promoter function with the low repressed transcription rate , the picture is very different . This promoter function is slightly bistable . Only when the external lactose concentration exceeds approximately 0 . 02 mM , does the operon switch on . This causes a slight delay in protein production , as can be seen in Figure 4B . In the stochastic model , the delay is , however , much larger . Only after approximately five hours , does the operon switch on . Because protein production occurs in bursts , and not in a gradual way , the cell has to wait for a sufficiently large burst to become induced . It is important to note that for the lactose concentrations the cells experience after approximately 4 . 5 hours , the operon is not bistable in the deterministic model , but , nevertheless , in the stochastic model the operon is not able to switch to the induced state . Figure 4 only shows the dynamics during one lactose pulse . In the evolutionary and competition experiments , many different pulses , of different height and length , are experienced . Also , the periods without influx have different lengths . Therefore , for every pulse the situation is somewhat different . During the periods without lactose influx , the promoter with the low repressed transcription rate has a higher growth rate than the promoter with the high repressed transcription rate , because its cost for operon activity is lower . When lactose influx starts , the operon with the high repressed transcription rate has a higher growth rate , because it can start lactose uptake earlier . Therefore , depending on the state of the environment ( whether there is lactose influx or not ) , the growth rate difference between the two promoters will be positive or negative . In an environment with the same parameters as the environment used for the evolutionary and competition experiments , we kept track of the growth rates during a period of time approximately equal to the time that is needed for one competition experiment . The result is shown in Figure 5 . The green line indicates that the promoter function with the high repressed transcription rate on average has a lower growth rate in the deterministic model than the promoter function with the low repressed transcription rate , although the growth rate difference indeed fluctuates , according to whether lactose is present in the environment or not . Again this proves that the promoter function with the low repressed transcription rate performs better in the deterministic model . When we compare both promoter functions in the stochastic model ( the black and the red line in Figure 5 ) , we see , however , that the average growth rate of the promoter function with the high repressed transcription rate is highest . This all shows that promoter functions evolved in the stochastic simulations evolve to a higher repressed transcription rate , which make them outcompete promoter functions with very low repressed transcription rates . In the deterministic simulations , the situation is the opposite , and these high repressed transcription rates are not optimal . Whether the cell is in equilibrium has considerable effect on the noise levels . In this section , we discuss two examples in which nonequilibrium dynamics is crucial for the amount of noise . Bistability in a deterministic system causes hysteresis . Depending on the history of the system , the system will be in one of the two equilibria . This hysteretic behavior disappears when stochasticity is added . In relatively short time scales , the system remains hysteretic , i . e . , when cells were induced , they remained induced , while when cells were repressed , they remained repressed . In longer time scales , transitions between the two equilibria are possible . Therefore , the probability distribution of the state of cells goes to a stable equilibrium , with some cells in the repressed and others in the induced state , depending not on the history but on the transition probability between the two equilibria . Therefore , if we would have waited long enough , the difference in noise levels between initially repressed and induced cells as shown in Figure 1 would have disappeared . Most of the cells would have gone to the induced state , because the transition probability to go from the induced to the repressed state is lower than vice versa , due to the lower noise for the induced operon . A second example is shown in Figure 6 . Here we show the population heterogeneity of a promoter function belonging to a last common ancestor of an evolutionary simulation . Both genetic and spatial heterogeneity are absent . We observe that the population heterogeneity is frequently lower than the intrinsic noise . Intrinsic noise is in general seen as inherent and therefore unavoidable for a cell . We would expect then that noise cannot be lower than the intrinsic noise . Here we see , however , that this is not necessarily so . This striking observation can be understood when we realize that the intrinsic noise was measured in cells that were in equilibrium . During our simulations , cells were very often not in equilibrium . Intuitively , we might expect that cells that are not in equilibrium have even higher population heterogeneity , but this is apparently not the case . If cells have a much higher protein concentration than the equilibrium value ( if , for example , external lactose has just been depleted ) , the protein concentration decreases . In such a situation , transcription can be neglected and the dynamics are purely determined by protein degradation and dilution . Noise caused by protein degradation and dilution is , however , much lower than transcriptional noise , because degradation does not occur in bursts . Noise due to degradation and dilution can be described as , Poisson noise ( solid line in Figure 6 ) . Translational noise is , however , much broader than Poisson: ( dashed line in Figure 6 ) , where b is the burst size . Indeed , we observe that Poisson noise does give a lower boundary to the population heterogeneity , whereas the intrinsic noise does not .
We study the influence of stochasticity in gene expression on evolutionary adaptation of the lac operon of E . coli . To this end , we used a detailed quantitative model of the lac operon in which stochasticity is incorporated on the protein level . This approach has the advantage that only one ( experimentally known ) parameter needs to be added to the model to make it stochastic , namely the average burst size of protein translation . We find good agreement between noise levels in our model and experimental noise measurements [2] . The experimentally observed noise levels , however , can only be explained when we realize that IPTG , which is not degraded by β-galactosidase , is used as inducer . We find that induction by IPTG leads to very different dynamics than induction by the natural inducer lactose ( see also [10] ) . When the operon is induced by lactose , stochasticity in gene expression is strongly reduced , and the total noise is mostly determined by the cell cycle . This is due to the fact that degradation of lactose by β-galactosidase reduces the strength of the positive feedback loop . Induction by TMG , however , leads to even higher noise values , because in contrast to IPTG , there is no protein-independent TMG influx . It would be very interesting to measure noise in gene expression of the lac operon , for TMG , IPTG , and lactose , to validate these results . In literature , different values for the Hill-coefficient are reported . For example , in [9] a Hill-coefficient of 2 is used , while in [12] a value of 4 . 0 was measured . When the Hill-coefficient is high , the positive feedback is strong and noise values are high . For IPTG , however , the noise curves would be very similar , because we chose the protein-independent inducer influx such that the maximal amount of noise corresponded to the experimentally measured noise values . For lactose we found that , all other parameters being equal , the promoter function only becomes bistable when the Hill-coefficient is larger than 52 , which is clearly unrealistically high . During the evolutionary experiments , the Hill-coefficient can be mutated and it mostly varies between 2 and 10 . To investigate the effect of different noise sources on population heterogeneity , we compared the amount of population heterogeneity in simulations with and without space , mutations , and stochasticity in gene expression . In these simulations the operons were induced by lactose , instead of by artificial inducers . We observed that only in the well-mixed simulations , without mutation , was the amount of population heterogeneity much larger in the stochastic than in the deterministic simulation . If spatial or genetic heterogeneity was added , stochasticity hardly influenced population heterogeneity . Surprisingly , we found that both genetic and spatial heterogeneity decrease with protein number , more or less in the same way as population heterogeneity by stochastic gene expression . Especially for genetic heterogeneity , we expect this also to be the case in nature . In nature , it seems likely that the spatial heterogeneity is very large . The gut is a highly diverse ecosystem [18] , and not at all well-mixed . It has been shown that E . coli is able to entirely change its lac operon promoter function in a few hundred generations [17] . This suggests that in nature the genetic diversity is also high . Therefore , we believe that the large genetic and spatial differences in our model are biologically realistic . However , we did check our results for a ten times lower mutation rate and ten times higher diffusion constant , which determine genetic and spatial heterogeneity , respectively . Even when both the mutation rate and the diffusion constant are modified , we observe that noise due to stochastic gene expression is only comparable to spatial and genetic heterogeneity at very low protein numbers ( unpublished data ) . Finally , we directly compared the promoter functions evolved in the deterministic and the stochastic evolutionary simulations . We observed that in the stochastic simulations cells evolve to higher repressed transcription rates and thus prevent stochasticity in gene expression . We show that this is because in the stochastic model low transcription rates cause longer delays than in the deterministic model . This is in striking contrast with the result found in [6] . There it was found that in Saccharomyces cerevisiae , bursts in gene expression enable a more rapid cell response . When we initialize cells in an environment with a fixed inducer concentration , for which the operon is bistable , we also see that larger bursts enable a more rapid response to this inducer concentration than smaller bursts . Indeed , in the deterministic model , cells would stay indefinitely in the “wrong” equilibrium . Because in our model , however , the inducer concentration varies over time , the inducer concentration quickly increases over the point at which the operon is still bistable , and in the deterministic model the cells then respond very fast . In the stochastic model , cells still need to wait for a sufficiently large burst to occur even when the inducer concentration has increased over the concentrations for which the cells are bistable . Even when the average burst size is large , this takes a long time , because then the frequency of the bursts is lower . This explains why the net effect of stochasticity in our model is negative for such promoters ( compare the growth rate of the promoter function with the low repressed transcription rate in the deterministic and the stochastic models; Figure 5 , red line ) . Furthermore , having a somewhat higher repressed transcription rate ensures that all cells have a rapid response . Both in the deterministic and the stochastic simulations , bistability with respect to lactose is most often avoided ( except at high glucose concentrations , which are very rare ) . Although the promoter function evolved in the deterministic model , which we used in Figure 4 , is slightly bistable , this has little influence on the behavior of this promoter function in the deterministic model , while the behavior in the stochastic model is changed drastically by the bistability . In [10] we already showed that when using a 10 times higher cost for lac operon activity , or a different environment ( with longer or shorter periods with and without lactose ) , bistability was also avoided . In the stochastic model , we also performed simulations in different environments , but again the results did not change essentially . We conclude that stochasticity cannot avoid the inherent disadvantages of bistability ( namely longer delays in protein dynamics [10] ) . Even more , we showed that bistability is even more deleterious in the presence of stochastic gene expression than in a deterministic system . These conclusions are in line with [4] , whose authors have shown that essential genes in S . cerevisiae have evolved to lower noise values than nonessential genes and thus that stochasticity for many genes appears to be detrimental during evolution .
The deterministic model is a spatially explicit , computational model of E . coli cells , growing on glucose and lactose while evolving their lac operon promoter function . It consists of an intracellular and an extracellular part . Intracellular dynamics . The intracellular dynamics is modeled using ten differential equations , following [19] . The following intracellular variables are incorporated: mRNA ( M ) β-galactosidase ( B ) , permease ( P ) , lactose ( L ) , allolactose ( A ) , glucose ( G ) , glucose-6-phosphate ( G6P ) , cAMP ( C ) , ATP ( ATP ) , and cell size ( X ) . Transcription of the lac operon , translation , lactose and glucose uptake and metabolism , cAMP and energy production , and cell growth are all modeled in detail . When possible , parameter values are taken from literature . All ten differential equations are integrated using a timestep of 0 . 2 seconds . Here we shortly discuss all differential equations . A list of all parameters is given in Table S1 . Transcription is modeled as a two-dimensional Hill-function , dependent on the cAMP and allolactose concentration ( see [12] ) . In this way , glucose , via cAMP , represses the operon , while lactose , via allolactose , induces the operon . This two-dimensional Hill-function depends on 11 biochemical parameters , such as the k-value and Hill-coefficient of allolactose binding to the repressor . In the evolutionary simulations , these are the only parameters that can mutate , all other parameters are fixed , because we are interested in the evolution of the promoter function , given realistic boundary conditions ( which are the other parameters in the intracellular model ) . a = RNAP/kRNAP , RNA-polymerase in units of its dissociation constant for binding to a free site . b = RNAP/kRNACP , RNA-polymerase in units of its dissociation constant for binding to a site with bound CRP ( cAMP receptor protein ) . c = LACIT/kLACI , the total LacI concentration in units of its dissociation constant for binding to its site . d = CRPT/kCRP , the total CRP concentration in units of its dissociation constant for binding to its site . α , the transcription rate when RNA Polymerase is bound to the DNA , but CRP and LacI are not . β , the transcription rate when both RNA Polymerase and CRP are bound , but LacI is not bound to the DNA . γ , “leakiness , ” the transcription rate when RNA Polymerase is not bound to the DNA . kA , k-value for allolactose binding to LacI . m , Hill-coefficient describing cooperativity in binding of allolactose to LacI . kC , k-value for cAMP binding to CRP . n , Hill-coefficient describing cooperativity in binding of cAMP to CRP . Protein production ( β-galactosidase and permease ) depends on the mRNA concentration , and proteins are slowly degraded . Lactose influx is permease-dependent , while lactose degradation and conversion to allolactose is dependent on β-galactosidase . Small protein-independent lactose and allolactose degradation terms also are added . Lactose is converted to glucose by β-galactosidase . Glucose uptake from the medium also is incorporated . Glucose metabolism ( via glycolysis and TCA-cycle ) produces ATP on which the cells grow . ATP production depends on the glycolytic fluxes and the fluxes through the TCA-cycle . ATP is consumed by basal metabolism , cell growth , and lac operon activity . The cAMP concentration is assumed to be dependent on the glucose influx rate ( see [19] ) . The spatial model . Cells , of which the dynamics are determined by the above-described intracellular model , are placed on a square grid of 25 × 25 grid points . These cells grow on extracellular glucose and lactose . The extracellular glucose and lactose concentrations are determined by a fluctuating influx of glucose and lactose into the grid , consumption of glucose and lactose by the cells , and diffusion over the grid . In the well-mixed simulations , we assume infinite diffusion , such that the glucose and lactose concentrations are equal over the whole grid . Furthermore , the cells are shuffled randomly over the grid . Glucose and lactose influx into the grid is modeled in pulses , independently of each other . Pulses of glucose and lactose influx both have an average duration of 11 hours . The total amount of carbon influx is on average equal for each pulse , such that short pulses have high influx rates , and vice versa . Every pulse has a probability of 10% of being instantaneous , such that very high glucose or lactose concentrations also sometimes occur . In 25% of these cases , simultaneous glucose AND lactose influx occurs , in order to enforce simultaneous high glucose AND lactose concentrations . These pulses are followed by a period without glucose or lactose influx , also on average 11 hours . This environment is chosen for two reasons . First , all combinations of glucose , lactose , glucose AND lactose , and neither one occur repeatedly . Second , the average length of the periods is chosen such that cells can just adapt their protein concentrations to the environment . The cells consume glucose and lactose , grow ( as the growth rate is given by the intracellular model ) , divide ( if their size has doubled and there is space to reproduce ) , and move randomly over the grid . The cells are diluted in a density-dependent way , and cells “die” if they have zero energy ( although this happens very rarely ) . Therefore , fitness differences are mainly caused by differences in growth rates between cells ( which are again dependent on glucose and lactose uptake ) instead of by differences in death rates . In the evolutionary experiments , we always start with a monomorphic population of approximately 60 cells that are bistable with respect to lactose and have a diauxic shift ( such that glucose and lactose are consumed sequentially ) . After each cell division , the daughter cell has a probability of 0 . 01 to mutate one of the 11 biochemical parameters that determine the shape of the lac operon promoter function . All other parameters are kept constant . This means that cells can adapt the precise shape of the promoter function to the ( fluctuating ) environment . We let the cells evolve for approximately 2 , 000 days , which is sufficient to adapt to the environment .
|
Gene expression is a process that is inherently stochastic because of the low number of molecules that are involved . In recent years it has become possible to measure the amount of stochasticity in gene expression , which has inspired a debate about the importance of stochasticity in gene expression . Little attention , however , has been paid to stochasticity in gene expression from an evolutionary perspective . We studied the evolutionary consequences of stochastic gene expression in one of the best-known systems of genetic regulation , the lac operon of E . coli , which regulates lactose uptake and metabolism . We used a computational approach , in which we let cells evolve their lac operon promoter function in a fluctuating , spatially explicit , environment . Cells can in this way adapt to the environment , but also change the amount of stochasticity in gene expression . We find that cells evolve their repressed transcription rates to higher values in a stochastic model than in a deterministic model . Higher repressed transcription rates means less stochasticity , and , hence , these cells appear to avoid stochastic gene expression in this particular system . We find that this can be explained by the fact that stochastic gene expression causes a larger delay in lactose uptake , compared with deterministic gene expression .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"evolutionary",
"biology",
"cell",
"biology",
"eubacteria",
"computational",
"biology"
] |
2007
|
The Effect of Stochasticity on the Lac Operon: An Evolutionary Perspective
|
Microsatellite expansions cause a number of dominantly-inherited neurological diseases . Expansions in coding-regions cause protein gain-of-function effects , while non-coding expansions produce toxic RNAs that alter RNA splicing activities of MBNL and CELF proteins . Bi-directional expression of the spinocerebellar ataxia type 8 ( SCA8 ) CTG CAG expansion produces CUG expansion RNAs ( CUGexp ) from the ATXN8OS gene and a nearly pure polyglutamine expansion protein encoded by ATXN8 CAGexp transcripts expressed in the opposite direction . Here , we present three lines of evidence that RNA gain-of-function plays a significant role in SCA8: 1 ) CUGexp transcripts accumulate as ribonuclear inclusions that co-localize with MBNL1 in selected neurons in the brain; 2 ) loss of Mbnl1 enhances motor deficits in SCA8 mice; 3 ) SCA8 CUGexp transcripts trigger splicing changes and increased expression of the CUGBP1-MBNL1 regulated CNS target , GABA-A transporter 4 ( GAT4/Gabt4 ) . In vivo optical imaging studies in SCA8 mice confirm that Gabt4 upregulation is associated with the predicted loss of GABAergic inhibition within the granular cell layer . These data demonstrate that CUGexp transcripts dysregulate MBNL/CELF regulated pathways in the brain and provide mechanistic insight into the CNS effects of other CUGexp disorders . Moreover , our demonstration that relatively short CUGexp transcripts cause RNA gain-of-function effects and the growing number of antisense transcripts recently reported in mammalian genomes suggest unrecognized toxic RNAs contribute to the pathophysiology of polyglutamine CAG CTG disorders .
Spinocerebellar ataxia type 8 ( SCA8 ) , is a slowly progressive neurodegenerative disease caused by a CTG•CAG expansion that primarily affects the cerebellum [1] , [2] . The disease is transmitted in an autosomal dominant pattern with reduced penetrance and the expansion mutation was originally shown to be expressed as a CUG expansion ( CUGexp ) transcript in the 3′region of an untranslated gene , ATXN8OS . BAC transgenic mice expressing the SCA8 expansion ( SCA8 BAC-EXP , [CTG]116 ) but not a control repeat ( SCA8 BAC-CRTL , [CTG]11 ) from the endogenous human promoter develop progressive motor deficits and a loss of cerebellar GABAergic inhibition [3] . Unexpectedly , SCA8 patients and BAC-EXP mice were found to have 1C2-positive intranuclear inclusions in Purkinje cells and brainstem neurons that result from the expression of a nearly pure polyglutamine protein ( ataxin 8 ) from a novel gene spanning the repeat in the opposite CAG direction ( ATXN8 ) . The expression of CUGexp transcripts from ATXN8OS in addition to CAGexp transcripts and a polyglutamine protein from ATXN8 suggests that SCA8 involves toxic gain-of-function effects at both the RNA ( CUGexp ) and protein ( PolyQ ) levels . An alternative hypothesis is that the SCA8 expansion affects the expression of an overlapping gene , KLHL1 and although Klhl1 knockout mice have a subtle phenotype , the relevance of this model to the human disease is unclear [4] . Substantial evidence that CUGexp RNAs are toxic comes from studies of the neuromuscular disease myotonic dystrophy ( DM ) where CUGexp ( DM1 ) and CCUGexp ( DM2 ) RNAs alter the activities of two families of alternative splicing factors , the MBNL and CELF proteins [5] , [6] . Multiple lines of evidence support the model that these expansion transcripts cause disease specific clinical features . First , transgenic mouse models of DM1 in which CUGexp transcripts are expressed in the 3′ UTR of the DM1 ( DMPK ) , or the unrelated human skeletal muscle actin ( HSA ) , gene cause skeletal muscle myotonia and myopathy similar to human DM skeletal muscle pathology [7] , [8] . In addition , CUGexp or CCUGexp transcripts accumulate as ribonuclear inclusions in DM1 and DM2 patient skeletal muscle [9]–[11] and alter the localization or regulation of RNA binding proteins CUGBP1 [12] , [13] and MBNL1 [14] , [15] . Additional studies show expression of CUGexp transcripts induce alternative splicing changes in skeletal muscle genes associated with disease symptoms including the chloride channel ( CLCN1 ) and insulin receptor genes ( IR ) [16]–[19] . Similar alternative splicing events in numerous genes have now been shown to be misregulated in myotonic dystrophy , often causing aberrant expression of fetal isoforms in adult tissue [5] . Evidence that MBNL1 loss-of-function plays a role in DM was suggested by studies showing MBNL1 co-localizes with CUG or CCUG ribonuclear inclusions [10] with additional strong support coming from Mbnl1 isoform knockout mice ( Mbnl1ΔE3/ΔE3 ) that recapitulate several aspects of the multisystemic disease pathology and misregulated splicing events characteristic of DM [20] . Furthermore , MBNL1 has been shown to bind directly to intronic elements of genes misregulated in DM cardiac and skeletal muscle ( cTNT and IR ) and to promote alternative splicing patterns normally found in adult tissue [21] . Taken together , these data suggest that expression of CUGexp or CCUGexp transcripts induces alternative splicing changes in DM skeletal and cardiac muscle by sequestration and functional loss of MBNL1 protein . In addition to skeletal and cardiac muscle disease , behavioral and cognitive changes in DM suggest that CUGexp and CCUGexp transcripts also cause CNS effects . DM1 CUGexp transcripts form ribonuclear inclusions that co-localize with MBNL1 in temporal lobe neurons of DM1 patients [22] . Additionally , alternative splicing changes in a number of CNS transcripts ( NMDA R1 , APP , MAPT/Mapt , Mbnl1 ) are found in humans and mice [22] , [23] . Although the mechanism for these CNS splicing changes is unknown , these data suggest that DM , and possibly the much shorter SCA8 CUG expansion transcripts , cause RNA gain-of-function effects in brain . In support of the model that SCA8 CUGexp transcripts affect MBNL1 regulated pathways , expression of human SCA8 cDNA transcripts ( ATXN8OS ) in Drosophila photoreceptor neurons induce a late-onset , neurodegenerative phenotype genetically enhanced by the loss of the fly MBNL1 orthologue muscleblind [24] . This result is similar to fly models of DM1 [25] , [26] and consistent with additional evidence for RNA gain-of-function effects in DM . In this study , we present the first evidence that SCA8 CUGexp transcripts cause “RNA gain-of-function” effects in the brain . First , we demonstrate that SCA8 CUGexp transcripts form hallmark ribonuclear inclusions that co-localize with MBNL1 in humans and mice and that genetic loss of Mbnl1 enhances motor coordination deficits in SCA8 BAC-EXP mice . Additionally , we show that expression of ATXN8OS CUGexp transcripts dysregulate MBNL1-CUGBP1 pathways in the CNS and trigger downstream molecular changes in GABA-A transporter 4 ( Gabt4 ) regulation through an RNA gain-of-function mechanism .
To determine if SCA8 ATXN8OS transcripts form ribonuclear inclusions similar to those found in DM , fluorescence in-situ hybridization ( FISH ) was preformed on SCA8 human and SCA8 BAC-EXP mouse cerebella using a Cy5-labeled ( CAG ) 10 oligonucleotide probe . In human SCA8 autopsy tissue , CUG positive inclusions were found in the cerebellar cortical layers ( red dots ) by confocal microscopy . These CUG ribonuclear inclusions are distinguishable from background lipofuscin auto-fluorescence which appears as a yellowish-brown perinuclear staining ( Figure 1A ) . Although ribonuclear inclusions were seen in all three SCA8 autopsy brains examined , the foci varied in distribution , size , and number between SCA8 cases and compared to ribonuclear inclusions found in DM1 cerebellum . In cerebellar tissue from SCA8 patients with 1000 and 400 CTG repeats , single CUG foci were frequently found in the nuclei of molecular layer ( ML ) interneurons and the Bergmann glia surrounding the Purkinje cells in the granule cell ( GC ) layer , while multiple smaller foci were typically found in Purkinje cells ( PC ) . Although qRT-PCR shows CUGexp transcripts are expressed at comparable levels in all three SCA8 autopsy cases , ribonuclear inclusions were not reproducibly found in the brain with 109 CTG repeats , and were only detected in a single molecular layer interneuron ( not shown ) . While these data suggest that ribonuclear inclusions are less likely to form in patients with shorter expansions , CUG RNA foci were readily detectable in SCA8 BAC transgenic mice which express similarly sized SCA8 ATXN8OS ( CUG ) 116 transcripts under the control of the endogenous human promoter ( Figure 1A , bottom row ) . Similar to the RNA foci found in SCA8 autopsy tissue with 400 and 1000 repeats , the foci in these mice have a similar distribution in the cerebellar cortex ( Purkinje cells , Bergmann glia , and molecular layer interneurons ) . Additionally , ribonuclear inclusions were found in the deep cerebellar nuclei in our mice , an area of the brain we were not able to examine in the human autopsy cases . No foci were seen in mice expressing normal repeat ( CTG ) 11 or non-transgenic littermates and additionally , no CAG foci were detected using a CTG oligonucleotide probe ( not shown ) . FISH combined with immunofluorescence ( IF ) was performed to determine if the SCA8 ribonuclear inclusions co-localize with MBNL1 . In human cerebellar sections , MBNL1 co-localizes with the CUG foci in molecular layer interneurons ( Figure 1B ) . In contrast , SCA8 and DM1 Purkinje cells have clearly detectable nuclear CUG foci but these foci do not co-localize with MBNL1 , which is predominantly expressed in the cytoplasm ( Figure 1B , middle and bottom row arrows ) . Similar to the human results , SCA8 mice expressing the expansion , but not the control repeat , have CUG-positive ribonuclear inclusions that co-localize with Mbnl1 in molecular layer interneurons and deep cerebellar nuclei ( Figure 1C ) but not in Purkinje cells ( not shown ) . Because co-localization of MBNL1 with CUGexp ribonuclear inclusions is thought to lead to functional impairment of nuclear MBNL1 activity in DM1 , we tested the hypothesis that RNA gain-of-function effects in SCA8 contribute to the motor deficits in SCA8 BAC-EXP mice via Mbnl1 depletion . Mice from a low copy SCA8 BAC-EXP5 line ( SCA8 BAC-EXP5+/− ) were crossed to heterozygous Mbnl1 isoform knockout mice ( Mbnl1+/ΔE3 ) . Mice from the SCA8 BAC-EXP5+/− line were selected for these studies because these animals , which have normal rotarod performance at 26 weeks of age , do not develop a movement disorder phenotype until >1 year of age [3] . Additionally , although homozygous Mbnl1ΔE3/ΔE3 knockout mice model the multisystemic features of DM pathology heterozygous Mbnl1+/ΔE3 mice are similar to wild type and do not develop myotonia or other skeletal muscle changes [20] . To test if genetic Mbnl1 loss enhances the SCA8 CNS phenotype we crossed heterozygous SCA8 BAC-EXP5+/− mice to heterozygous Mbnl1+/ΔE3 knockout mice and tested the F1 offspring [ ( SCA8+/− ( n = 11 ) ; Mbnl1+/ΔE3 ( n = 13 ) ; SCA8+/−; Mbnl1+/ΔE3 ( n = 17 ) and non-transgenic littermates ( n = 9 ) ] for motor deficits at 26 weeks of age by rotarod analysis . The latency to fall in seconds ( sec ) was recorded for 4 trials per day over 4 consecutive days . Mean differences between groups for each testing day were determined by taking the average of the last 3 trials . Consistent with previous results [3] , no significant difference in mean latency to fall was found between the SCA8 BAC-EXP5+/− mice ( 409 . 52±22 . 51 sec ) and non-transgenic littermates ( 414 . 83±22 . 18 Sec . ; P = 0 . 86 ) . Although heterozygous Mbnl1+/ΔE3 mice did have a significantly different mean latency to fall ( 342 . 15±18 . 75 sec . ) compared to non-transgenic littermates [F ( 1 , 86 ) = 6 . 22; P = 0 . 012] , double mutant offspring ( SCA8+/−; Mbnl1+/ΔE3 ) performed significantly worse ( 273 . 69±14 . 13 Sec ) than either the singly mutant Mbnl1+/ΔE3 ( P = 0 . 003 ) or SCA8 BAC-EXP5+/− littermates [F ( 1 , 102 ) = 31 . 27; P<0 . 00001] ( Figure 2 ) . These data provide genetic evidence that loss of Mbnl1 plays a role in SCA8 pathogenesis and support the hypothesis that SCA8 CUGexp transcripts affect Mbnl1 regulated pathways in the brain . The discovery that CUGexp RNA and MBNL1 protein co-accumulate in ribonuclear foci in neurons suggests that , similar to the dysregulation of MBNL/CELF pathways in DM muscle , the accumulation of SCA8 CUG expansion transcripts might lead to the dysregulation of developmentally regulated splicing patterns in SCA8 brain . Further support for this hypothesis comes from alternative splicing changes found in MBNL1 and NMDAR1 in SCA8 autopsy brains ( Figure 3A ) which mimic previously reported changes in DM1 [22] , [27] . To identify novel splicing targets that might be affected by Mbnl1 loss or increased Cugbp1 activity we used cross-linking and immunoprecipitation ( CLIP ) analysis [28] on mouse postnatal day 8 ( P8 ) hindbrains . CLIP was successful in identifing RNA targets of Cugbp1 but not Mbnl1 . Because CUG-BP1 and MBNL1 have been shown to be antagonistic regulators of alternative splicing of a number of different targets from work done in the myotonic dystrophy field , we reasoned that CUG-BP1 CLIP tags would be also be good candidate targets for MBNL1 regulation . At P8 , hindbrain can be readily dissociated into a cell suspension which was then exposed to UV-light to fix RNA-protein complexes formed in vivo . This procedure avoids artifacts associated with immunopurification of unfixed proteins which may redistribute during in vitro manipulation . Cross-linked RNA-protein complexes were treated with RNase T1 , to generate relatively short ( 60–200 nt ) CLIP RNA tags , and immunopurified with a monoclonal antibody against Cugbp1 . Complexes were then subjected to electrophoresis/electroblotting and filter-retained RNA tags were identified following cDNA conversion , amplification and DNA sequencing . A total of 315 Cugbp1-associated RNA tags were identified representing 206 genes with 53 multi-hit tags ( Figure 3B and Table S1 ) . While the majority of these tags were intronic ( 64% ) , in agreement with a previous study on the splicing regulator Nova1 [28] , a significant number ( 25% ) were positioned within 3′ untranslated regions ( UTRs ) . Sequence analysis revealed that Cugbp1 CLIP tags were enriched in UG repeats consistent with previous three-hybrid and SELEX studies which have indicated that the highest affinity sites for CUGBP1 , CUGBP2 and CELF4 are ( UG ) n , and ( UGUU ) n repeats [29] , [30] . A recent study identified a GU-rich 11-mer element ( GRE ) UG ( U ) 3G ( U ) 3GU , which is enriched in the 3′-UTRs of short-lived transcripts , as a CUGBP1 binding motif [31] . Interestingly , only one Cugbp1 3′ UTR CLIP tag ( Slc22a5 ) , and two intronic tags ( Lrrc9 , Samd12 ) , contained this GRE motif ( see Table S1 ) . From this list of potential Cugbp1 targets we focused on the gamma-aminobutyric acid ( GABA-A ) transporter 4 gene , Gabt4 , because previous in vivo functional imaging studies showed a loss of cerebellar GABAergic inhibition in our SCA8-BAC-EXP mice [3] . Further analysis of the Gabt4 CLIP tag showed that the putative Cugbp1 binding site maps to exon 7 ( and overlaps the exon 7 5′ splice site ) which is highly conserved in human and mouse . The gene name in mouse is Gabt4/Slc6a11 and in human GAT3 but for clarity referred to here as Gabt4 and GABT4 , respectively . Since Gabt4 was identified as a putative target of CUGBP1 by CLIP analysis and because increased Gabt4 expression could explain the increased cortical loss of GABAergic inhibition previously reported in our mice by reducing GABA at the synapse , we hypothesized that SCA8 CUGexp RNA dysregulates Cugbp1 and Mbnl1 pathways resulting in an increase in Gabt4 expression . Consistent with this idea , we found significant increases in cerebellar Gabt4 protein and RNA levels by protein blot ( 5 . 36±1 . 11 fold; p = 0 . 003 ) and qRT-PCR ( 2 . 72±0 . 68 fold p = 0 . 0015 ) in SCA8 BAC-EXP1 mice compared to non-transgenic littermates ( Figure 4A and 4B ) while no increase was seen in SCA8 BAC-CTRL animals ( Figure 4B ) . Further supporting the hypothesis that Gabt4 increases in SCA8 are caused by sequestration of Mbnl1 by SCA8 CUGexp transcripts , Gabt4 protein and transcript levels are also higher in Mbnl1ΔE3/ΔE3 knockout mice compared to strain matched ( 129/B6 ) non-transgenic littermate controls with a 2 . 49±0 . 083 ( P<0 . 001 ) and 2 . 29±0 . 32 ( P = 0 . 013 ) mean fold increase by protein blot and qRT-PCR , respectively ( Figure 4A and 4B ) . The increases in Gabt4 protein were reproducible in both high copy number SCA8 BAC-EXP lines studied ( BAC-EXP1 and BAC-EXP2 ) but more variable in mice from the lower copy number BAC-EXP5 line ( data not shown ) consistent with the decreased penetrance reported previously in this line . In addition , no similar changes were seen in Gabt1 levels , another member of GABA-A transporter family expressed in cerebellum indicating that changes seen in Gabt4 are not caused by a general upregulation of GABA-A transporters ( Figure 4C ) . Immunofluorescence studies show expression of Gabt4 protein is primarily localized to the granular cell layer of the cerebellar cortex and the deep-cerebellar nuclei ( DCN ) and that no overt change in this distribution is seen ( Figure 4D ) between wildtype and SCA8 BAC-EXP1 mice . To further characterize the loss of inhibition phenotype and the possible role of Gabt4 we examined the SCA8 BAC mice for functional changes in the granular layer of the cerebellum , the site of highest Gabt4 expression using flavoprotein optical imaging in response to whisker pad stimulation in SCA8 mice [32] , [33] . Activation of cerebellar granule cells by mossy fibers is in part controlled by Golgi cell mediated feedback that produces GABAergic inhibition of granule cells [34] . Because the clearance of synaptically released GABA in the granular layer is predominately mediated by Gabt4 , up-regulation of Gabt4 would be expected to reduce this Golgi mediated inhibition and enhance the responses to mossy fiber input . The imaging data collected confirm this prediction . Whisker pad stimulation evokes a patch-like response ( Figure 4E ) consistent with previous electrophysiological and imaging studies [32] , [35] . There is a significant increase in both the intensity and area of the response in Crus II in the SCA8 BAC-Exp ( n = 5 ) compared with FVB mice ( n = 7 ) ( Figure 4 , p<0 . 05 ) . Therefore , up-regulation of Gabt4 in the granular layer is associated with the expected increase in the response to mossy fiber inputs activated by peripheral stimulation . In summary , these data demonstrate that Gabt4 , a gene identified by CLIP analysis as a Cugbp1 target , is upregulated at the RNA and protein levels in both our SCA8 BAC-EXP1 and Mbnl1ΔE3/ΔE3 knockout animals but not in the SCA8 BAC-CTRL mice expressing a normal length CUG11 repeat . Additionally , Gabt4 upregulation in these mice is associated with the predicted loss of GABAergic inhibition within the granular cell layer . To determine if GABT4 upregulation also occurs in humans , we examined steady-state RNA and protein levels in human SCA8 autopsy brains . While expression of the SCA8 CUGexp and CAGexp transcripts and GABT4 overlap in both the cerebellum and the frontal lobe , only frontal lobe tissue was suitable for analysis because of significant cell loss in the cerebellum caused by neurodegeneration in SCA8 patients . Total RNA extracted from adult SCA8 , DM1 and control and from 26-week fetal frontal cortex was examined by qRT-PCR using primers to exons 1 and 2 . SCA8 autopsy brains ( n = 3 ) showed increased GABT4 transcripts levels compared to adult control ( p<0 . 01 ) ( Figure 5A ) and similar levels to those found in fetal tissue . A similar trend of increased GABT4 protein was also seen in SCA8 and fetal brain compared to adult control or DM1 tissue ( Figure 5B ) . Because exon 7 of Gabt4 was identified as a potential Cugbp1 binding site , we investigated if increases in GABT4 expression could be related to misregulation of exon 7 alternative splicing . Consistent with this hypothesis , all three human SCA8 autopsy brains showed a shift in alternative splicing favoring inclusion of GABT4 exon 7 containing transcripts compared to adult control ( Figure 5C ) . Preferential exon 7 inclusion was also seen in control human fetal autopsy tissue but not in adult DM1 tissue and is correlated with the increases in GABT4 RNA and protein seen in SCA8 adult and control fetal brain ( Figure 5 ) . Interestingly , no exon 7 splicing shifts or increases in GABT4 expression were found in DM1 autopsy tissue , possibly reflecting differences in spatial or temporal expression of the SCA8 and DM1 CUGexp transcripts . Similar to the dysregulation of other genes in DM , SCA8 CUGexp transcripts trigger a shift in the ratios of alternatively spliced human GABT4 ( +/− exon 7 ) transcripts that resemble those found during fetal development . Sequence analysis shows that transcripts skipping exon 7 would lead to the introduction of a premature termination codon ( PTC ) which would be predicted to target ( − ) exon 7 transcripts for nonsense mediated decay ( NMD ) . These data suggest a model in which developmental alternative splicing changes in human normally lead to lower levels of GABT4 in adult tissue and higher levels in SCA8 and during fetal development . Similar to humans , Gabt4 is also up-regulated in the SCA8 BAC-Exp and Mbnl1ΔE3/ΔE3 mice . Although alternative splicing of exon 7 has not yet been detected in the mouse by RT-PCR , exclusion of mouse exon 7 would also create a PTC ( in exon 10 ) that would be predicted to lead to NMD . Further studies are needed to determine if NMD in the mouse leads to more efficient degradation of ( − ) exon 7 transcripts which prevent their detection or if Gabt4 upregulation in the mouse occurs via another mechanism . To test directly if increases in GABT4 are induced by SCA8 CUGexp or CAGexp transcripts , we examined their effects in human neuroblastoma SK-N-SH cells . Transient transfections were performed using minigenes ( Figure 6A ) designed to express SCA8 CUG ( SCA8-CTGexp ) or CAG expansion transcripts ( SCA8-REV CAGexp ) . Cells expressing SCA8 CUGexp transcripts show significant increases in GABT4 RNA levels by qRT-PCR relative to untransfected cells ( p = 0 . 007 ) or to cells transfected with vector alone ( p = 0 . 006 ) while expression of the SCA8 CAGexp construct compared to vector alone had no effect ( p = 0 . 72 ) ( Figure 6B ) . Additionally , transient transfections of minigenes without ATXN8 and ATXN8OS flanking sequence show poly-CUG105 but not poly-CAG105 transcripts up-regulate GABT4 RNA ( p = 0 . 001 ) ( Figure 6C ) . Further analysis using primers flanking exon 7 , show cells expressing higher levels GABT4 transcripts also preferentially express higher ratios of exon 7 included transcripts . ( Figure 6B and 6C ) . To test if GABT4 expression is antagonistically regulated by CUGBP1 and MBNL1 , we transfected SK-N-SH cells using GFP-tagged human MBNL1/41 and CUGBP1 minigenes capable of inducing alternative splicing changes in cell culture [21] , [36] . As above , alternatively-spliced exon 7 transcripts were assayed by RT-PCR with primers located in GABT4 exons 6 and 8 and primers spanning exons 1 and 2 were used to assess endogenous levels of GABT4 by qRT-PCR . Cells overexpressing CUGBP1 ( p<0 . 0001 ) or SCA8 exon A CTGexp minigenes show increases in GABT4 RNA ( p = 0 . 003 ) and a concomitant increases in protein and a splicing shift favoring exon 7 inclusion compared to vector alone ( Figure 7A–7C ) . While no change in GABT4 RNA or protein was seen in cells overexpressing MBNL1/41 alone , overexpression of MBNL1/41 and SCA8 CUGexp transcripts reverses the increase in GABT4 RNA ( p<0 . 0001 ) and protein ( Figure 7A and 7C ) triggered by SCA8 CUGexp transcripts alone and restores exon 7 alternative splicing ratios to the normal adult pattern ( Figure 7B ) . Sequence analysis of the GABT4 RT-PCR products confirm the upper and lower bands include and exclude exon 7 respectively and that the ( - ) exon 7 transcripts have a premature stop codon in exon 8 . Taken together , these results are consistent with a model in which SCA8 CUGexp transcripts alter the regulation of GABT4 by sequestration of MBNL1 and/or an increase in the expression or activity of CUGBP1 .
The SCA8 CTG CAG mutation is bidirectionally expressed and produces both CUG and CAG-expansion transcripts and a nearly pure polyglutamine expansion protein [3] . We investigated RNA gain-of-function effects in SCA8 and present three lines of evidence that CUGexp transcripts play a role in SCA8 . First , we demonstrate CUGexp transcripts accumulate as ribonuclear inclusions in selected cells and that these RNA foci co-localize with Mbnl1 in a subset of neurons in SCA8 patients and mice . Second , we show genetic loss of Mbnl1 enhances motor coordination deficits in low-copy SCA8 BACexp mice . Third , we demonstrate SCA8 CUGexp transcripts trigger increased expression of a CUGBP1-MBNL1 regulated CNS target , GABT4 , in both mice and humans as well as in a human cell culture model and demonstrate the predicted loss of GABAergic inhibition within the granular cell layer occurs in these animals . Although CNS effects are a clinically important feature of myotonic dystrophy , mechanistic studies have focused on skeletal and cardiac muscle and little is known about the effects of CUGexp transcripts in brain . In addition , the antagonistic relationship between CUGBP1 and MBNL1 , previously documented in heart and skeletal muscle , has not been demonstrated in the CNS . In this study we used CLIP analysis to identify putative Cugbp1 CNS targets . To explore the molecular basis of the loss of GABAergic inhibition we investigated changes in expression levels and alternative splicing of the Cugbp1 CLIP target , Gabt4 , in our mice . We show that overexpression of CUGBP1 in human SK-N-SH cells , or depletion of MBNL1 , result in an upregulation of GABT4 that mimics the in vivo changes in steady state levels caused by CUGexp transcripts . These data provide the first evidence that CUGexp RNA gain-of-function effects in the brain involve the dysregulation of CUGBP1-MBNL1 pathways . These data also suggest that MBNL1 overexpression , which has been demonstrated to be therapeutic in skeletal muscle , might also be an effective treatment to reverse pathological changes associated with expression of CUGexp transcripts in the CNS . Additionally , the long list of other putative Cugbp1 targets identified in mouse brain provides an important future resource for the identification of additional CNS genes dysregulated in SCA8 and DM . Previous studies in myotonic dystrophy suggest that the expression of CUGexp transcripts , together with MBNL1 and CELF proteins and their downstream target genes need to be coordinated temporally and spatially for disease pathogenesis . Therefore , defining the temporal and cell specific ATXN8OS expression pattern will be crucial for understanding disease pathogenesis and interpreting future results . We identified for the first time in SCA8 , CUG-RNA foci in Purkinje cells , molecular layer interneurons and the deep cerebellar nuclei . In addition , we show MBNL1/Mbnl1 colocalizes with SCA8 CUG foci in molecular layer interneurons and the DCN . Interestingly , although Purkinje cells had nuclear CUG foci in both DM1 and SCA8 , co-localization with nuclear MBNL1 was not observed in these cells . Further studies will need to be done to determine if MBNL1 is expressed in the nucleus of these cells or if changes in MBNL1 splice forms affect its ability to bind to the CUG expansion transcripts . Taken together , these results suggest that temporal and spatial expression patterns of expansion transcripts and the overlap in expression of specific RNA binding proteins and their downstream target genes are likely to underlie the susceptibility of specific cells to RNA gain-of-function effects and the clinical difference between DM and SCA8 . To investigate the broader significance of MBNL1 in SCA8 , we tested the effects of Mbnl1 depletion on a behavioral phenotype in our SCA8 mice by crossing heterozygous Mbnl1+/ΔE3 animals with a low-copy SCA8-BAC-EXP5 line . The phenotypic enhancement of the rotarod deficits found in the doubly transgenic animals with reduced Mbnl1 suggests that expression of CUGexp transcripts in SCA8 and the subsequent downstream effects on Mbnl1 are sufficiently significant to contribute to the movement disorder phenotype found in SCA8 . GABT4 upregulation in human SCA8 autopsy tissue and cell culture studies supports a model in which GABT4 levels are regulated by alternative splicing changes and the NMD pathway . Similar to the dysregulation of other genes in DM1 ( CLCN-1 , IR ) [5] , the expression of CUGexp but not CAGexp transcripts [37] triggers a shift in the ratios of alternatively spliced GABT4 transcripts that resembles those found during fetal development . Further studies are needed to determine if GABT4 alternative splicing changes occur throughout the brain and to directly show if alternative splicing of exon 7 and NMD regulate GABT4/Gabt4 expression levels in both humans and mice . Additional studies are also required to directly test the hypothesis that GABT4 upregulation causes the decrease in GABAergic inhibition seen in our SCA8 BAC-EXP mice . While these data provide the first evidence for RNA gain-of-function effects in SCA8 , it is also possible that the polyQ expansion protein contributes to the disease . In contrast to other disorders in which polyQ expansions are expressed as part of a mature protein , the SCA8 CAGexp is expressed as a nearly pure polyQ tract . While previous studies have shown that pure polyQ repeats are toxic in Drosophila and mice , the relative contribution of RNA and protein gain-of-function effects in SCA8 still needs to be assessed . In this study we provide several lines of evidence that RNA gain-of-function effects play a significant role in SCA8 and show that SCA8 CUGexp transcripts affect alternative splicing patterns controlled by MBNL1 and CUGBP1 in the mammalian brain . While SCA8 is the first reported disease in which a single expansion mutation expresses both a polyglutamine protein and CUGexp transcripts , bidirectional expression has also been recently described in other triplet expansion disorders . For example , sense ( CUGexp ) and antisense ( CAGexp ) transcripts have been reported at the DM1 locus [38] . Similarly , Huntington disease Like 2 ( HDL2 ) , another disease in which CUGexp transcripts form ribonuclear inclusions [39] , also has 1C2-positive inclusions ( 1C2 is an antibody that recognizes polyglutamine expansions ) , suggesting bidirectional expression may also occur in that disease [40] , [41] . Additionally , bidirectional expression across the FMR1 CCG CGG repeat in Fragile X tremor ataxia patients has also been reported [42] . Our data showing that relatively short ( ∼110 repeats ) CUGexp transcripts can cause dysregulation of MBNL1/CUGBP1 regulated pathways and the growing number of antisense transcripts recently reported in mammalian genomes [43] , highlight the need to look for CUG transcripts expressed at other loci traditionally associated with polyglutamine expansion disorders .
Fluorescent in-situ hybridization ( FISH ) and immunofluorescence ( IF ) was performed on frozen parasagittal cerebellar sections ( 6 µm ) as described [10] . Sections were fixed in 4% paraformaldehyde for 30 , permeabilized in 2% acetone for 5′ , incubated in 30% formamide/2XSSC pre-hybridization for 1 hr at RT and hybridized with a Cy5- ( CAG ) 10 for 2 hr at 42°C and post-hybridized at 45°C for 30′ . Sections were coverslipped and stained with DAPI to identify CUG RNA foci or incubated overnight at 4°C with Mbnl1 antibody ( polyclonal A2764 gift of C . A . Thornton ) at 1∶1000 and visualized by Alexa 488 secondary antibody ( 1∶2000 ) at RT for 30′ . A2764 is a polyclonal antibody , directed against a C-terminal peptide of Mbnl1 , specificity was demonstrated by lack of reactivity on immunoblots from Mbnl1ΔE3/ΔE3 mice and the antibody recognizes all known alternative splice isoforms of Mbnl1 [27] . Co-localization of Cy5- ( CAG ) 10 RNA foci with MBNL1 was done by confocal microscopy ( Olympus , Fluoview 1000 ) in 3 fluorescent channels with ≥5 layers ( 0 . 5 µm ) compressed along the Z-axis and then merged . Gabt4 immunofluorescence staining was done using 10 µm frozen sections fixed using 4% paraformaldehyde for 30′ , permeabilized with 2% acetone for 5′ , briefly washed with PBS and then blocked in 5% goat serum , 0 . 3% Triton X-100 in PBS for 2 hr at 4°C . Sections were incubated with anti-Gabt4 polyclonal antibody ( human GAT3 , Sigma , St . Louis , MO ) diluted 1∶1000 in the same blocking solution for 24 hr at 4°C , washed in PBS four times for 20′ and incubated with secondary antibody ( Alexa 488 goat anti mouse IgG , Molecular Probes , Eugene , OR ) using a 1∶2000 dilution for 2 hr at RT . Sections were washed 5 times for 5′ each in PBS and DAPI counterstained and coverslipped using Vectastain with DAPI ( Vectashield , Ca ) Rotarod training was performed at 26 weeks of age using an accelerating rotarod ( Ugo Basille , Comerio , Italy ) as described [3] . All mice tested were F1 littermate progeny of single copy integrant SCA8 BAC-EXP5+/− and Mbnl1+/ΔE3 mice . Four trials were run per day for four days: averages of the four trials on day four are presented . rmANOVA followed by post-hoc analysis ( Tukey's HSD ) was performed to assess differences in rotarod performance between groups ( SCA8 BAC-EXP5+/−; Mbnl1+/ΔE3; SCA8 BAC-EXP5+/−/Mbnl1+/ΔE3 and non-transgenic littermates ) . For RNA analysis , 15 µg of total RNA isolated from non-transgenic FVB or Bl6/129 wild type , SCA8 BAC-EXP1 , SCA8 BAC-CTRL and Mbnl1ΔE3/ΔE3 cerebellum was separated on a Northern Max-Gly glyoxal gel ( Ambion ) , transferred to a nitrocellulose membrane , cross-linked by ultraviolet radiation and hybridized at 65°C in Rapid-Hyb buffer ( Amersham ) using a [32P]dUTP in vitro transcribed RNA probe to nucleotides 576–996 ( NM_172890 . 3 ) of the mouse Gabt4 gene . Expression analysis was performed relative GAPDH by densitometry and analyzed by one-way ANOVA with Tukey's HSD post-hoc comparisons when necessary . Two step qRT-PCR was performed on an ABI Prism 7500 Real Time PCR System ( Applied Biosystems , Foster City , CA ) . Total RNA was isolated from mouse cerebellum , human frontal lobe autopsy tissue or transiently transfected SK-N-SH cells . cDNA was generated from 5 µg of total RNA using 1st Strand Synthesis Supermix primed with random hexamers ( Invitrogen ) . Relative qRT-PCR was performed on 1 µl of cDNA with qRT-PCR SYBER Green Master Mix UDG with ROX ( Invitrogen ) using mouse specific primers ( RSD1013 5′- CCT CTG AAG GCA TCA AGT TCT ATC TGT ACC-3′ ) ( RSD1014 5′-GTT GTT GTA ACT CCC CAG AGC GGT TAG-3′ ) or human specific primers for autopsy and SK-N-SH cells ( RSD1009 5′- AAC AAG GTG GAG TTC GTG CT-3′ ) ( RSD1010 5′- ACT TGT GAA CTG CCC CAG AG-3′ ) . Two stage PCR was performed for 40 cycles ( 95°C – 15″ , 60°C– 1′ ) in an optical 96 well plate with each sample cDNA/primer pair done in triplicate . Relative quantification compared to strain specific control was estimated using the threshold cycle ( Ct ) of GABT4/Gabt4 normalized to the Ct of the housekeeping gene Hprt or GAPDH for mouse and human , respectively . Dissociation curve analysis and ethidium bromide gel analysis was used to assess PCR product purity at the end of each qRT-PCR run . Statistical analysis was done using rmANOVA on the mean normalized Ct value of the 3 trials per sample and compared by Tukey's HSD post-hoc analysis for differences between groups when necessary . Animals were sacrificed and half of the cerebellum was rinsed with PBS and lysed in 450 µl of RIPA buffer ( 150 mM NaC1 , 1% sodium deoxycholate , 1% Triton X-100 , 50 mM Tris-HC1 pH 7 . 5 , 100 ug/ml PMSF ) for 45′ on ice . Cell lysates were centrifuged at 16 , 000×g for 15′ at 4 °C and the supernatant was collected . 20 µg of protein were separated on a 10% NuPAGE Bis-Tris gel ( Invitrogen ) , transferred to nitrocellulose membrane ( Amersham ) , blocked in 5% dry milk in PBS containing 0 . 05% Tween 20 and probed with an anti-GAT4 antibody ( human GAT3; Sigma; 1∶1000 ) or anti-GAT1 antibody ( ABcam; 1∶1 , 000 ) in blocking solution and then incubated with anti-rabbit or anti-mouse HRP conjugated secondary antibody ( Amersham ) . Mean fold increase in protein levels were determined by densitometry normalized to GAPDH and compared to non-transgenic littermates for each blot . Statistical significance was determined by one-way ANOVA with between group comparisons evaluated by Tukey's HSD when necessary . ExonA of containing the SCA8 expansion was amplified by PCR from the BAC transgene construct , BAC-exp , using the 5′ primer ( 5′CGAACCAAGCTTATCCCAATTCCTTGGCTAGACCC-3′ ) containing an added HindIII restriction site and the 3′ primer ( 5′ACCTGCTCTAGATAAATTCTTAAGTAAGAGATAAGC-3′ ) containing an added XbaI restriction site . This HindIII/XbaI fragment was cloned into the pcDNA3 . 1/myc-His A vector ( Invitrogen , CA ) . The SCA8 ExonA cDNA was placed under the control of the CMV promoter of plasmid pcDNA3 . 1/myc-His . To construct ExonA-Rev , the HindIII/XbaI fragment of SCA8 ExonA was subcloned in the reverse orientation into pcDNA3 . 1/myc-His vector . The pCTGexp and pCAGexp clones ( 108 repeats ) were generated by PCR amplification of SCA8 Exon A with added EcoRI restriction sites . This EcoRI/EcoRI fragment was cloned into the pcDNA3 . 1/myc-His A vector . The integrity of all constructs was confirmed by sequencing . Both pEGFP-N1-CUGBP1 and pEGFP-C1-MBNL1/41 have been described [27] , [44] , [45] . SK-N-SH cells were cultured in DMEM medium with 10% fetal bovine serum at 37°C with 5% CO2 . Transient transfections were performed using 1 µg of SCA8 repeat expressing plasmids , pEGFP-N1-CUGBP1 , or pEGFP-C1-MBNL1/41 minigenes and Lipofectamine 2000 Reagent ( Invitrogen , Carlsbad , CA ) . Cells were collected 48 hrs post-transfection for expression analysis . Analysis of NMDAR1 exon 5 and MBNL1 exon 7 alternative splicing were conducted using total RNA collected from SCA8 , DM1 and control human brain autopsy tissue using Trizol ( Invitrogen , CA ) reagent according to the manufactures procedures . Human NMDAR1 exon 5 splicing was determined by amplification of exon 4-5-6 using PCR primers hsGRIN1 ex4 For 5′- GCGTGTGGTTTGAGATGATG -3′; hsGRIN1 ex6 Rev5′-GGTCAAACTGCAGCACCTTC -3′ . Similarly , exon 7 alternative splicing of human MBNL1 was determined by amplification of exon 6-7-8 using PCR primers hsMBNL1 ex6 For 5′-GCTGCCCAATACCAGGTCAAC -3′; hsMBNL1 ex8 Rev 5′-TGGTGGGAGAAATGCTGTATGC -3′ . Determination of GABT4 exon 7 alternative splicing was done by reverse transcribing 2 µg total RNA from frontal lobe or SKN-S-H cells as described above . 10% of this reaction was subjected to PCR using primers ( RSD1004 5′-GTTGTATACGTGACTGCGACATT-3′; RSD1011 5′-GTTCAGGCAACAGAGCATGA-3′ ) to amplify nucleotides 791 – 1057 of human GAT3 ( NM_014229 . 1 ) for 25 cycles at 94°C-45″ , 54°C-30″ , 72°C-1′ followed by 72°C for 6′ . PCR products were run out on a 1 . 5% agarose gel with ethidium bromide and bands containing exons 6-7-8 ( 267 bp ) or 6–8 ( 163 bp ) were cut and verified by sequencing . The crosslinking and immunoprecipitation protocol ( CLIP ) was performed as described [28] , [46] with minor modifications . Hindbrains were dissected from mouse postnatal day 8 ( P8 ) C57BL/6J pups followed by dissociation in 1X Hank's balanced salt solution containing 10 mM HEPES , pH 7 . 3 and UV-irradiated to crosslink RNA-protein complexes . Cells were lysed and RNA was partially digested with RNase T1 to produce 30–200 nt fragments . Lysates were cleared by ultracentrifugation and Cugbp1-bound fragments were immunoprecipitated using mAb 3B1 and Protein G Dynabeads ( Invitrogen , Carlsbad , CA ) . Following 3′-end addition of RNA linkers , 5′ ends were labeled with g32P-ATP , protein-RNA complexes were eluted , separated by electrophoresis and protein-RNA complexes transferred to nitrocellulose . Bands corresponding to 60–70 kDa ( 10–20 kDa larger than the 50 kDa Cugbp1 protein ) were excised from the nitrocellulose and the RNAs released by proteinase K digestion . RNA fragments were size fractionated by denaturing PAGE followed by 5′-end RNA linker ligation , RT-PCR and DNA sequencing . All animal experimentation was approved by and conducted in conformity with the Institutional Animal Care and Use Committee of the University of Minnesota . Experimental details on the animal preparation , optical imaging and stimulation techniques are only briefly described as these have been provided in previous publications [32] , [33] . Mice ( 3–8 months old ) were anesthetized with urethane ( 2 . 0 mg/g body weight ) , mechanically ventilated , and body temperature feedback-regulated . The electrocardiogram was monitored to assess the depth of anesthesia . Crus I and II of the cerebellar cortex were exposed and the dura removed . An acrylic chamber was constructed around the exposed folia and superfused with normal Ringer's solution . The animal was placed in the stereotaxic frame on an X-Y stage mounted on a modified Nikon epifluorescence microscope fitted with a 4×objective and a 100 W mercury-xenon lamp . Images of Crus I and II were acquired with a Quantix cooled charge coupled device camera with 12 bit digitization ( Roper Scientific ) . The images were binned ( 2×2 ) to 256×256 pixels with a resolution of ∼10 µm×10 µm per pixel . Flavoprotein autofluorescence was monitored using a band pass excitation filter ( 455±35 nm ) , an extended reflectance dichroic mirror ( 500 nm ) , and a>515 nm long pass emission filter [33] . To evoke peripheral responses , the ipsilateral 3C vibrissal pad was stimulated with a bipolar electrode ( tips ∼1 mm apart ) using 20 V , 300 µs pulses at 10 Hz for 10 s [32] . Parallel fiber stimulation was performed throughout the experiment to test the general physiological condition of the cerebellar cortex . To activate parallel fibers and their postsynaptic targets ( Purkinje cells and interneurons ) , an epoxylite-coated tungsten microelectrode ( ∼5 M ) was placed just into the molecular layer [33] . The basic imaging paradigm consists of collecting a series of 10 control frames ( background ) followed by a series of 500 experimental frames with an exposure time of 200 ms for each frame ( Metamorph Imaging System , Universal Imaging Corp . ) . Whisker pad stimulation was initiated at the onset of the experimental frames . This was repeated 4 times and the 4 series were averaged . Each pixel in this series of average images was converted into the change in fluorescence above background ( ΔF/F ) [32] . The maximal response to whisker stimulation occurred in frames 31–75 and these frames were averaged to generate the response image . A pixel was defined as responding to the peripheral stimulation by the following threshold procedure . First , the response image was low-pass filtered ( 3×3 ) and the mean and standard deviation ( SD ) of the pixels in a control region ( usually a corner of the image ) were determined . The pixels above the mean + 1 SD of the control region were considered to respond to the stimulus . The response area was defined as total area of all the pixels responding and the response intensity as the sum of the DF/F of all responding pixels . Differences in the area and intensity of the responses between the SCA8 and FVB mice were evaluated using a Student's t-test ( α = 0 . 05 ) . For display the pixels above or below this mean±1 SD were pseudo-colored and superimposed on an image of the folia studied .
|
We describe several lines of evidence that RNA gain-of-function effects play a significant role in spinocerebellar ataxia type 8 ( SCA8 ) and has broader implications for understanding the CNS effects of other trinucleotide expansion disorders including myotonic dystrophy type 1 , Huntington disease like-2 , and spinocerebellar ataxia type 7 . The SCA8 mutation is bidirectionally transcribed resulting in the expression of CUGexp transcripts from ATXN8OS and CAGexp transcripts and polyglutamine protein from the overlapping ATXN8 gene . These data suggest that SCA8 pathogenesis involves toxic gain-of-function effects at the RNA ( CUGexp ) and/or protein ( PolyQ ) levels . We present three lines of evidence that CUGexp transcripts play a significant role in SCA8: 1 ) CUGexp transcripts accumulate as ribonuclear inclusions that co-localize with MBNL1 in selected neurons; 2 ) loss of Mbnl1 enhances motor deficits in SCA8 mice; 3 ) SCA8 CUGexp transcripts trigger alternative splicing changes and increased expression of the CUGBP1-MBNL1 regulated CNS target , GABA-A transporter 4 ( GAT4/Gabt4 ) which is associated with the predicted loss of GABAergic inhibition within the granular cell layer in SCA8 mice . Additionally , alternative splicing changes and GAT4 upregulation are induced by CUGexp but not CAGexp transcripts . From a therapeutic viewpoint , it is promising that this change is reversed in cells overexpressing MBNL1 .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience/behavioral",
"neuroscience",
"neuroscience/motor",
"systems",
"neurological",
"disorders/neuroimaging",
"neurological",
"disorders/movement",
"disorders",
"neurological",
"disorders",
"genetics",
"and",
"genomics/disease",
"models",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/gene",
"function",
"neurological",
"disorders/neurogenetics",
"neuroscience/neuronal",
"and",
"glial",
"cell",
"biology"
] |
2009
|
RNA Gain-of-Function in Spinocerebellar Ataxia Type 8
|
Plant volatiles play important roles in attraction of certain pollinators and in host location by herbivorous insects . Virus infection induces changes in plant volatile emission profiles , and this can make plants more attractive to insect herbivores , such as aphids , that act as viral vectors . However , it is unknown if virus-induced alterations in volatile production affect plant-pollinator interactions . We found that volatiles emitted by cucumber mosaic virus ( CMV ) -infected tomato ( Solanum lycopersicum ) and Arabidopsis thaliana plants altered the foraging behaviour of bumblebees ( Bombus terrestris ) . Virus-induced quantitative and qualitative changes in blends of volatile organic compounds emitted by tomato plants were identified by gas chromatography-coupled mass spectrometry . Experiments with a CMV mutant unable to express the 2b RNA silencing suppressor protein and with Arabidopsis silencing mutants implicate microRNAs in regulating emission of pollinator-perceivable volatiles . In tomato , CMV infection made plants emit volatiles attractive to bumblebees . Bumblebees pollinate tomato by ‘buzzing’ ( sonicating ) the flowers , which releases pollen and enhances self-fertilization and seed production as well as pollen export . Without buzz-pollination , CMV infection decreased seed yield , but when flowers of mock-inoculated and CMV-infected plants were buzz-pollinated , the increased seed yield for CMV-infected plants was similar to that for mock-inoculated plants . Increased pollinator preference can potentially increase plant reproductive success in two ways: i ) as female parents , by increasing the probability that ovules are fertilized; ii ) as male parents , by increasing pollen export . Mathematical modeling suggested that over a wide range of conditions in the wild , these increases to the number of offspring of infected susceptible plants resulting from increased pollinator preference could outweigh underlying strong selection pressures favoring pathogen resistance , allowing genes for disease susceptibility to persist in plant populations . We speculate that enhanced pollinator service for infected individuals in wild plant populations might provide mutual benefits to the virus and its susceptible hosts .
Insects pollinate many plant species , including several major crops [1] . Bees are the single most important insect pollinator group and can be a limiting factor for the success of plant reproduction [1–3] . Consequently , there is strong inter- and intra-specific competition among plants for the attention of pollinators [2 , 3] . With respect to insect-pollinated crops , pollinator visitation ( or artificial pollination ) is required to obtain maximal seed and fruit production [4 , 5] . Consequently , pollination facilitates higher yields even when a crop plant is self-compatible [4 , 5] . Tomato ( Solanum lycopersicum ) provides a good example of the relationship between pollination and yield . Bumblebees are important pollinators of tomato and other Solanum species that utilize an unusual pollination system called ‘buzz-pollination’ [6] . Buzz-pollinated flowers provide excess pollen as a reward to foraging bumblebees that feed it to their developing larvae [6] . Although domesticated tomato is to a large extent ‘self-fertilizing’ , buzz-pollination by bumblebees or by manual application of mechanical vibration ‘wands’ is required for maximal seed production , which in turn promotes increased fruit yield ( see [5] and references therein ) . Cucumber mosaic virus ( CMV ) , one of the major viral pathogens of tomato , is a positive-sense RNA virus that encodes five proteins including the 2b protein , which is a viral suppressor of RNA silencing ( VSR ) [7 , 8] . Bees do not transmit CMV but the virus is vectored by several aphid species [7 , 8] . Virus infection causes dramatic changes in plant host metabolism ( reviewed in [9] ) . CMV-induced metabolic changes include qualitative and quantitative alterations in the emission of volatile compounds and in certain host species this makes infected hosts more attractive to aphid vectors [10 , 11] . It is not known if the virus-induced alterations in host volatile emission that influence aphid behavior can also affect plant-pollinator interactions . Most bee-plant interaction studies have focussed on the effects of visual cues . Therefore , the influences of floral and non-floral volatiles on bee-mediated pollination are less well understood [12–14] . In contrast , the floral odors that attract moth pollinators have been more extensively researched [15–17] . In this study we determined that CMV infection induced changes in olfactory cues emitted by Arabidopsis thaliana ( hereafter referred to as Arabidopsis ) and tomato plants in ways that could be perceived by the bumblebee Bombus terrestris , and confirmed in tomato that this was associated with quantitative and qualitative changes in the blend of plant-emitted volatile organic compounds ( VOCs ) . We also elucidated a role for the host microRNA ( miRNA ) pathway in regulating the emission of bee-perceivable olfactory cues . Our data indicated that bumblebees possess an innate preference for olfactory signals emitted by CMV-infected tomato plants and we mathematically modeled what the possible wider implications of this might be if a similar preference occurred in wild host plants under natural conditions .
In ‘free-choice’ assays , bumblebees encountered flight arenas containing ten tomato plants ( five plants/treatment ) concealed within towers designed to allow odors to diffuse out but prevent the bees from seeing or touching the plants ( Fig 1A ) . Cups that were placed on top of towers hiding plants of both treatment groups offered bumblebees the identical ‘incentive’ of a 30% sucrose solution . Nonetheless , when presented with mock-inoculated and CMV-infected tomato plants , bumblebees preferred to visit the towers that were hiding infected plants ( Fig 1B ) ( S1 Table ) . Bumblebees showed similar preferences for flowering and non-flowering CMV-infected plants , indicating that leaves were the main source of attractive volatiles ( Fig 1B ) . Bumblebees also displayed a preference for CMV-infected tomato plants over plants infected with CMVΔ2b , a viral mutant lacking the gene for the 2b VSR ( Fig 1B ) , a factor that also influences CMV-plant-aphid interactions [18 , 19] . The results obtained in free-choice assays with tomato plants infected with CMVΔ2b suggested that the 2b protein , which is a VSR , may be exerting effects on the metabolism of plant volatiles by interfering with host small RNA pathways . The model plant Arabidopsis is the best higher plant system to use to investigate the effects of small RNA pathways . However , whilst Arabidopsis plants emit potentially pollinator-influencing volatiles , this species is not bee-pollinated [20] . Consistent with this , bumblebees showed no significant difference in preference for volatiles emitted by CMV-infected versus mock-inoculated Arabidopsis plants in free-choice assays ( Fig 1B ) . An alternative approach to investigate the ability of bees to recognise differences in olfactory or other stimuli is to set up a differential conditioning or ‘learning curve’ assay [14 , 21] . A differential conditioning assay can reveal whether bees can perceive cues that would not normally induce any behavioural responses and that could not be studied in free-choice assays . In our differential conditioning assays , cups on towers offered bumblebees either a 30% sucrose solution ‘reward’ for choosing one treatment group or a ‘punishment’ ( 0 . 12% quinine ) for choosing the other group [14 , 21] . Bumblebees cannot distinguish quinine from sucrose except by taste [22] . Thus , increasing frequency of visits to sucrose-offering towers over the course of an experiment indicated that bees have learned to use plant odor as a cue to identify and avoid drinking from cups placed on towers offering quinine solutions . In these assays , a steep learning curve shows that bumblebees can easily distinguish between two treatment groups , and indicates that the volatile blends are likely to be qualitatively and/or quantitatively very distinct , whereas less steep curves indicate that differences between blends are less marked , and that bees find it more difficult to learn to distinguish between them based on odor . An illustration of the power of this approach is shown in Fig 2 ( S2 Table ) . Although bumblebees displayed an innate preference for volatiles emitted by CMV-infected tomato plants in free choice assays ( Fig 1A ) , they could be trained by differential conditioning to overcome their innate preference and instead preferentially visit mock-inoculated tomato plants and avoid CMV-infected plants ( Fig 2A ) . Although we had observed that bumblebees had no innate preference for , or aversion to , volatiles emitted by Arabidopsis plants ( Fig 1B ) , differential conditioning assays revealed that the insects could recognize differences between volatiles emitted by Arabidopsis plants that had been mock-inoculated and by plants that were infected with CMV ( Fig 3A ) ( S2 Table ) . Bumblebees could also distinguish between CMV-infected and CMVΔ2b-infected Arabidopsis plants ( Fig 3B ) . Hence , although they exhibit no innate behavioural response to the volatile blends emitted by Arabidopsis plants , differential conditioning assays showed that bumblebees could perceive differences in volatiles emitted by these plants . This meant that differential conditioning assays could permit further dissection of the mechanisms underlying CMV-induced changes in volatile emission using Arabidopsis as a model system . Bumblebees could learn to differentiate transgenic plants constitutively expressing the 2b VSR from non-transgenic plants ( Fig 3C ) and from control-transgenic plants that were expressing an untranslatable 2b transcript ( Fig 3D ) . However , the insects displayed less ability to learn to distinguish mock-inoculated from CMVΔ2b-infected plants ( Fig 3E ) . Comparison of the learning curves in Fig 3A versus Fig 3E by logistic regression ( see Methods ) indicated that bumblebees were better at distinguishing mock-inoculated plants from CMV-infected plants than from CMVΔ2b-infected plants ( χ2 ( 1 ) = 40 . 17 , p < 0 . 0001 ) . Bees could not be trained to differentiate non-transgenic plants from control-transgenic plants expressing a non-translatable 2b transcript ( Fig 3F ) . The results with CMVΔ2b suggested that the 2b VSR plays an important role in altering the emission of bee-perceivable olfactory cues emitted by tomato and Arabidopsis plants ( Figs 1A and 3E ) . However , CMVΔ2b accumulates to lower levels in plants than wild-type CMV and in previous work it was found that viral titer , as well as the presence of the 2b protein , plays a role in modification of the interactions of Arabidopsis with aphids [19] . Hence , it was conceivable that differences in virus titer might affect the emission of bee-perceivable volatiles by plants infected by CMV or CMVΔ2b and explain why the bees found it difficult to distinguish CMVΔ2b-infected plants from mock-inoculated plants . However , it is known that CMVΔ2b accumulates to levels comparable to those of wild type CMV in Arabidopsis plants carrying mutations in the genes encoding the Dicer-like ( DCL ) endoribonucleases DCL2 and DCL4 , which are important factors in antiviral silencing [19] . Therefore , we examined the ability of bumblebees to learn to distinguish between volatile blends emitted by CMVΔ2b-infected and mock-inoculated dcl2/4 double mutant plants ( Fig 3G ) . The resulting learning curve ( Fig 3G ) was not significantly different from that obtained using wild-type plants that had been mock-inoculated or infected with CMVΔ2b ( Fig 3E ) ( χ2 ( 1 ) = 0 . 66 , p = 0 . 42 ) , indicating that an increase in CMVΔ2b titer did not enhance bee learning . Although we cannot rule out a role for other CMV gene products , the results indicate that the 2b VSR is the most significant viral factor conditioning changes in the emission of bee-perceivable volatiles . One of the host molecules that interact with the 2b VSR is the Argonaute 1 ( AGO1 ) ‘slicer’ protein . AGO1 is required for silencing directed both by short-interfering RNAs ( which can be generated de novo ) and by miRNAs , which are generated by a specific host endoribonuclease ( DCL1 ) from miRNA precursor transcripts encoded by nuclear genes [23 , 24] . In differential conditioning assays , bumblebees were able to learn to distinguish between volatiles emitted by wild-type plants versus those emitted by ago1 mutant plants ( Fig 3H ) and those emitted by dcl1 mutant Arabidopsis plants ( Fig 3I ) . However , bumblebees showed little or no ability to learn to distinguish between volatile blends emitted by ago1 and dcl1 mutant plants , indicating that the volatile blends emitted by plants of these two mutant lines were very similar ( Fig 3J ) . Thus , the miRNA-directed silencing pathway regulates the emission of bee-perceivable volatile compounds . Double mutant dcl2/4 plants are unable to generate CMV-derived short-interfering RNAs but are not affected in miRNA biogenesis . In CMV-infected dcl2/4 plants a higher proportion of the 2b protein is available to bind AGO1 and inhibit its miRNA-directed activity [19] , which is likely to enhance virus-induced changes in emission of bee-perceivable volatiles . In line with this , bumblebees were able to learn to distinguish between volatiles emitted by CMV-infected wild-type and dcl2/4 double mutant Arabidopsis plants ( Fig 3K ) . As an additional control we showed that bumblebees could not learn to distinguish between volatiles emitted by mock-inoculated plants covered by towers offering sucrose rewards or quinine punishments ( Fig 3L ) . The responses of bumblebees to CMV-infected tomato plants that were hidden from the insects indicated that changes in the emission of volatiles were affecting bee behavior and were responsible for the innate preference of these insects for CMV-infected plants ( Fig 1B ) . To confirm that CMV infection caused changes in the emission of VOCs , tomato plant headspace volatiles were collected and analysed by gas chromatography coupled to mass spectrometry ( GC-MS ) . VOCs were collected from non-flowering mock-inoculated plants , plants infected with CMV-Fny and plants infected with the 2b gene deletion mutant of CMV-Fny , CMVΔ2b . The emitted VOCs were distinct from each other when compared by principal component ( PC ) analysis on the relative intensity of ions ( over 75 Da in size ) within the samples ( Fig 4A ) . PC1 explained 80 . 3% of the variation and discriminated between samples from mock-inoculated and CMV-infected plants , whereas PC2 discriminated between samples from mock-inoculated and CMVΔ2b-infected plants ( Fig 4A ) . Thus , the VOC blend emitted by CMV-infected tomato plants was more distinct from that released by mock-inoculated plants than it was from the volatiles emitted by CMVΔ2b-infected plants . Nevertheless , VOC emission by CMVΔ2b-infected tomato plants was distinct from either mock-inoculated plants or CMV-infected plant VOC emission ( Fig 4A ) , despite this mutant virus accumulating to markedly lower levels than CMV ( S1 Fig ) . Although CMV-infected plants were smaller than either mock-inoculated or CMVΔ2b-infected plants , the emission of the combined volatiles on a whole plant basis was similar between mock-inoculated and CMV-infected plants ( Fig 4B ) . Indeed , expressing the emission of the combined VOCs by mass of tissue revealed that CMV-infected plants released greater quantities of volatiles compared to mock-inoculated and CMVΔ2b-infected plants ( Fig 4C ) . Thus , despite being stunted , CMV-infected plants generated a greater total quantity of VOC than either mock-inoculated or CMVΔ2b-infected tomato plants . Identification by GC-MS of the most abundant VOC by g dry weight of tomato plant tissue showed that terpenoids dominated the profile , with α-pinene , 2-carene , p-cymene , β-phellandrene and the sesquiterpene ( E ) -caryophyllene being apparent ( Fig 4D and 4E ) . CMV infection caused quantitative changes in the profile of these VOCs; α-pinene and p-cymene emission increased markedly , whereas 2-carene and β-phellandrene did not , and ( E ) -caryophyllene almost disappeared from the profile ( Fig 4E ) . Isomeric composition was not further determined than that stated here . When VOC emission was compared on a whole plant basis , α-pinene and p-cymene emission rates from CMV-infected plants appeared similar to mock-inoculated or CMVΔ2b-infected plants , while 2-carene and β-phellandrene levels appeared to be lower ( although this was not statistically significant in a one-way ANOVA: Fig 4D ) . Bumblebees of a closely related species ( B . impatiens ) are known to be repelled by β-phellandrene and 2-carene [25] . Thus , lower emission values of these VOCs from CMV-infected plants may explain why bumblebees displayed an innate preference for CMV-infected tomato plants over mock-inoculated plants in free choice assays ( Fig 1B ) . The VOC profiles of mock-inoculated and CMVΔ2b-infected plants were similar , although not identical ( Fig 4A ) , and this could explain the bees’ lack of preference in free choice assays ( Fig 1B ) . Domesticated tomato plants are often said to be self-fertilizing . However , optimal self-fertilization requires sonication of the flower to release pollen from the anthers onto the stigma , which is provided either by buzz-pollination from a bee ( typically a bumblebee ) or simulated buzz-pollination using mechanical vibration [5] . This is illustrated in Fig 5A , which shows how mechanical buzz-pollination of flowers increased seed production by around a third . Seed production by tomato was very dramatically decreased in plants infected with CMV-Fny to less than 10% of the yield in mock-inoculated plants ( Fig 5A ) . Remarkably , artificial buzz-pollination of flowers of CMV-infected plants rescued seed production to a significant degree with seed numbers reaching approximately half the level seen for non-buzzed flowers of mock-inoculated plants and about 6- to 7-fold greater than the number of seeds produced in non-buzzed , CMV-infected plants . The difference in seed yield between mock-inoculated and CMV-infected plants that had been vibrated was less marked than between non-buzzed , mock-inoculated and CMV-infected plants ( Fig 5A ) . Although CMV-infected plants produced fewer seeds , the mass of individual seeds was unaffected by infection and was not affected whether or not flowers were vibrated ( Fig 5B ) . Additionally , the number of flowers produced by CMV-infected plants was similar to the number produced by mock-inoculated plants , and tomato flower morphology was also not markedly altered by infection ( S2 Fig ) . Overall plant growth was stunted by CMV infection ( S2 Fig ) but , interestingly , virus infection appeared to accelerate the appearance of flowers by a small but statistically significant degree ( S2 Fig ) . A recent report indicated that flowers of squash ( Cucurbita pepo ) plants infected with the potyvirus zucchini yellow mosaic virus yielded decreased quantities of pollen [26] . However , we found no significant differences in the quantity or viability of pollen released from mock-inoculated and CMV-infected tomato flowers ( S3 Fig ) . We investigated the effects of CMV infection on bumblebee-mediated pollination under glasshouse conditions in which the insects could see and interact with flowers ( Fig 6 ) . A European CMV isolate , PV0187 , which is 99% identical in RNA sequence to CMV-Fny and which encodes a 2b VSR that is identical in amino acid sequence to that of CMV-Fny ( S4 Fig ) , was used for these experiments in order to comply with UK quarantine and containment regulations . CMV-PV0187 had similar effects on growth of tomato plants as CMV-Fny ( S5 Fig ) and volatiles emitted by tomato plants infected with CMV-PV0187 were attractive to bumblebees in the free choice assay ( Fig 1B ) . When CMV-infected and mock-inoculated tomato plants were exposed to bumblebees , a higher proportion of the insects made their initial floral visits to CMV-infected plants and spent longer sonicating the flowers of CMV-infected plants ( S6 Fig ) . As had been seen for artificial buzz-pollination ( Fig 5 ) , when bumblebees buzzed flowers , seed yield was increased ( Fig 6 ) . For CMV-infected plants , when bees did not visit flowers or where flowers were on plants not exposed to bees ( untouched plants ) , the seed yield was significantly decreased ( Fig 6 ) . However , although CMV infection decreased seed number in fruits derived from unvisited flowers , buzz-pollination by bumblebees negated this effect; indeed , bee-pollinated flowers on CMV-infected plants yielded fruit that contained seed numbers similar to those found in fruit that developed from bee-pollinated flowers on mock-inoculated plants ( Fig 6 ) . The results imply that there was greater buzzing activity on flowers of CMV-infected plants ( S6 Fig ) , resulting in a greater amount of seed production . We have seen that under controlled conditions CMV infection made tomato plants more attractive to bumblebees ( Fig 1B ) . We also found that although infected plants yielded fewer seeds , simulated buzz-pollination could to some extent rescue seed production ( Fig 5A ) and when bees were allowed access to CMV-infected plants they caused a greater increase in seed production per fruit compared to simulated buzz-pollination ( Fig 6 ) . The results obtained with this domesticated plant under controlled conditions prompted us to wonder what would be the consequences for a wild buzz-pollinated plant growing under natural conditions , if virus infection resulted in greater pollinator visitation and/or seed production and whether this might result in any benefits for the host plant or the virus . To investigate this idea further we developed a mathematical model to test whether increased pollinator service to virus-infected plants could allow genes for virus susceptibility to persist in a host plant population , despite the significant fitness cost of infection for plants as female ( seed producing ) parents ( cf . Figs 5 and 6 ) . Our model tracks the long-term dynamics of the interaction between resistant and susceptible phenotypes in a population of annual plants ( see also Materials and Methods ) . We focused on resistance as a dominantly inherited trait and attached no fitness penalty to the presence of resistance , which is a conservative approach given that recessive resistance is a commonly observed antiviral defense mechanism and that resistance may incur fitness costs in the absence of infection [27] . We assume infected susceptible plants produce fewer seeds , with the parameter δ controlling the proportionate number of viable seeds produced per fertilized ovary on a virus-infected plant . However , we also assume that , all other things being equal , an individual visit by a pollinator is ν times more likely to be to a flower on an infected versus an uninfected plant . This pollinator bias makes infected plants more likely to reproduce as both male ( pollen donor ) and female ( seed producing ) parents , potentially out-weighing the deleterious effect of infection on seed production . We focus initially on the trade-off between pollinator bias ( ν ) and reduction in seed production ( δ ) , for different levels of pollinator service ( which we parameterize via γ , the mean number of pollinator visits per flower over the plant’s reproductive season ) . In indicative examples of both low ( γ = 0 . 25 ) and high ( γ = 2 . 5 ) pollination regimes , ( ν , δ ) parameter space can be divided into three regions: resistance takes over in the long-term , susceptibility takes over in the long-term , or resistant and susceptible plants coexist ( Fig 7A and 7B ) . For both values of γ , at high values of ν and δ ( i . e . if infected plants are strongly preferred by pollinators but do not suffer a great reduction in seed production ) , then genes conferring susceptibility will take over in the plant population . For low values of ν and δ the situation is reversed , and resistance is favored . At intermediate values of ν and δ , resistant and susceptible plants coexist . For fixed baseline values of ν = 3 . 0 and δ = 0 . 5 , the proportion of susceptible alleles in the population first increases then decreases as the level of pollinator service ( γ ) is increased ( Fig 7C ) . At very low values of γ , although virus-infected plants benefit from additional pollinator service on both male and female sides , the vast majority of fertilizations do not involve pollinator visits ( instead being via self-pollination ) . The cost to susceptible plants of reduced seed production as female parents is therefore more important than increased pollinator visitation , and so virus resistance takes over . As γ is increased , the proportion of fertilizations caused by pollinators goes up , which allows the benefits to virus-infected plants on both male and female sides to outweigh the cost of infection , and so the genes for susceptibility are favored . As γ is increased still further , the benefit on the female side becomes smaller ( since pollinator visitation is not limiting and almost all ovules are fertilized ) , but on the male side proportionately more pollen still comes from infected plants . For these values of the parameters , alleles conferring virus susceptibility persist in the plant population , but at reduced density . The maximum density of susceptible genotype plants is therefore realised at intermediate pollinator densities . The broad pattern of a rise then fall in the proportion of plants carrying the susceptible allele is repeated for a range of values of the proportion of susceptible plants that are infected ( i . e . the parameter α in our model: Fig 7D ) . However , for our default parameterization at low levels of infection the eventual fall with increasing pollinator levels is not apparent , and susceptible plants exclude resistant plants even for very high values of γ . A full sensitivity scan around default parameter values ν = 3 . 0 , δ = 0 . 5 , γ = 1 . 0 , σ = 0 . 5 , φ = 0 . 75 and α = 0 . 5 , ( Fig 8; S7 Fig ) shows the behaviour of the model over large regions of parameter space . The susceptible genotype is able to persist under many combinations of parameters . Our model therefore suggests preferential visitation of infected plants by pollinators could in principle provide a robust mechanism allowing susceptible genotype plants to be retained in the host population for a wide range of conditions .
Infection with CMV altered the volatile profile of tomato plants and made them more attractive to bumblebees , indicating that these insects possess an innate preference for the blend of volatile compounds emitted by CMV-infected tomato . Although bumblebees showed no innate preference for CMV-infected or mock-inoculated Arabidopsis plants , differential conditioning experiments showed that bumblebees were able to perceive alterations in volatiles emitted by these plants . Experiments with the 2b gene deletion mutant virus , CMVΔ2b , in tomato and Arabidopsis , and with 2b-transgenic and ago1 and dcl1 mutant Arabidopsis plants , implicate small RNA pathways in the regulation of the production of bee-perceivable volatile compounds . The inability of bees to learn to effectively distinguish between volatiles emitted by ago1 and dcl1 mutant plants causes us to conclude that miRNAs are the predominant class of small RNAs involved in regulating the metabolism of bee-perceivable compounds . The rationale for this conclusion is that AGO1 , a target for the CMV 2b VSR , utilizes both short-interfering RNAs and miRNAs to guide RNA cleavage , while DCL1 is involved in miRNA biogenesis but is not involved in production of short-interfering RNAs ( see refs . [23 , 24] and references therein ) . As far as we are aware , an effect of miRNAs on plant volatile production ( presumably through regulation of stability or translation of specific plant mRNAs ) has not been previously reported . The work also indicates that olfactory signals emitted by non-floral tissue may have a more important effect than previously thought in plant-bee interactions and may play roles in bee attraction , presumably at longer ranges than visual features such as the optical or tactile qualities of flowers . Thus , foliar volatile signals may affect bee choices or synergize with and reinforce visual floral cues , as has been seen with floral odors [28 , 29] . How do changes in the output of volatiles increase the attractiveness of CMV-infected plants for bumblebees ? Much of the existing bee perception literature is focused on the effects of visual stimuli ( e . g . color and other optical properties of flowers [14] ) , whereas the effects of olfactory stimuli have been relatively neglected . But it is known , for example , that the VOC output from flowers decreases after they have been pollinated [12] . Pollination can also trigger qualitative changes in the volatile blend . For instance , following pollination by bees , blueberry ( Vaccinium corymbosum ) flowers emit an increased proportion of their volatiles as ( E ) -caryophyllene [30] . It is thought that decreased volatile emission by pollinated flowers decreases their saliency to bees and prevents damage from over-visitation [12] and a similar explanation was offered by Rodriguez-Saona and colleagues [30] to explain the post-visit increase in ( E ) -caryophyllene emission . In the case of tomato plants infected with CMV , it may be that the virus is both ‘turning up the volume’ of plant volatile emission ( making these plants more apparent to the bumblebees ) whilst ‘tuning’ volatile blend composition so as to diminish levels of a signal ( ( E ) -caryophyllene ) , that at higher levels might indicate a previous bee visitation , and levels of β-phellandrene and 2-carene that might discourage visitation [25] . When the bumblebees were allowed access to flowering tomato plants under glasshouse conditions we found that buzz-pollination by bumblebees was more effective at enhancing seed yield on CMV-infected plants . This result suggests that additional foliar or floral cues , for example visual or tactile stimuli , do not negate the effects on the bees of CMV-induced changes in volatile emission . It is possible that our findings may have implications for transmission of viruses vectored by bees . However , pollinators transmit very few viruses and CMV is not one of them ( discussed on page 522 in reference [31] ) . Nevertheless , is it possible that a virus that is not bee transmitted gains some advantage by re-paying a susceptible host by altering its volatile cues to attract pollinators ? In our mathematical model it was assumed that a hypothetical population of wild plants included some hosts that possessed genetic resistance to the virus . It might then be assumed that pathogen-imposed selection pressure would favor the takeover of the plant population by any plants possessing one or more resistance genes . This outcome , causing a decrease in the population , or even the extinction , of susceptible plants would clearly not be beneficial either for the pathogen or for the susceptible hosts . However , our mathematical model shows that in the case where pollinators show increased bias towards pathogen-infected plants , the increased reproductive success of infected plants means that the outcome might be different . Thus , it is plausible that if the attractiveness of infected plants to pollinators is increased , this might inhibit or negate the selective advantage of resistant individuals and prevent them from taking over the population ( represented conceptually in Fig 9 ) . We also noted that CMV infection accelerated the appearance of flowers in tomato . If such an effect occurred in a wild plant population , it is conceivable that this may give infected , susceptible plants a further advantage over resistant or uninfected plants in the competition for limited pollinator services . Indeed , there are examples in which earlier flowering increases pollination and enhances yield ( for example in the oil crop plant Echium plantagineum ) [32] . However , the relationship between flowering time and pollination is complex and there may be environments in which it is more advantageous for plants to flower in a concerted fashion . However , in certain contexts earlier flowering may result in flowers being produced before pollinators are available ( reviewed in [33] ) . At this stage , it may be imprudent and premature to propose that increased pollinator attraction to infected , susceptible hosts represents some sort of specific viral strategy to inhibit selection for resistance , and there are difficulties in envisaging how this might initially arise . However , it seems plausible to suggest that in principle increased pollinator attraction to virus-infected plants could favor the persistence of susceptible plants in the environment and this could be seen as payback or compensation to the host . It is worth noting that other forms of payback by viruses to their hosts have been observed in a number of systems . This has led to the suggestion that our general view of viruses has been overly colored by their pathogenic properties and that we should view them as symbionts in the classical sense ( viz . on a spectrum that ranges from parasitic to mutualistic [34] ) . For plant viruses it has been shown that virus infection can enhance the endurance of susceptible host plants to drought or in one case to cold [35 , 36] and that plants of several species were protected from herbivory by virus infection [37–40] . It has been suggested that resistance to drought is a conditional phenotype that could act as a payback to the host . In the case of CMV-induced drought resistance in Arabidopsis and other plants [35 , 36] and in the present study , in which CMV enhances a tomato plant’s attractiveness to bumblebees , we may be seeing examples of ‘extended phenotypes’ . An extended phenotype emerges from the action of a parasite gene when it alters the phenotype of a host; potentially to the benefit of the parasite [41] . In both examples , drought resistance in Arabidopsis [36] and pollinator attraction in tomato ( the present study ) , the parasite gene controlling these extended phenotypes is the CMV 2b gene . A potential result of these extended phenotypes would be to increase the odds of continued survival of susceptible host plant populations , which would be beneficial to both host and pathogen . Our mathematical modeling results indicated that , for the areas of the parameter space that are most salient to our experimental findings , the most likely outcome of long-term selection would be coexistence of resistant and susceptible genotypes , i . e . the long-term maintenance of R gene polymorphisms . Several mechanisms have been proposed that could contribute to the maintenance of balanced R gene polymorphisms such as the ratio of costs versus benefits of resistance , and diffuse interactions between hosts and attackers [27 , 42 , 43] . Our data suggest that the enhanced attraction of pollinators to infected susceptible plants might add to these mechanisms and contribute to the long-term maintenance of R gene polymorphisms in insect-pollinated species . Production of many important crops depends on bee-facilitated pollination . Worryingly , bee populations are endangered by disease , environmental change [44 , 45] and , more controversially , by anthropogenic factors [46] . To mitigate the ensuing loss of pollination activity requires among other things a deeper understanding of the mechanisms shaping bee-plant interactions . Our data show that non-floral plant volatiles can be perceived by bumblebees and affect their behaviour and that emission by plants of bee-perceivable compounds is regulated in part by miRNA activity . This information may be useful in developing strategies to increase pollinator services for crops under conditions of cultivation , as well as for a better understanding of the interplay of plant pathogens , wild plants and pollinators under natural conditions .
Plants used were Arabidopsis thaliana ( Heynh . ) accession Col-0 and Solanum lycopersicum ( L . ) cv . Moneymaker ( Suttons Seeds Ltd . , Paignton , UK ) . Plants were grown in a growth chamber at 22°C in M3 compost ( Levingtons Ltd . , Northampton , UK ) . Tomato and Arabidopsis plants were grown under 16hr light/8hr dark and 8hr light/16hr dark photoperiods , respectively . Fny-CMV [47] , Fny-CMVΔ2b [48] , the 2b-transgenic plant line 2 . 30F [49] , and the dcl1-9 , dcl2/4 , and ago1-25 mutant plant lines have been described elsewhere [19 , 50 , 51] . CMV isolate PV0187 was obtained from the German Collection of Microorganisms and Cell Cultures ( DSMZ , www . dsmz . de ) . RNAs1 , 2 and 3 of CMV isolate PV0187 were sequenced and submitted to GenBank under accession numbers KP165580 , KP165581 , and KP165582 , respectively . Inoculations were carried out at the seedling stage and were performed as described previously [49] . Plants were used in experiments when the virus had spread systemically and infection was confirmed routinely by double-antibody sandwich enzyme-linked immunosorbent assays ( BioReba , Reinach , Switzerland ) . Quantification of CMV and CMVΔ2b RNA accumulation was carried out as previously described [52] . Leaf tissue from systemically infected leaves was harvested at 10 and 18 dpi . Total RNA for reverse transcription coupled polymerase chain reaction analysis was extracted using an RNeasy Plant Kit ( Qiagen ) and treated with TURBO-DNase ( Ambion ) according to the manufacturers’ instructions . Reverse transcription was carried out with 0 . 5 μg total RNA using Goscript ( Promega ) with random hexamer primers according to the manufacturer's instructions . Following the reaction , cDNA was diluted 1/10 for subsequent use . Semi-quantitative PCR was performed using Biomix Red ( Bioline ) and products were separated electrophoretically on a 1 . 5% agarose gel . Reverse transcription coupled to quantitative polymerase chain reaction analysis was performed using SYBR Green JumpStart Taq ReadyMix ( Sigma ) in 15 μl reactions according to the manufacturer's instructions . Reactions were performed in triplicate . Primers described in [52] were designed against the conserved 3’ non-translated regions of the CMV genomic RNAs and the stable transcript elongation factor 1 alpha ( EF1α ) was used as the reference RNA . Data were analyzed using LinRegPCR to give Ct values . Relative viral RNA accumulation was calculated using ΔΔCt methodology , incorporating the EF1α transcript to control for variation in loading [53] . Bombus terrestris ( L . ) colonies ( obtained from Syngenta-Bioline , Leicester , UK and Koppert Biological Systems , Berkel en Rodenrijs , The Netherlands ) were connected by gated transparent tubing to flight arenas with the dimensions 72 x 104 x 30 cm [22] containing 11 cm tall feeding towers ( to conceal plants ) formed from black card sitting within ‘Aracon’ bases ( Lehle , Roundrock , TX ) , roofed by plastic mesh supporting a microcentrifuge tube lid ( Fig 1A ) containing sucrose solution . Tower height was selected because bumblebees cannot effectively resolve visual cues beyond 10 cm [54] . Seven days prior to carrying out conditioning or free choice assays bees were allowed to feed on sucrose solution from cups placed on empty towers for three days to familiarize them with the arena . Foraging bees were marked on the thorax with water-soluble paint and used once . Initially , cups on towers offered 30% sucrose , conditioning bees to associate towers with a reward . For differential conditioning and free-choice experiments , five plants per treatment group were individually covered by towers . For differential conditioning experiments , towers hiding plants from one treatment group provided 0 . 3 ml quinine hemisulfate ( 0 . 12% ) , whilst the others offered 0 . 3 ml of 30% sucrose . Individual foraging bumblebees were released into the arena and allowed to forage until satiated . Aborts following landing or hovering over towers offering quinine or drinking on towers offering sucrose were scored as correct choices . Between foraging bouts , towers were re-arranged randomly to inhibit spatial learning and meshes cleaned ( 30% ethanol ) to remove scent marks . One hundred choices for each bee tested for each pair-wise comparison were recorded . In free-choice preference assays towers covering plants from both treatment groups offered equal sucrose rewards and only the first feeding choice was recorded . The learning curve data were analysed using binomial logistic regression [55] . The experimental protocol did not record individual choices made by the bees , but instead the number of ‘correct’ choices made by each bee was grouped into sets of 10 successive choices for ease of scoring . Exploratory analyses suggested no pronounced differences between individual bees within treatment groups , and so we fitted the following fixed effect model to these data bij~Bin ( 10 , pi ) , log ( pi1−pi ) =α0+α1 ( i−0 . 5 ) , where bij is the number of correct choices made by the jth bee in its ith set of ten choices , pi is the probability of choosing correctly in each successive batch of ten choices , and where α0 and α1 are the parameters to be estimated . We used Hosmer-Lemeshow tests to assess model goodness-of-fit [56]: in all cases there was no evidence for lack-of-fit . We therefore went on to assess whether the parameter α1 was different to zero via a likelihood ratio test against the simpler nested model with α1 fixed to be zero [57] . Since the parameter α1 controls how the ( logit ) of the probability of making a correct choice pi increases with i , positive values of α1 correspond to the bees ‘learning’ over time . Any systematic differences in the rate at which bees learn between pairs of experiments was assessed by simultaneously fitting a single regression model to the results of both experiments , allowing the probabilities of making a correct choice to depend on the experiment via log ( pi ( E ) 1−pi ( E ) ) =α0+ ( α1+α2E ) ( i−0 . 5 ) , in which E is an indicator variable which is equal to zero for the first experiment , and equal to one in the second experiment . A value α2 ≠ 0 corresponds to bees learning at a different rate in the different experiments: again , this was tested via a likelihood ratio test against the simpler nested model in which α2 was fixed to be zero . Artificial buzz-pollination was carried out using an electrically actuated toothbrush ( ‘Oral-B’: Proctor and Gamble , Cincinnati , USA ) . Mean seed mass was obtained by dividing the mass of seeds by the total seed number for a total of five fruits per plant , with three plants per treatment group . Pollen viability was assessed by staining with fluorescein diacetate [58] and pollen grains viewed under blue light and bright field using an epi-fluorescent microscope ( DMRXA , Zeiss ) connected to a digital camera ( DFC425 , Zeiss ) . For bumblebee pollination experiments two-week-old tomato seedlings were inoculated with CMV ( isolate PV0187 ) or mock-inoculated and grown in a controlled environment room for 4 weeks . At this time , the plants began flowering and were transferred to a glasshouse . Two weeks later single bumblebees ( released from a small flight arena ) were allowed to buzz pollinate flowers on three mock-inoculated and three CMV-infected tomato plants within a larger flight arena ( 125 x 370 x 90cm , H x W x D ) constructed from nylon netting ( S8 Fig ) ( JoTech-Insectopia Ltd . , Austrey , UK ) . Two inflorescences of two to three flowers per plant were left accessible to the bee ( any more inflorescences were covered with a paper bag ) . When each bee had made 10 visits to flowers ( or had ceased pollinating ) , any buzz-pollinated flowers were labeled with a jeweler’s tag and all plants that had been visited by the bee were removed from the arena and replaced with another . A new bee was then released from the small arena into the larger arena containing plants . In total , 8 bees freely pollinated flowers from 17 mock-inoculated and 14 CMV-infected tomato plants . Bumblebee visitation to mock-inoculated versus CMV-infected plants was noted and , using a stopwatch , the duration of flower sonication was recorded for each bee . The plants were left in the greenhouse for a further 8 weeks to allow fruits to develop . Further flower development on the plants was permitted . To release seeds , fruits were harvested individually into 60 ml screw-cap pots and left to ferment for 1–2 weeks before washing and counting . Fruits were either from flowers that were not buzz-pollinated by a bumblebee ( fruit from flowers not visited by bee ) or from flowers that were buzz-pollinated ( fruit from bee-pollinated flowers ) . A further category of fruit was from flowers that were not buzz-pollinated , but were adjacent to fruit from buzz-pollinated flowers ( fruit from flowers adjacent to bee-pollinated flowers ) . Fruits were also harvested from eight mock-inoculated and eight CMV-infected plants that were not exposed to bees in the flight arena , but had otherwise experienced the same growth conditions as the plants used in the bee pollination experiment ( fruit from untouched plants ) . Headspace volatiles were collected from tomato plants ( 4 weeks-old ) by dynamic headspace trapping over a period of 24 hours onto Porapak Q filters [50 mg , 60/80 mesh size , Supelco ( Sigma-Aldrich ) ] as described by Beale and colleagues [59] . The tomato plants were contained in a 1 . 0 liter bell jar clamped to two semi-circular metal plates with a hole in the center to accommodate the stem . Charcoal-filtered air was pumped in at the bottom of the container at a rate of 750 ml . min-1 and drawn out through the Porapak Q filter at the top , at a rate of 700 ml . min-1 . Leaf fresh weight and dry weight were measured to enable normalization of the volatile abundance . Trapped organic chemicals were eluted from the Porapak Q filter with diethyl ether for analysis by gas chromatography coupled to mass spectrometry ( GC-MS ) . For initial investigation of volatiles by principal component analysis , volatiles were separated on a capillary GC column ( TG-SQC , 15 m by 0 . 25mm; film thickness , Thermo Scientific , UK ) . The injection volume ( splitless ) was 1μl , the injector temperature was 200°C , and helium was used as the carrier gas at a constant flow rate of 2 . 6 ml min−1 in an oven maintained at 30°C for 5 minutes and then programmed at 15°C . min-1 to 230°C . The column was directly coupled to a mass spectrometer ( ISQ LT , Thermo Scientific , UK ) with a MS transfer line temperature of 240°C . Ionization was by electron impact with an ion source temperature of 250°C in positive ionization . Mass ions were detected between 30 and 650 m/z . Data were collected using Xcalibur software ( Thermo Scientific ) . Principal component analysis on the mass spectra was performed with MetaboAnalyst 2 . 0 [60] using binned m/z and per cent total ion count ( %TIC ) values . Confirmation of identities of specific organic compounds comprising the blends emitted by mock-inoculated and virus-infected plants was carried out by re-analysis of trapped organic compounds using a Thermo-Finnigan Trace GC directly coupled to a mass spectrometer ( MAT-95 XP , Thermo-Finnigan , Bremen , Germany ) equipped with a cold on-column injector . Two microliters of collected volatiles were separated on an HP1 capillary gas chromatography column ( 50 m x 0 . 32 mm I . D . ) in an oven maintained at 30°C for 5 min and then programmed at 5°C . min-1 to 250°C [61] . The carrier gas was helium . Ionization was by electron impact at 70 eV at 220°C . Compounds were identified by comparison of spectra with mass spectral databases ( National Institute of Standards and Technology: http://www . nist . gov/ ) , as well as by co-injection with authentic standards on a Hewlett-Packard 6890 gas chromatograph with two different columns of different polarity ( HP1 and DB-WAX ) . Our model tracks the interaction over evolutionary time between virus resistant and virus susceptible phenotypes in a population of diploid annual plants . The plant population size is assumed to be large and to remain constant over generations . Since CMV is a broad host-range pathogen , we can reasonably make the simplifying assumption that within-generation pathogen prevalence is not affected by the density of resistance in the focal host plant species . The proportion of susceptible plants that become virus infected in each generation is therefore held constant as a parameter ( α ) in our model . We model resistance as controlled by a single bi-allelic locus , with resistant ( R ) and susceptible ( r ) forms , and we assume R is dominant . We assume infected plants produce fewer seeds , with the parameter δ controlling the proportionate number of viable seeds produced per ovary on a virus-infected plant . We additionally assume that virus resistance carries no fitness penalty when compared to uninfected susceptible hosts . If the reduction in seed number were the only consequence of virus infection , resistance would certainly fix in the plant population under such a conservative assumption on the cost of virus resistance for the plant . However , we also assume that increased attractiveness to pollinators means infected plants are more likely to reproduce , as both male ( pollen donor ) and female ( seed producing ) parents . In particular , we assume the pollinator density remains constant over generations , and that this pollinator density leads to an average of γ pollinator visits per flower averaged over all plants over the entire reproductive season . We assume that flowers visited by pollinators will certainly be pollinated: by cross-pollination ( proportion φ ) or by self-pollination ( proportion 1 – φ ) . Self-pollination after a visit by a pollinator can be due to either geitonogamous pollen transfer from flowers on the same plant , or via autogamous buzz-pollination ( cf . Figs 5 and 6 ) . A proportion σ of the remaining ovules in flowers that are not visited by pollinators also go on to self-pollinate . The potential selective benefit to virus-infected plants is caused by pollinator preference . We assume that an individual pollinator is ν times more likely to visit a flower on an infected vs . an uninfected plant than would be expected by chance alone . This potentially increases female ( seed producing ) fitness by making ovules on infected plants more likely to be fertilized , and male ( pollen donor ) fitness by increasing rates of pollen transfer from infected plants . Given these assumptions , our model tracks the proportion of the plant population in generation n with genotype RR , Rr or rr , which we denote by xn , yn and zn , respectively . The equations linking populations over generations are xn+1=ζn ( ϵR ( xn+yn4 ) +κR ( βRR+βRr2 ) ( xn+yn2 ) ) , yn+1=ζn ( ϵRyn2+κR ( βRr2+βrr ) xn+κRyn2+κr ( βRR+βRr2 ) zn ) , zn+1=ζn ( ϵRyn4+ϵrzn+ ( βRr2+βrr ) ( κRyn2+κrzn ) ) , in which η=11+ ( ν−1 ) αzn , βRR=ηxn , βRr=ηyn , βrr=η ( 1+ ( ν−1 ) α ) zn , ω−=1−e−γη , ω+=1−e−νγη , θ−= ( 1−ω− ) σ+ω− ( 1−ϕ ) , μ−=ω−ϕ , θ+= ( 1−ω+ ) σ+ω+ ( 1−ϕ ) , μ+=ω+ϕ , ϵR=θ− , κR=μ− , ϵr=αδθ++ ( 1−α ) θ− , κr=αδμ++ ( 1−α ) μ− and where ζn is chosen in each generation to ensure xn+1 + yn+1 + zn+1 = 1 . A full derivation of the model showing how it follows from the underlying assumptions is given in S1 Text . The majority of the results presented in the main text are relative to our default parameterization of the model . By default we take the following parameter values: ν = 3 . 0 , δ = 0 . 5 , γ = 1 . 0 , σ = 0 . 25 , φ = 0 . 75 and = 0 . 75 . However , as described above , we perform a full two-way sensitivity analysis of pairs of parameters around these default values ( Fig 8 ) to test the robustness of our results to our choice of parameterization . The behaviour of the model can most easily be characterised in terms of which genotypes persist in the long-term . This classification follows from a stability analysis of the susceptible-free ( i . e . xn = 1 , yn = zn = 0 ) and resistance-free ( i . e . xn = yn = 0 , zn = 1 ) equilibria . Since we are working in discrete time , an equilibrium is stable if the magnitude of the largest Eigenvalue of the Jacobian matrix evaluated at the equilibrium is less than unity [62] . If neither equilibrium is stable then both susceptible and resistant plants are able to invade a population consisting almost exclusively of the other when rare , and so the genotypes are predicted to coexist . If only the susceptible-free equilibrium is stable , then resistance dominates . If only the resistance-free equilibrium is stable , then susceptibility dominates . But if both equilibria are stable , then the long term outcome depends on the initial densities of each genotype . Extensive numerical simulations of the model were performed to verify that local stability analyses could be used to infer the long-term outcome for all initial conditions . In particular we tested 10 , 000 combinations of parameters and initial conditions ( 1 , 000 sets of randomly-chosen parameters , each simulated starting from 10 independent and randomly-selected sets of initial conditions ) . In all cases the outcome after 10 , 000 generations of the model matched that predicted by the ( purely local ) stability analysis described above . We also performed a number of individual tests for pairs of sets of parameters chosen to cross stability boundaries: the stability analysis predicted behaviour in full simulations of the model in the large number of cases we tested .
|
Cucumber mosaic virus , an important pathogen of tomato , causes plants to emit volatile chemicals that attract bumblebees . Bumblebees are important tomato pollinators , but do not transmit this virus . We propose that under natural conditions , helping host reproduction by encouraging bee visitation might represent a ‘payback’ by the virus to susceptible hosts . Although tomato flowers can give rise to seed through self-fertilization , bumblebee-mediated ‘buzz-pollination’ enhances this , increasing the number of seeds produced per fruit . Buzz-pollination further favors reproductive success of a plant by facilitating pollen export . Mathematical modeling suggests that if self-fertilization by infected plants , as well as pollen transfer from these plants ( cross-fertilization ) to surrounding plants is increased , this might favor reproduction of susceptible over that of resistant plants . This raises the possibility that under natural conditions some viruses might enhance competitiveness of susceptible plants and inhibit the emergence of resistant plant strains . We speculate that it may be in a virus’ interest to pay back a susceptible host by enhancing its attractiveness to pollinators , which will likely increase fertilization rates and the dissemination of susceptible plant pollen and may compensate for a decreased yield of seeds on the virus-infected plants .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"invertebrates",
"plant",
"anatomy",
"chemical",
"compounds",
"brassica",
"animals",
"organic",
"compounds",
"plant",
"science",
"model",
"organisms",
"crops",
"volatile",
"organic",
"compounds",
"plants",
"flowering",
"plants",
"bees",
"arabidopsis",
"thaliana",
"research",
"and",
"analysis",
"methods",
"hymenoptera",
"crop",
"science",
"chemistry",
"tomatoes",
"seeds",
"insects",
"fruits",
"agriculture",
"arthropoda",
"bumblebees",
"flowers",
"plant",
"and",
"algal",
"models",
"organic",
"chemistry",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2016
|
Virus Infection of Plants Alters Pollinator Preference: A Payback for Susceptible Hosts?
|
To assess the burden of neurocysticercosis ( NCC ) in California we examined statewide hospital discharge data for 2009 . There were 304 cases hospitalized with NCC identified ( incidence = 0 . 8 per 100 , 000 ) . Cases were mostly Latino ( 84 . 9% ) , slightly more likely to be male than female ( men 57 . 6% , women 42 . 4% ) with an average age of 43 . 5 years . A majority of cases were hospitalized in Southern California ( 72 . 1% ) and many were hospitalized in Los Angeles County ( 44 . 7% ) . Men were more likely than women to have severe disease including hydrocephalus ( 29 . 7% vs . 18 . 6% , p = 0 . 027 ) , resulting in longer hospitalizations ( >4 days , 48 . 0% vs . 32 . 6% , p = 0 . 007 ) that were more costly ( charge>$40 thousand men = 46 . 9% vs . woman = 4 . 1% , p = 0 . 026 ) . Six deaths were recorded ( 2 . 0% ) . The total of NCC-related hospital charges exceeded $17 million; estimated hospital costs exceeded $5 million . Neurocysticercosis causes appreciable disease and exacts a considerable economic burden in California .
Neurocysticercosis ( NCC ) is associated with severe disease morbidity and mortality across the globe , including the United States where its impact has been studied primarily in the states of California [1] , [2] , [3] , [4] , Texas [5] and Oregon [6] . Though NCC mortality is rare in the United States ( 3 . 9 per million population ) , nearly 60% of U . S . deaths occurs in California ( 126/221 deaths over 12 year study period ) [7] . This parasitic disease is preventable , causing premature death globally and has been identified by the WHO as a potentially eradicable disease . Taenia solium eggs in the feces of a tapeworm carrier ( taeniasis ) are the source of the infection . By identifying and treating taeniasis tapeworm carriers , the risk for exposure can be eliminated . Taeniasis infection is acquired through the consumption of undercooked pork containing the larval form of Taenia cysts . NCC is not reportable in most jurisdictions and data on the burden of this parasitic disease in the United States are lacking . Moreover , few population-based data sources are available . A recent published article on neglected infections of poverty has drawn attention to the need for additional data for NCC [8] . In addition , the economic impact of NCC can be sizable , with the average charge of a NCC hospitalization in Los Angeles County ( $66 thousand ) [1] considerably more costly than the average hospital charge in the U . S . in 2008 ( $29 thousand ) [9] . The clinical presentation of NCC generally takes on two forms depending on the location of the cerebral lesions [10] . Lesions appearing in the parenchymal area of the brain are characterized by seizures and associated with cerebral edema . Occasionally , these cases may develop intracranial hypertension which can be lethal without interventions such as a decompression craniotomy to allow the brain to swell . However , parenchymal NCC is generally a more benign form of NCC . In comparison , lesions forming in the extraparenchymal area of the brain often lead to a more serious form of the disease due to the obstruction and accumulation of cerebrospinal fluid in the ventricles , or cavities , of the brain causing hydrocephalus . This condition may also lead to intracranial hypertension and can be lethal without intervention such as a shunting procedure to remove excess fluid . Several studies have pointed out that some individuals identified with NCC lesions have minimal inflammatory response to these lesions and little or no symptoms of infection , indicating that the degree of immune response and severity of illness may vary considerably by individual . An earlier study of parenchymal NCC in Mexico indicated that women with this infection present more frequently with severe illness than men [11] . Another study in Mexico also identified that females with parenchymal NCC present with more evidence of focal edema around cysticerci in CT scans as compared to men with the same infection , suggesting more severe illness among females and potentially some difference in immune response by gender to this organism [12] . In addition , this same study found little difference in the immune response by gender when reviewing extraparenchymal NCC cases , suggesting similar levels of illness severity by gender for extraparenchymal NCC . We examined the distribution and burden of NCC in California using hospital discharge data and explored the incidence of NCC hospitalizations , demographics of those hospitalized , and factors leading to lengthy and costly hospitalizations . We also reviewed the demographics of those presenting with diagnosis suggestive of a parenchymal NCC infection as well as extraparenchymal NCC infection . Common procedures performed on NCC hospitalizations were also reviewed and the total NCC-related hospitalization charges were computed . In addition , we also describe new methodology to identify individual cases from a de-identified hospital discharge dataset that potentially contains multiple visits of the same case .
We defined a NCC case hospitalized in California as a resident having a hospital discharge diagnosis of cysticercosis ( ICD9 123 . 1 ) as a primary or additional diagnosis ( 1–24 ) AND also diagnosed with one of the following 1 ) hydrocephalus , seizures ( including epileptic convulsions ) , cerebral edema or cerebral cyst . We used the 2009 California Office of Statewide Health Planning and Development ( OSHPD ) annual hospitalization discharge dataset for this analysis . The dataset included demographic information , primary diagnosis , up to 24 additional diagnoses and as many as 20 procedures performed while hospitalized . Also included in this dataset are length of each hospitalization and resulting hospitalization charge . This is a public use dataset and identifiers such as names have been removed . We initially extracted data on all hospitalizations with a discharge diagnosis of cysticercosis as a primary or additional diagnosis ( 1–20 ) ( ICD9 = 123 . 1 , N = 805 ) from the 2009 OSHPD data set . Because this dataset contained multiple records in instances where the same case was hospitalized more than once , a sub-set was created containing one record for each case hospitalized for cysticercosis . When possible , this is done by reviewing the record linkage number ( RLN; based on a coded social security number ) for each visit and eliminating duplicates . Unfortunately , many of the RLN numbers for the cysticercosis hospitalization dataset created were missing ( 31 . 6% , n = 254 ) . This may be due to the fact that cysticercosis primarily impacts an immigrant population in which many cases may not have social security numbers . To overcome this problem , an algorithm was created to assign a linkage number ( RLN2 ) to hospital visits lacking an RLN number . This new number was constructed from demographic information available in the dataset such as age ( +/−1 year ) , race , gender and home zip code . This algorithm was tested against cysticercosis hospitalizations having an actual RLN number ( n = 551 ) to assess its ability to capture duplicate records from the same case . The algorithm was successful in correctly assigning a hospitalization as either a single visit or an additional visit for 94 . 9% of hospitalizations; 94 . 7% of single hospitalizations were identified correctly and 95 . 3% of duplicate hospitalizations were identified correctly . We then assigned this newly constructed linkage number ( RLN2 ) to hospitalizations previously missing a linkage number , which allowed us to create a dataset with one record for each case hospitalized , even in instances where multiple hospitalizations had been recorded ( N = 670 ) . Finally , we applied the NCC case definition described earlier to screen out any possible miscoding or incidental findings . The resulting dataset containing 304 individual NCC cases hospitalized in 2009 in CA was used for the study analysis . The diagnostic and procedural information from cases with multiple visits was consolidated into one entry to facilitate analysis of these variables at the individual level . Incidence rates were calculated using population estimates from the California Department of Finance . SAS version 9 . 1 software was used for the analysis . Demographic differences in clinical illness and lengthy and costly hospitalizations were analyzed using a nonparametric chi-square testing . A lengthy hospitalization was defined as having hospitalization for more than 4 days and a costly hospital stay was defined as having a hospital charge greater than $40 , 000 . We also analyzed the demographics of those presenting with diagnosis suggestive of a parenchymal NCC infection ( seizures ) as well as extraparenchymal NCC infection ( hydrocephalus ) using non-parametric chi-square testing . The demographics used in these analyses included gender ( male , female ) , ethnicity ( Latino , non-Latino ) , age ( 0–19 , 20–39 , 40–59 , 60 or more years ) and language spoken ( Spanish , non-Spanish ) . Common procedures performed on NCC cases were also reviewed as well as computation of the total NCC-related hospitalization charges .
In 2009 , there were 805 hospitalizations in California listing cysticercosis as a discharge diagnosis . There were 670 individual cysticercosis cases that made up these hospitalizations . Of these , there were 304 hospitalized persons that met the NCC case definition ( 0 . 8 per 100 , 000 persons ) . By comparison , 8 years earlier ( 2001 ) there were 792 total cysticercosis hospitalizations , 632 individual cysticercosis cases hospitalized and 386 persons hospitalized with NCC ( incidence 1 . 1 per 100 , 000 persons ) . Of the 304 NCC cases hospitalized in 2009 , there were 113 cases ( 37 . 2% ) discharged with a primary diagnosis of cysticercosis . Cases hospitalized for NCC were primarily Latino ( 84 . 9% ) and those who primarily speak Spanish ( 56 . 9% ) ( Table 1 ) . The prevalence of Spanish language preference was similar by gender ( men = 58 . 3% , women = 55 . 0% ) and by age ( greater than 40 years 50 . 9% vs . 49 . 1% ) . Those that were age 20 to 39 years had the highest incidence of hospitalization of the age groups reviewed ( 1 . 3 per 100 , 000 ) . There were 13 cases 19 years of age or younger; nine were Hispanic ( 69 . 2% ) and 7 were male ( 53 . 9% ) . Among all NCC cases , the rate of hospitalization was slightly higher for men as compared to women ( 0 . 9 and 0 . 7 per 100 , 000 respectively ) . The average age of hospitalization ( 43 . 5 years ) was similar by gender ( men = 42 . 6 years , women = 44 . 7 years ) . In comparison , the average age of a NCC case hospitalized in 2001 was 38 . 0 years . Many of the NCC cases hospitalized in 2009 had multiple hospitalizations within the year reviewed ( 20 . 1% ) . The majority of NCC cases in California were hospitalized in the Southern California region ( 72 . 4% , 0 . 9 per 100 , 000 ) which included the counties of Los Angeles ( 44 . 7% , 1 . 3 per 100 , 000 ) , Orange ( 8 . 5% , 0 . 8 per 100 , 000 ) and San Diego ( 7 . 9% , 0 . 8 per 100 , 000 ) . Many of the NCC cases hospitalized in Northern California occurred in the counties of Santa Clara ( 4 . 6% , 0 . 8 per 100 , 000 ) and Alameda ( 2 . 6% , 0 . 5 per 100 , 000 ) . Additional diagnoses identified among NCC cases included seizures ( 74 . 3% ) , hydrocephalus ( 25 . 0% ) , cerebral cyst ( 7 . 6% ) or cerebral edema ( 6 . 6% ) . Some cases were diagnosed with both hydrocephalus and seizures ( 7 . 5% ) . Many patients received a neurological procedure while hospitalized ( 41 . 0% ) which included cranial procedures ( 25 . 0% ) such as a ventricular shunt procedure ( 14 . 8% ) or the excision [or destruction of] a brain lesion ( 8 . 2% ) . Some of these shunt procedures involved the removal and replacement of a ventricular shunt ( 6 . 3% ) . The average length of a hospital stay was 6 . 5 days and the average hospital charge was $57 . 8 thousand dollars . Six deaths occurred during the period of hospitalization ( 2 . 0% ) . Deaths occurred among both men ( n = 3 ) and women ( n = 3 ) and all were Latino . Ages of deaths ranged from 33–71 years ( median 50 . 5 years ) . One death had hydrocephalus and had received a shunt procedure . One death had cerebral edema and 4 others had diagnosis of seizure . Other health conditions frequently listed among NCC cases include diabetes ( 15 . 5% ) and heart disease ( 12 . 8% ) . One case was also diagnosed with taeniasis . Overall , 41 . 5% of NCC cases had lengthy hospital stays ( >4 days ) . Men were more likely to have a lengthy stay as compared to women ( men = 48 . 0% vs . woman = 32 . 6 , p = 0 . 007 , Figure 1 ) , but age and race ethnicity were not associated with lengthy stays , nor was language spoken . NCC cases with hydrocephalus were more likely to have a lengthy stay as compared to cases without hydrocephalus ( 63 . 2% vs . 34 . 2% , p<0 . 001 ) , but cases with seizures were not associated with a lengthy stay ( 37 . 2% vs . 53 . 9% ) , nor were those with cerebral edema ( 35 . 0% vs . 41 . 9% ) . Chronic health conditions such as heart disease were associated with a lengthy hospital stay ( 59 . 0% vs . 38 . 9% , p = 0 . 017 ) . Of interest , heart disease was no more likely to be diagnosed among men than women with NCC ( men = 10 . 9% , women = 15 . 5% ) . Other chronic health conditions such as diabetes were not associated with a lengthy hospital stay ( 46 . 8% vs . 40 . 5% ) . Multiple hospitalizations , another measure of health care utilization , was not associated with gender ( men = 22 . 3% , woman = 17 . 2% , Figure 1 ) or age , race ethnicity or language spoken . The percent of NCC cases having a costly hospitalization ( charges >$40 , 000 ) was 41 . 4% . By gender , men were also more likely than woman to have had a costly hospitalization ( 46 . 9% vs . 34 . 1% , p = 0 . 026 , Figure 1 ) , but other demographics such as age , ethnicity and language spoken were not associated with a costly hospitalization . NCC cases with hydrocephalus were associated with a costly hospitalization as compared to those without hydrocephalus ( 68 . 4% vs . 32 . 5% , p <0 . 001 ) , however cases with seizures were not associated with a costly hospitalization ( 32 . 3% vs . 68 . 0% ) , nor were those with cerebral edema ( 40 . 1 vs . 60 . 0% ) . NCC cases with heart disease were also associated with a costly hospitalization ( 59 . 0% vs . 38 . 9% , p = 0 . 002 ) , however those with diabetes were not associated with a costly hospitalization ( 46 . 8% vs . 40 . 5% ) . NCC cases with hydrocephalus were associated with male gender ( men = 29 . 7% , women = 18 . 6% , p = 0 . 027 , Figure 1 ) and associated with having an age greater than 20 years [no cases under the age of 20 years identified] . Aside from this finding , hydrocephalus was not associated with ethnicity , language spoken or other age groups . Among those diagnosed with hydrocephalus ( n = 76 ) , men were more likely to be associated with a lengthy hospitalization ( >4 days ) than women ( men = 71 . 2% vs . women = 45 . 8% , p = 0 . 03 ) . However , there was no association with costly hospitalizations ( >$40 , 000 ) among NCC cases with hydrocephalus by gender ( men = 71 . 2% , women = 62 . 5% ) . Age , race ethnicity and language spoken were not associated with lengthy stays among NCC cases with hydrocephalus . NCC cases with seizures were not associated by gender , age , ethnicity or language spoken . By gender specifically , seizures appeared at very similar rates ( men = 73 . 7% , women = 75 . 2% , Figure 1 ) . Among NCC cases with seizures ( n = 226 ) , men appeared to have a more lengthy hospital stay ( 44 . 2% vs . 27 . 8% , p = 0 . 011 ) , but this finding was not significant when removing NCC cases who also had hydrocephalus from the analysis ( n = 209 , 39 . 1% vs . 28 . 7% ) . Among NCC cases with seizures ( n = 226 ) , a costly hospitalization was not associated with gender ( men = 36 . 4% vs . women = 26 . 8% ) or by age , race ethnicity or language spoken . NCC cases with cerebral edema ( n = 20 ) , also a diagnosis suggestive of a parenchymal infection , was not associated by gender ( men = 55 . 0% vs . women = 45 . 0% , Figure 1 ) or age , ethnicity or language spoken . NCC cases with cerebral edema were not associated with more lengthy hospital stays ( 35 . 0% vs . 41 . 9% ) or more costly hospital stays ( 60 . 0% vs . 40 . 1% ) . Cranial procedures performed on hospitalized NCC cases was not associated with gender ( men = 28 . 6% vs . women = 20 . 2% , Figure 1 ) , age , ethnicity or language spoken . Shunting procedures were not associated by gender ( men = 17 . 1% vs . women = 11 . 6% ) age , ethnicity or language spoken . No shunting procedure was performed on NCC cases less than 20 years of age . Shunting procedures were almost exclusively performed on cases with hydrocephalus ( 97 . 8% ) . A shunting procedure was associated with having multiple hospitalizations ( 29 . 5% vs . 11 . 1% , p<0 . 001 ) . NCC cases having a brain lesion extraction were not associated by gender ( men = 8 . 6% , women = 7 . 8% ) age , ethnicity or language spoken . NCC cases requiring a brain lesion extraction ( n = 25 ) involved cases with hydrocephalus ( 44 . 0% ) , seizures ( 52 . 0% ) and cerebral edema ( 8 . 0% ) , conditions not mutually exclusive . In summary , the demographic that appeared the most consistently associated with illness severity in this analysis was male gender . A summary of severity indicators such as diagnosis , procedures and hospital utilization are shown in Figure 1 .
Our findings support the notion that NCC remains a significant problem in California , especially in Los Angeles County ( LAC ) , and the burden of severe disease and economic costs are considerable . The number of hospitalized cases reported here ( n = 304 ) far exceeds the number reported to the California State Health Department ( n = 32 ) for 2009 [13] and illustrates the considerable under-reporting of NCC . The incidence of NCC hospitalization in 2008 in California is comparable to that found in 2001 , indicating that the burden of disease has not changed much in the last 8 years in California . The largest majority of cases were among Latinos from Southern California counties . However , NCC hospitalizations were observed across racial ethnic groups and in 29 California counties . The proportion of hospitalized NCC cases that were identified as Latino in this study ( 84 . 9% ) was slightly lower than that identified in another study of NCC mortality in California ( 92 . 7% ) [2] . The average age of a NCC case hospitalized in California for 2009 ( 43 . 5 years ) was older than that seen among hospitalized NCC cases in 2001 ( 38 . 0 years ) and also much older than that found in earlier studies of hospitalized cases in LAC . One study of hospitalized NCC cases in four hospitals in LAC ( 1973–83 ) identified an average age of 31 . 1 years [3] . A study of cysticercosis deaths in California ( 1989–2000 ) identified a mean age of 34 . 5 years [2] . The number of repeat hospitalizations per year identified in this study is also larger than that found in another more detailed study of hospitalized NCC cases in a single Texas hospital ( 20% vs . 14% ) [5] . These findings underscore the chronic and debilitating nature of this disease . The results of our study may indicate a shift in the demographics of the disease from a younger age , more acute illness , to older age , more chronic illness . It may also indicate that the incidence of NCC in endemic regions where cases are migrating from may be declining [14] . Our study found that the severity of NCC in California was associated with male gender . Men were slightly more likely to be hospitalized for NCC ( 58% ) . Other studies of reported cases have also shown similar results; 54% male ( LAC 1973–83 ) [3] and 58% male ( LAC 1981–88 ) [7] . Our study revealed that men with NCC have longer and more costly hospitalizations as compared to women hospitalized with the same infection . In addition , men were found to have more severe disease that includes hydrocephalus , suggestive of a higher incidence of extraparenchymal infection among men . These findings are consistent with the finding of another study of NCC mortality in California indicates that men have a mortality rate nearly twice that of women in CA ( 5 . 2/106 vs . 2 . 7/106 ) [2] . This study also found that 31% of deaths occur outside of a hospital setting and would not be captured in our analysis . However , our findings differ from an earlier study involving a Mexican study population suggesting that the inflammatory response to extraparenchymal NCC infection does not differs by gender , implying that severity of extraparenchymal NCC should not differ by gender [12] . This earlier study also suggests that the inflammatory response to parenchymal NCC is more severe among women , where our study found little evidence to support differences in severity of parenchymal infection by gender . Some of these discrepancies may be the result of very different study populations . The study group in Mexico [12] was more likely comprised of NCC cases with locally acquired infection whereas our study population is more likely to be comprised of immigrants with exposure outside of the U . S . The more urgent economic necessity for men to migrate to California for employment may explain some of the difference in the study findings . This difference in NCC severity by gender does not appear to be the result of gender differences in other chronic health conditions such as heart disease or diabetes that could results in more lengthy and costly hospitalizations . Could it be that men delay seeking medical attention until their condition is more severe ? However , if this were true , we might expect the average age of men hospitalized for NCC to be older than women , which was not the case . Is it due to a language barrier in understanding the management of this complicated disease ? Is it that men receive less care , or opt for less care once hospitalized ? Our findings suggest that this is also not the case . It may also be possible that the difference in disease severity by gender identified in our study is due to a difference in immune response to NCC as suggested in this earlier study [12] . This is the first study that documents a method of managing a de-identified public use hospital discharge dataset that is missing a large number of RLN's . This is especially important for analyzing severe chronic diseases such as NCC where persons may have multiple visits within the same year . As was identified in an earlier study [1] , the economic impact of NCC is sizeable . We found that the total charges associated with NCC hospitalizations in California exceeding $17 . 1 million annually . The average charge for a NCC hospitalization in California for 2009 ( $57 . 8 thousand ) was significantly higher than that of the average hospitalization charge in the U . S . ( $30 . 6 thousand ) [9] . Based on the average charge to cost to ratio for a hospitalization nationally in 2009 ( 3 . 0∶1 , n = 39 , 434 , 956 ) [9] , the estimated total cost of NCC hospitalizations in California would be $5 . 1 million . The estimated average cost of a NCC hospitalization in California in 2009 would be $17 . 3 thousand . The average length of a NCC hospitalization in 2009 ( 6 . 5 days ) was also considerably longer than that of the average hospitalization in the U . S . ( 4 . 6 days ) [12] . The additional costs , beyond hospitalization , of emergency room visits , outpatient care , disability and days of work lost all contribute to the overall financial impact of NCC . Any conclusions based on hospital discharge data may be limited for several reasons . It is possible that persons without an RLN ( SSI ) number may be more likely to relocate and have a change in zip code within a given year . This may cause multiple hospitalizations by the same cases to be interpreted as separate persons hospitalized when using the described method . Also , cases listed here as having only one hospitalization may have other hospitalizations in the previous year that would not be captured due to the selected study period . In addition , hospitalizations of California residents outside of California would not be captured by our data . The hospital charges presented here may overestimate the actual cost of hospitalization , as the actual amount received for hospital services may be less than the amount charged . However , we feel that the calculation of hospital charges can be useful when used in comparison to an average hospital charge in the U . S . and to hospital charges for other specific diseases . It would be helpful to know how much transmission is occurring in the US; however , this dataset does not have any information on the case's country of origin and how long cases have resided in the US . Thus there is no way to know if any of these infections were acquired locally verses their country of origin or through possible exposure when traveling . There is no information in the dataset regarding the type of laboratory diagnostics used to make these diagnoses , making it difficult to know how confident we can be in these diagnoses . It is also possible that a patient may have been clinically diagnosed with cysticercosis , but were not identified as such in the ICD9 discharge codes . Finally , while we identified the number of incident hospitalizations for cysticercosis in this study , a distinction should be made between incident hospitalizations and incident cases . This dataset does not allow us to determine whether the hospitalization represents a newly diagnosed case or a previously identified case . In Conclusion , neurocysticercosis exacts a significant disease and economic toll in California , especially in Los Angeles County . Hospitalized NCC cases are primarily working age Latinos . As compared to women , men are hospitalized with more severe disease and more costly and longer hospital stays . The reason for this difference in severity by gender is unclear , but potentially gender should be considered by physicians for proper disease management . NCC is a preventable disease and public health efforts to identify and treat the tapeworm carrier could be improved and adopted by other public health systems . Increased reporting from hospitals and clinicians in California will allow for public health follow-up of cases , identification and treatment of T . solium tapeworm carriers , prevention of additional cases , and reduction of the disease burden in California .
|
Neurocysticercosis ( NCC ) is considered one of the major neglected infections of poverty in the United States , with mortality studies indicating that California bears the highest burden of this disease . Although NCC is a reportable disease in California , studies indicate that this disease goes largely under-reported , contributing to the lack of information about the disease distribution and burden . In this manuscript , we reviewed the distribution of NCC hospitalizations in California , demographics of those hospitalized and total hospital-related charges for 2009 . This study revealed that a majority of persons hospitalized with NCC in California receive their medical service in Southern California hospitals , primarily in the County of Los Angeles . As compared to women hospitalized for this disease , men had a longer and more costly hospitalization with more severe symptoms such as hydrocephalus , a diagnosis suggestive of extraparenchymal infection . The reasons for this difference in NCC severity by gender are not clear , but do not appear to be due to delay in seeking medical care or a language barrier . The intensity of hospital care needed to manage these cases and the sizable NCC hospitalization charge underscores the considerable economic burden this disease presents in California .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"mathematical",
"computing",
"mathematics",
"neurological",
"disorders",
"neurology",
"mathematical",
"economics"
] |
2012
|
The Impact of Neurocysticercosis in California: A Review of Hospitalized Cases
|
In China , dengue remains an important public health issue with expanded areas and increased incidence recently . Accurate and timely forecasts of dengue incidence in China are still lacking . We aimed to use the state-of-the-art machine learning algorithms to develop an accurate predictive model of dengue . Weekly dengue cases , Baidu search queries and climate factors ( mean temperature , relative humidity and rainfall ) during 2011–2014 in Guangdong were gathered . A dengue search index was constructed for developing the predictive models in combination with climate factors . The observed year and week were also included in the models to control for the long-term trend and seasonality . Several machine learning algorithms , including the support vector regression ( SVR ) algorithm , step-down linear regression model , gradient boosted regression tree algorithm ( GBM ) , negative binomial regression model ( NBM ) , least absolute shrinkage and selection operator ( LASSO ) linear regression model and generalized additive model ( GAM ) , were used as candidate models to predict dengue incidence . Performance and goodness of fit of the models were assessed using the root-mean-square error ( RMSE ) and R-squared measures . The residuals of the models were examined using the autocorrelation and partial autocorrelation function analyses to check the validity of the models . The models were further validated using dengue surveillance data from five other provinces . The epidemics during the last 12 weeks and the peak of the 2014 large outbreak were accurately forecasted by the SVR model selected by a cross-validation technique . Moreover , the SVR model had the consistently smallest prediction error rates for tracking the dynamics of dengue and forecasting the outbreaks in other areas in China . The proposed SVR model achieved a superior performance in comparison with other forecasting techniques assessed in this study . The findings can help the government and community respond early to dengue epidemics .
Dengue is a serious infectious disease and remains rampant across tropical and subtropical regions [1] . Primary dengue infection in humans often leads to a variety of clinical symptoms , from mild fever to potentially fatal dengue shock syndrome , and effective antiviral agents capable of treating dengue infection are not available at present [1] . Aedes mosquitoes , including Aedes aegypti and Aedes albopictus , serve as the main transmission vector of dengue viruses [2] . The impacts of variability in climate conditions such as temperature and precipitation on development rates and habitat availability for Aedes aegypti and Aedes albopictus larvae and pupae have been identified [3] . By affecting agent development and transmission vector dynamics , climate factors influences the spread of dengue . According to a recent analysis of the global distribution and burden of dengue virus , the number of dengue infections per year is estimated to be 390 million , of which nearly 96 million are symptomatic [4] . The estimated number of dengue infections has sharply increased over the past 50 years , resulting in a huge impact on human health around the world . In China , dengue is a notifiable disease , and in recent years the area affected by dengue has expanded and the incidence has steadily increased [5] . According to the China Center for Disease Control and Prevention ( CDC ) , the range of dengue incidence is from 0 . 0091 to 3 . 4581 per 100 , 000 people , with a total of 52 , 749 new cases of dengue having been reported during 2009–2014 [6] . In particular , a succession of dengue outbreaks occurred in several provinces including Guangdong , Yunnan , Fujian , and Guangxi during 2014 ( S1 Fig ) [6] . All of these provinces are located close to Southeast Asian countries including Laos , Vietnam , Thailand , Singapore and Malaysia , where dengue has been hyperendemic for decades and poses a large burden of disease [7–10] . However , dengue is still characterized as an imported disease in China due to localized transmission sparked by regular virus importations from returned travelers or visitors , rather than endemic transmission [5] . Guangdong , the most developed province located in southern China , experienced an unprecedented outbreak in 2014 , and the number of cases reached the highest level over the past 25 years [5] . Our previous study showed that most of indigenous dengue cases occurred in the autumn of 2014 , and the Pearl River Delta Region accounted for the majority of cases [11] . In addition to this remarkable spatial heterogeneity of cases , we observed a wide temporal variation of weekly dengue incidence ranging from 0 to 9 , 660 cases , which makes predicting dengue incidence difficult [11] . In the absence of an effective vaccine against dengue in China , accurate and early forecasts of dengue epidemics might allow for more effective targeting of control measures for the government . Since 2008 , the China CDC has introduced the China Infectious Disease Automated-alert and Response System ( CIDARS ) , which uses a time series moving percentile method based on historical data , for detecting dengue outbreaks in China [12] . This traditional method is overly dependent on the numbers of the routine surveillance data [12] . However , routine surveillance data is typically available with a 1- to 2-week lag [13] . Recently , several studies have explored the application of internet search terms to timely monitor disease outbreak and verify the usefulness and effectiveness of the approach [13–16] . The idea of applying internet search query data may contribute to enhancing predictability for dengue in Guangdong where dengue poses a great temporal cycling of incidence . For dengue surveillance , several attempts have been made to develop robust predictive models for dengue incidence worldwide . Althouse et al . comprehensively assessed three regression models including step-down linear regression , gradient boosted regression tree model ( GBM ) and negative binomial regression model ( NBM ) for dengue incidence prediction in Singapore , and suggested the linear model selected by AIC step-down was superior to other models compared [16] . A more recent study achieved good performance by applying the least absolute shrinkage and selection operator ( LASSO ) algorithm to develop a real-time model to forecast dengue in Singapore [17] . In addition , generalized additive models ( GAMs ) were also used as valuable tools of risk assessment for dengue dynamics in previous studies [18 , 19] . Furthermore , as a kind of the state-of-the-art and powerful machine learning algorithm , support vector regression ( SVR ) [20] displayed excellent performances in time series prediction . However , thorough comparisons of different predictive models and thus identifying an optimal model in China are still lacking . We aimed to construct an accurate forecast model to track the epidemic trajectory of dengue by comparing different prediction algorithms . This work addressed the gap by a ) rigorously evaluating predictive performance of a variety of state-of-the-art algorithms using different assessment strategies and determining the optimal model , and b ) combining dengue surveillance data , meteorological and internet query information with the proposed model for dengue incidence prediction in China .
Temporal characteristics of dengue cases , DSI , mean temperature , rainfall and relative humidity for each city in Guangdong province during 2011–2014 are presented in S3–S12 Figs . There was a sharp increase in dengue cases in the autumn of 2014 for each city . In particular , the Pearl River Delta cities had the most obvious increase in the number of the notified dengue cases in September and October , and most areas in Guangdong have hotter temperatures and more rain during the summer season . The fluctuating trend in DSI was fairly consistent with the epidemic activity of dengue . In 2014 , Guangdong accounted for about 96 . 3% of all notified dengue cases nationwide ( S1 Fig ) . Spatiotemporal dynamics of dengue infections and DSIs during 2011–2014 in Guangdong is presented in Fig 1 . Most of the dengue cases occurred in the Pearl River Delta region of Guangdong , especially for Guangzhou , Foshan , Zhongshan , Zhuhai and Shenzhen ( Fig 1A ) . There was a close correlation between the number of dengue cases and the DSI in Guangdong ( Fig 1 and S13 Fig ) . The relative predictive accuracy of dengue incidence and goodness-of-fit assessment for each model are shown in Table 1 . The standardized RMSE and R-squared values for each city in Guangdong are shown in Fig 2 . According to the model performance for the two prediction periods , the SVR model had the smallest RMSE values , irrespective of city . The results suggested that the SVR model outperformed other compared models and was chosen as the optimal model in this study . Results of goodness-of-fit suggested that the discrepancy between observed incidence and the incidence expected under the SVR model was smallest . Forecasts of the SVR model for the last 12 weeks and the outbreak period of dengue incidence in 2014 , including 95% prediction intervals , for Foshan are presented in Fig 3 . The epidemic during the last 12 weeks and the peak of the large 2014 outbreak were accurately forecasted by the SVR model . SVR model forecasts for the other four cities including Guangzhou , Zhongshan , Zhuhai and Shenzhen with a high risk of dengue infection are displayed in S14–S17 Figs , respectively . The ACF and PACF plots revealed that there was no autocorrelation in the residuals from the SVR approach established , and thus ensured the validity of the models ( Fig 3 and S18 Fig ) . SVR algorithm consistently yielded the smallest prediction error rates for all the studied cities among the models compared , supporting the use of SVR to perform the forecasts . Additionally , the forecast accuracy of the SVR model increased as the value of parameter C got larger , and then quickly converged to a stable level , indicating the model had a good stability predictive ability ( S19 Fig ) . Predictions of dengue incidence in 2014 using an out-of-sample forecasting approach ( 1-week-ahead prediction for each forecast window ) for the best fitted SVR model are shown in Fig 4 . We observed an outstanding performance of the SVR model for detecting the peak of the large 2014 outbreak for the cities with a high risk of dengue infection ( Fig 4A ) . Dynamic forecasts of dengue incidence for the five cities are presented in S1–S5 Videos . The estimated map of dengue incidence in 2014 for Guangdong province by the SVR model well described the truly epidemic proportions of this disease ( Fig 4B ) . The ACF and PACF plots of the residuals from the fitted SVR models also revealed that there was no any autocorrelation in the residuals and the models had captured the patterns in the data quite well ( S20 Fig ) . To further validate the established models , we used dengue data from five other provinces , Yunnan , Guangxi , Hunan , Fujian and Zhejiang ( S1 Fig ) , with a high risk of dengue infection in southern China . There was a high correlation between the epidemic activity of dengue infection and the trend in DSI in these areas ( Fig 5A–5F ) . The assessment of predictions for single observations that were left out of the data set used to fit the model is presented in Fig 6 . The results demonstrated a more competitive prediction by the SVR model relative to the other models , because the RMSE values of the SVR model were consistently smallest for the 1-month-ahead predictions in 2014 , irrespective of the region investigated ( Fig 6 ) . The proposed SVR model had satisfactory prediction performance with large R-squared values for Yunnan ( R-squared = 0 . 976 ) , Guangxi ( R-squared = 0 . 970 ) , Hunan ( R-squared = 0 . 997 ) , Fujian ( R-squared = 0 . 981 ) and Zhejiang ( R-squared = 0 . 985 ) ( Fig 6 ) . It shows that the SVR model is a practical method to predict dengue dynamics in the five provinces .
This study demonstrates an efficient tool using a SVR algorithm to predict dengue outbreaks and track the epidemic trajectory in China . To the best of our knowledge , it is the first attempt to thoroughly evaluate the state-of-the-art algorithms for dengue prediction , and identify an optimal model that may help to complement the traditional surveillance for dengue dynamics . Located in southern China , Guangdong has a subtropical humid monsoon climate and has frequent economic and cultural communication with the nations of Southeast Asia where dengue poses a great burden of disease . The climate , combined with Guangdong’s highly urbanized environment , favors the presence of Aedes mosquitoes and the transmission of dengue virus , thus making the area highly vulnerable to dengue outbreaks . In the absence of an effective vaccine against dengue in China , conducting a rapid survey on mosquito vector density and suppressing the vector population comprise the core of dengue-control programs at present [38] . Though a community-based integrated intervention strategy has been carried out to control dengue outbreaks in Guangdong [39] , it is still important to enhance the predictability of dengue outbreaks that exhibit strong temporal cycling . Although the China CDC has introduced the CIDARS for detection of dengue outbreaks , this method is overly dependent on numbers of notified dengue cases , and there is room to improve the predictive performance [12] . Moreover , due to an inherent defect in the routine surveillance approach , reports of the spread of dengue are delayed [13] . This may slow the quickly public health response to an impending outbreak of infectious disease to some degree . Taking these points into account , we believe that a statistical model holds the promise of being able to provide near real-time quantitative predictions of the occurrence and evolution of an outbreak of dengue , and may be used to efficiently guide the deployment of vector-control operations . Recent studies have exploited digital surveillance based on internet search behavior to timely monitor infectious diseases that have substantial seasonal and geographic variation [13–16] . Due to the increased availability and use of internet over the last decade , the behavior of people seeking information about health has been greatly changed by the availability of health-related information on the internet [40] . In China , according to the 39th Statistical Report on Internet Development , there are 73 . 1 million internet users in China until 2016 , accounting for about 53 . 2% of the national population [41] . The remarkable increase in the internet use and search trends data of people is the basis for us being able to detect and track dengue outbreaks in the country . However , evidence for a working statistical model that exhibits robust ability in the practice of dengue dynamics forecasting is still not available in China , especially for near real-time estimates of dengue epidemic activity in Guangdong , where the risk of dengue infections is high . Our study aimed to develop an accurate prediction tool for dengue outbreaks using machine learning in conjunction with internet search queries and meteorological data in China . Marcel et al . recently discussed the importance of internet-based disease surveillance for rapid disease outbreak detection , and proposed it as a powerful tool to complement traditional disease surveillance [42] . Our analysis found that specific search terms from Baidu are highly correlated with dengue incidence in China . Particularly , for Guangdong , the included search keywords showed a correlation of 0 . 91 with observed dengue incidence , which is basically consistent with previous studies [16] . We further demonstrate the feasibility of applying SVR in dengue incidence forecasting and show that the established SVR model is superior to the other models compared according to the results of the empirical analysis of this study . Our results , based on dengue surveillance data from five other high risk provinces of Yunnan , Guangxi , Hunan , Fujian and Zhejiang also demonstrate a more competitive performance by the SVR model . Our proposed method exhibited itself as a highly efficient tool to predict dengue incidence , and should have predictable positive impacts on the development of an early forecasting system for dengue outbreaks in China . Previous studies also show that a support vector machine-based model has high generalization performance and outperforms classical models in terms of prediction accuracy in Malaysia and Thailand , where the incidence of dengue outbreaks is also high [43 , 44] . Our proposed SVR model further supports the support vector machine-based model as a highly efficient tool to predict dengue incidence . The proposed SVR is a machine learning algorithm implementing the structural risk minimization inductive principle to minimize the generalized error bound and achieve good generalization in complex and noisy data [45] . In comparison to the considered models including step-down linear regression , GBM , NBM , LASSO and GAM , one of the main features of the SVR model is that it performs linear regression in the high-dimension feature space using ε-insensitive loss and tries to reduce model complexity , and handle different types of data sets with high prediction accuracy [46] . Although good generalization performance with SVR has been presented in this study when compared with other five models considered , this model can be abysmally slow in large-scale tasks since it has the extensive memory requirements [47] . Also , another important practical question of SVR lies in choice of the kernel [47] . Regarding the establishment of the SVR model herein , the most suitable kernel function for the dengue data should be considered . It has been suggested that linear kernel function is more robust to multicollinearity , and using the linear kernel function could achieve better performance than the RBF kernel function in case where the number of predictors is relatively large [48] . Additionally , the linear kernel has less complexity than other kernel functions because it has fewer hyperparameters and will be easier to understand . Therefore , the linear kernel function in SVR was used because it could effectively handle many variables in this analysis . Carefully tuning the cost parameter C for the established SVR model and selecting the most suitable value was also an important practical question to avoid overfitting and enhance predictive performance . In practice , the cost parameter C was varied through a wide range of values and the optimal performance assessed using cross-validation for verifying performance [49] . In this study , we applied a cross-validation technique to search the optimized value for the parameter C . By training several SVR models for different values of the parameter C , we chose the best model with the smallest RMSE . Baidu is the most popular search engine in China , making it the most representative data source for tracking online behavior of Chinese people . However , several limitations related to internet search query based surveillance for infectious diseases should be mentioned . First , according to the 39th Statistical Report on Internet Development , the percentage of internet users in the rural areas has steadily increased and is responsible for 27 . 4% until 2016 [41] . Although the availability and popularity of the internet has grown greatly in the rural areas in recent years , the differences in the internet penetration between the rural and urban areas still exist and may influence the internet search queries based surveillance for dengue . Second , internet searching behavior is susceptible to the impact of media reports , which may affect the performance of the internet search term-based predictive model [50] . For example , due to a loss of resolution occurring as a result of media-driven interest that change search behavior , Google Flu Trends was reported to over-estimate the seasonal influenza [40] . In this study , we retrospectively assessed the performance of the proposed SVR model for dengue prediction . Prospective studies should be conducted to evaluate the impacts of media-driven interest or other events that change search behavior of people on the model in the future . In addition , although the variables of dengue case data , internet search surveillance data , meteorological data , and human population data were integrated and analyzed in this work , other sources of information on relevant indicators of risk , particularly evidence on mosquito density and herd immunity [16] , may subsequently be incorporated in future studies . Furthermore , since annual population data in Guangdong province during the study period could not be obtained , the latest data of the 6th population census in 2010 was used to calculate the observed and predicted dengue incidence . The variation of population during the study period might affect the estimates of dengue incidence in this study . In conclusion , the present study demonstrates the utility of using SVR model to track dynamics of dengue outbreaks in China . The proposed SVR model achieves a superior performance in comparison with other forecasting techniques we assessed . The findings of this study will be useful for the government in identifying initiatives needed to strengthen dengue control .
|
Dengue epidemics have posed a great burden expanding of disease , with areas expanding and incidence increasing in China recently . It has remained challenging to develop a robust and accurate forecast model and enhance predictability of dengue incidence . Several state-of-the-art machine learning algorithms , including the support vector regression algorithm , step-down linear regression model , gradient boosted regression tree algorithm , negative binomial regression model , least absolute shrinkage and selection operator linear regression model and generalized additive model , were compared and evaluated to forecast dengue incidence in this study . The SVR model , based on selection by a cross-validation technique , was superior to other models assessed using weekly dengue surveillance data , Baidu search query data and meteorological data during 2011–2014 in Guangdong province . The high accuracy and robustness of the proposed SVR model to predict the occurrence of an outbreak was also validated using data from other provinces , including Yunnan , Guangxi , Hunan , Fujian and Zhejiang , spanning southern China . To the best of our knowledge , this is the first attempt to thoroughly evaluate different algorithms for dengue incidence prediction . Our identification of the optimal model will help to precisely track dengue dynamics in the country .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[
"machine",
"learning",
"algorithms",
"medicine",
"and",
"health",
"sciences",
"china",
"atmospheric",
"science",
"applied",
"mathematics",
"geographical",
"locations",
"simulation",
"and",
"modeling",
"algorithms",
"regression",
"analysis",
"mathematics",
"forecasting",
"statistics",
"(mathematics)",
"artificial",
"intelligence",
"internet",
"infectious",
"disease",
"control",
"humidity",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"epidemiology",
"mathematical",
"and",
"statistical",
"techniques",
"computer",
"networks",
"people",
"and",
"places",
"infectious",
"disease",
"surveillance",
"asia",
"meteorology",
"linear",
"regression",
"analysis",
"earth",
"sciences",
"disease",
"surveillance",
"physical",
"sciences",
"statistical",
"methods",
"machine",
"learning"
] |
2017
|
Developing a dengue forecast model using machine learning: A case study in China
|
Profiling amino acids and acylcarnitines in whole blood spots is a powerful tool in the laboratory diagnosis of several inborn errors of metabolism . Emerging data suggests that altered blood levels of amino acids and acylcarnitines are also associated with common metabolic diseases in adults . Thus , the identification of common genetic determinants for blood metabolites might shed light on pathways contributing to human physiology and common diseases . We applied a targeted mass-spectrometry-based method to analyze whole blood concentrations of 96 amino acids , acylcarnitines and pathway associated metabolite ratios in a Central European cohort of 2 , 107 adults and performed genome-wide association ( GWA ) to identify genetic modifiers of metabolite concentrations . We discovered and replicated six novel loci associated with blood levels of total acylcarnitine , arginine ( both on chromosome 6; rs12210538 , rs17657775 ) , propionylcarnitine ( chromosome 10; rs12779637 ) , 2-hydroxyisovalerylcarnitine ( chromosome 21; rs1571700 ) , stearoylcarnitine ( chromosome 1; rs3811444 ) , and aspartic acid traits ( chromosome 8; rs750472 ) . Based on an integrative analysis of expression quantitative trait loci in blood mononuclear cells and correlations between gene expressions and metabolite levels , we provide evidence for putative causative genes: SLC22A16 for total acylcarnitines , ARG1 for arginine , HLCS for 2-hydroxyisovalerylcarnitine , JAM3 for stearoylcarnitine via a trans-effect at chromosome 1 , and PPP1R16A for aspartic acid traits . Further , we report replication and provide additional functional evidence for ten loci that have previously been published for metabolites measured in plasma , serum or urine . In conclusion , our integrative analysis of SNP , gene-expression and metabolite data points to novel genetic factors that may be involved in the regulation of human metabolism . At several loci , we provide evidence for metabolite regulation via gene-expression and observed overlaps with GWAS loci for common diseases . These results form a strong rationale for subsequent functional and disease-related studies .
High-throughput metabolomics experiments using mass spectrometry platforms are becoming an integral part of clinical and systems biology research . Profiling of amino acids and acylcarnitine species in dried whole blood samples of newborns is used worldwide in neonatal screening programs to identify rare inborn errors of metabolism [1] . These diseases are generally caused by rare mutations , leading to loss of function of an enzyme that catalyzes the biochemical reaction of the respective trait . Recently , many of the amino acid and fatty acid metabolites utilized in newborn screening were also implicated in common complex diseases of adults such as cardiovascular disease , insulin resistance and obesity . Exemplarily , obesity is accompanied by an increase in circulating levels of multiple amino acids , including branched chain amino acids [2 , 3] , and in type 2 diabetics , altered levels of acylcarnitines were described [4 , 5] . Amino acids and acylcarnitines show substantial inter-individual variation [6] and a strong genetic contribution to their blood concentrations has been reported [7] . Thus , the integration of genetic and metabolic profiling holds the promise for providing novel insights into the regulation of metabolic homeostasis in health and disease . Indeed , recent studies have identified common genetic variants associated with a variety of circulating metabolites in serum , plasma or urine using different analytical platforms ( LC-MS/MS , NMR ) [8–24] . However , the complexity of the metabolome cannot be captured by a single technology . Since differences in metabolite abundance have been described between plasma and whole blood [25] , we hypothesized that additional genetic determinants affecting the blood metabolome are yet to be discovered . Thus , we performed an integrated study combining genetics , gene expression and metabolom data ( see S1 Fig for the study design ) . We applied a targeted LC-MS/MS method to measure the abundance of amino acids and acylcarnitines in dried whole blood spots of 2 , 107 individuals and performed genome-wide association analysis . Top findings were replicated in a second independent European Caucasian cohort of 923 Sorbs . Further , going beyond plain genetic associations , we integrated analyses of mRNA levels in leukocytes to establish causal links between genetic variations , gene-expression levels and metabolites . Finally , we explored whether SNP-metabolite associations identified in our study overlap with previously identified genetic loci for other complex traits or diseases .
Quantitative concentrations of 26 amino acids , 36 acylcarnitines and 34 metabolite ratios were determined in dried whole blood spots of 2 , 107 participants of the LIFE Leipzig Heart Study using LC-MS/MS . Metabolites and their ratios reflect metabolic function of various biochemical pathways e . g . urea cycle , branched chain amino acid metabolism or cellular fatty acid oxidation ( see S1 Table for complete list of phenotypes and their categories ) . We performed a genome wide association study ( 2 , 619 , 023 SNPs ) for whole blood metabolites and identified 2 , 261 SNP-metabolite associations ( 119 after pruning ) with p-values <10-7 . These associations comprise 42 metabolites ( including 19 ratios ) and 866 SNPs ( 54 lead-SNPs after pruning ) at 25 unique genomic locations ( Fig 1 , S2 Table ) . QQ-plots and regional association plots for all loci demonstrating valid quality control are presented in the supplemental material ( S2 and S3 Figs ) . Next , replication of top SNPs was sought in an independent cohort of 923 individuals from the Sorb study , where genome-wide SNP and metabolite datasets were available . Good proxies ( r2>0 . 8 ) for replication analysis in the Sorbs were available for 858 ( 99 . 1% ) of our 866 top-SNPs , covering 21 of the 25 identified loci and comprising 2 , 227 associations ( well-imputed proxies were not available for the loci at 1q32 . 3 , 3p24 . 1 , 5p15 . 2 , 20q13 . 2 , see S3 Table for complete results ) . We observed identical directions of effects for 2 , 133 ( 95 . 8% ) combinations of SNPs and metabolites in the replication cohort , resulting in a replication rate of 88 . 3% , when applying a FDR ( false discovery rate ) of 5% ( Fig 2 ) . Replicated lead-SNPs were distributed over 14 of the 21 genomic loci eligible for replication analysis ( Table 1; see S3 Table for results of non-replicated loci ) . In addition , we considered associations at locus #4 ( 2q34 ) with glycine and locus #14 ( 12q24 . 31 ) with C4 as validated results , since these loci were already reported in other GWAS for serum metabolites [8 , 9 , 13–15] . Moreover , non-lead-SNPs at 12q24 . 31 were replicated in the Sorbs at FDR 5% level . None of the other non-replicated loci or loci without proxies in the Sorb study achieved a p-value <10−8 in our initial GWAS . In total , our study led to the identification of 16 unique , validated loci for 36 whole blood metabolites ( Table 1 ) . At six of the 16 loci we identified associations for blood metabolites for the first time i . e . these loci represent novel findings of our study . Also , we successfully validated ten loci previously reported for serum , plasma , and urine metabolites ( Table 1 and S4 Table ) . At three of these loci , associated metabolites were different from those previously reported . In detail , at locus #3 ( 2p13 . 1 ) we detected associations with Arg and related metabolite ratios , whereas earlier associations were reported for plasma N-acetylornithine and related compounds [8 , 13 , 14 , 16] . Further , at loci #11 ( 9q34 . 11 ) and #15 ( 15q22 . 2 ) , we identified associations with methylmalonyl-carnitine , whereas earlier studies reported associations involving the isobaric compound succinyl-carnitine [13 , 14] . To investigate if associated variants have gene regulatory effects , we analyzed our validated lead-SNPs for correlations with gene expression in peripheral blood mononuclear cells ( PBMC ) . Transcriptome data ( 28 , 295 eligible transcripts ) was available for 2 , 112 subjects of the LIFE Leipzig Heart study . At an FDR of 5% , 132 eQTLs were identified for 38 of the 45 validated lead-SNPs , affecting the expression of 69 transcripts . Explained variances of eQTLs ranged between 0 . 4% ( corresponding p-value = 3 . 9x10-3 ) and 28 . 0% ( corresponding p-value = 8 . 0x10-153 , S5 Table ) . We observed eQTLs at 14 of the 16 validated loci , including the six novel loci identified in our study ( Fig 3 and S7 Fig , Table 2 ) . All 14 loci included lead-SNPs with cis-regulatory effects on gene expression . In addition , novel loci #2 ( 1q44 ) and #12 ( 19q11 ) , as well as reported locus #14 ( 12q24 ) also included trans-regulated eQTLs . The trans-eQTLs at locus #2 ( 1q44 ) regulating JAM3 expression were inter-chromosomal and particularly strong , explaining about 13 . 0% of variance ( Fig 3 and S7 Fig , Table 2 ) . We next aimed to assess whether changes in expression of identified eQTL genes can explain observed SNP-metabolite associations in our study . Therefore , we analyzed the relationship between expression levels of these genes and metabolites . We found 40 study-wide significant associations between gene expressions and metabolites , corresponding to 9 loci and 18 eQTL transcripts ( 16 unique genes , see Table 3 and S6 Table ) . We then integrated information from SNP-metabolite ( mQTL ) , SNP-gene expression ( eQTL ) and expression-metabolite associations to form association triangles . A triangle is defined by a triple of SNP , transcript and metabolite showing pair-wise associations ( see methods for details ) . We constructed a network of all pairs of associations and their strengths ( see Fig 4 ) to illustrate the multiple relationships between associated genetic loci , genes and metabolites . An interactive html-document to explore the network is provided as supplement material ( S4 Fig ) . Certain overlaps with previously reported molecular interactions exist . These known relationships are summarized in S11 Table . We identified 177 relations containing 21 unique primary associations between features analysed in our study . Additionally , we identified 16 unique molecules potentially connecting features analysed in our study . As expected , these molecules include Proinsulin and Ubiquitin . Association triangles were further used to test whether variances in gene expression are causally related to variances of metabolite levels . We discovered 38 association triangles mapping to six unique loci including the two novel loci #2 and #10 at 1q44 and 8q24 . 3 , respectively ( S7 Table ) . To estimate the number of such triangles identified by chance , we performed a comprehensive permutation analysis including mQTL , eQTL and expression-metabolite association analysis ( S8 Fig ) . From this , the empirical likelihood of the reported six triangles obtained by chance was estimated to be <1x10-15 . Particularly , in only two of 100 permutations we obtained a single triangle while in 98 of our 100 permutations , no triangles were observed . Next , we used Mendelian randomization to establish a causal link between gene expression and the metabolite . We identified 15 metabolite-gene pairs included in 36 triangles ( S7 Table ) . Next , we investigated whether identified eQTLs explained a significant part of the SNP-metabolite association which we could demonstrate for a total of five loci ( Table 4 ) . Strongest causal effects were found for novel locus #10 at 8q24 . 3 associated with several Aspartic acid traits ( strongest causal effect for ratio Aspartic acid / Acetylcarnitine via cis-regulation of PPP1R16A ) and locus #11 at 9q34 . 11 associated with MMA via PPP2R4 . Finally , we explored whether SNP-metabolite associations identified in our study overlap with genetic loci for clinically relevant traits published in the National Human Genome Research Institute ( NHGRI ) GWAS Catalog . At nine of the 16 validated loci , metabolite associated SNPs matched SNPs previously associated with clinical traits or diseases ( S9 Table ) . We observed associations with platelet and red blood cell properties at three loci associated with acylcarnitines in our study ( 1q44 ( C18 ) , 10q11 ( C3 ) and 15q22 ( MMA ) ) [26–28] . Further , we found that several of our variants were associated with clinical chemistry traits , e . g . fibrinogen ( 2q34 ) [29] , homocysteine ( 2q34 ) [30] and traits reflecting lipid metabolism ( HDL-cholesterol at 2q34 and 15q22 ) [31] , purine catabolism ( uric acid at 10q21 ) [32] , and kidney function ( creatinine at 2p13 and 2q34 ) [33] . At the 2p13 and 2q34 loci , reported associations for creatinine were also linked to chronic kidney disease [34] . In addition , variants at the 2q34 locus for glycine also convey risk for non-small cell lung cancer [35] . Interestingly , recent studies described a key role for glycine in cancer cell proliferation and tumorigenesis [36 , 37] . Further , metabolite associations at 3q27 ( C5OH+HMG ) , 5q31 ( AC-total ) , 9q34 ( MMA ) and 15q22 ( MMA ) overlapped with associations for Parkinson’s Disease [38] , Asthma [39] , Hypersomnia [40] and orofacial cleft [41] , respectively . These co-localizations may implicate a shared genetic basis ( pleiotropy ) between complex traits and aid in forming new hypothesis regarding molecular pathomechanisms .
At two of the six newly identified loci ( 6q23 , ARG1 and 21q22 , HLCS ) , rare variants are known to cause autosomal recessive inborn errors of metabolism , providing a strong biological plausibility for the SNP-metabolite associations . Mutations in ARG1 ( 6q23 ) , encoding arginase , the enzyme which catalyzes the hydrolysis of arginine , are the cause of Argininemia ( OMIM #207800 ) . Here , we report common variants of ARG1 to be associated with arginine levels . Likewise , defects in HLCS ( 21q22 ) are responsible for holocarboxylase synthetase deficiency ( OMIM #253270 ) with affected individuals displaying elevated levels of C5OH+HMG . In line with this observation , the lead SNP at the HLCS locus exhibited a strong cis-eQTL and the allele responsible for higher HLCS expression was associated with lower C5OH+HMG levels . A third novel locus ( #8; 6q21 ) associated with multiple acylcarnitines ( lead phenotype: AC-total ) also contained a gene with direct biochemical relationship to the associated metabolites , namely SLC22A16 , encoding an organic cation/ carnitine transporter . Gene expression of SLC22A16 was regulated in cis at this locus , but SLC22A16 gene expression was not correlated with acyl-carnitine concentrations in whole blood . In fact , the strongest SNP metabolite association at this locus was observed for a non-synonymous coding SNP ( rs12210538 ) in SLC22A16 , which is predicted to be damaging by Polyphen and SIFT [44 , 45] . These findings suggest that associations at 6q21 are more likely driven by this non-synonymous coding mutation than by gene expression of SLC22A16 . The remaining three novel loci relate to candidate genes with no prior connection to metabolism to the best of our knowledge . For the locus at 10q11 . 21 , associated with C2 and C3 , we observed cis-effects on ANUBL1 and FAM21C expression , but gene expressions of both transcripts were not correlated with either C2 or C3 . Thus , additional work will be required to explore the causal link between genetic variation at the 10q11 . 21 locus and C2 and C3 blood concentrations . At novel locus 8q24 . 3 , integration of SNP , eQTL and gene-expression data let to the identification of PPP1R16 as putative causal gene for the association with aspartic acid and corresponding ratios ( lead phenotype: alanine / aspartic acid ) . While we detected strong cis-effects on expression of two local genes , PPP1R16A and LRRC14 , only the eQTL of PPP1R16A partly explained the observed SNP-phenotype associations . Future studies need to address how PPP1R16A , a gene involved in signal transduction [46] , may be affecting blood levels of aspartic acid . Finally , we identified JAM3 encoding the junctional adhesion molecule C ( JAM-C ) as a novel candidate gene of acylcarnitine metabolism . Top associated SNP rs3811444 ( 1q44 ) exhibited an exceptionally strong trans-eQTL for JAM3 , located at 11q25 . This trans effect was also described by other eQTL studies [47] . Gene expression of JAM3 correlated with several long chain acyl-carnitines ( i . e . C16 ) and explained a significant part of the SNP-metabolite association . JAM-C participates in cell-cell adhesion , leukocyte transmigration and platelet activation . The soluble form of JAM-C has been shown to mediate angiogenesis [48] . Homozygous mutations in JAM3 cause hemorrhagic destruction of the brain , subependymal calcification , and congenital cataracts ( HDBSCC , OMIM #613730 ) . At present , the potential functional role of JAM3 in acyl-carnitine metabolism remains elusive . In addition to the identification of novel loci , we replicated and extended functional evidence for SNP-metabolite associations at ten loci previously described in GWAS for serum or plasma metabolites ( Table 1 ) . The majority of these loci contain highly plausible candidate genes based on their biologic function in metabolism ( MCCC1 , ETFDH , SLC22A4/5 , ACADM , ACADS , CPS1 , CRAT ) . Rare loss of function mutations in these genes cause Mendelian inborn errors of metabolism and measuring the respective marker metabolites in whole blood spots is part of neonatal screening programs throughout the world [1] . Here , we validated common variants located in non-coding DNA with modest effect sizes on blood metabolites . Additionally , we found blood eQTLs for MCCC1 , ETFDH , SLC22A4/5 , ACADM , and CRAT . This is in line with evidence from other complex genetic traits , demonstrating that most associations for common variants arise in non-coding DNA and emphasizes the importance of regulatory variants in modulating gene expression [49 , 50] . A striking example is the ACADM locus , where SNPs have been associated with C8 and C10 levels [13 , 14 , 20 , 21] . In our study , gene-expression of ACADM was associated with C8 and C10 blood levels and we showed for the first time that this relationship was causal explaining a part of the observed SNP association . In conclusion , our study expanded the current knowledge on the genetic regulation of human blood metabolites by adding six novel genetic loci . Furthermore , by integrative analysis of SNP , gene expression and metabolite data , we derived mechanistic insights into the molecular regulation of blood metabolites . At several loci , we provide evidence for metabolite regulation via gene-expression and observed overlaps with GWAS loci for other complex traits and diseases , pointing towards potential pathomechanisms via metabolic alterations . Additional functional studies are required to elucidate the cellular mechanisms how the discovered candidate genes affect metabolic pathways and relate to disease pathology .
LIFE Leipzig Heart is an observational study in a Central European population designed to analyze genetic and non-genetic risk factors of atherosclerosis and related vascular and metabolic phenotypes [51] . Patients undergoing first-time diagnostic coronary angiography due to suspected stable CAD with previously untreated coronary arteries , patients with stable left main coronary artery disease and patients with acute myocardial infarction were recruited . The latter were excluded for the present analysis . The study meets the ethical standards of the Declaration of Helsinki . It has been approved by the Ethics Committee of the Medical Faculty of the University of Leipzig , Germany ( Reg . No 276–2005 ) and is registered at ClinicalTrials . gov ( NCT00497887 ) . Written informed consent including agreement with genetic analyses was obtained from all participants . In this analysis , we considered a total of 2 , 464 individuals . From these , 2 , 107 had complete genotype , metabolite and covariate data qualifying them for GWAS analysis ( descriptive statistics can be found in S9 Table ) . A subset of 1 , 856 individuals had complete data of genotypes , gene expression , metabolites and covariates . These individuals were used for integrative analyses ( see study design , S1 Fig ) . The Sorbs were recruited from the self-contained Sorbs population in Germany [52–54] . All individuals were at fasting state . Phenotyping included standardized questionnaires for past medical history and family history , collection of anthropometric data ( weight , height , waist-to-hip ratio ) and results from an oral glucose tolerance test . A complete set of high-quality genotype data , metabolites and covariates was available for 923 subjects ( S9 Table ) . The study was approved by the ethics committee of the University of Leipzig and all subjects gave written informed consent before taking part in the study . An overview of the study design is presented in S1 Fig . In brief , we first performed a genome-wide metabolite quantitative trait ( mQTL ) analysis in the LIFE Leipzig Heart cohort , with replication of the top-SNPs in the Sorbs cohort . Following this two-stage design , we applied a liberal cut-off of 1 . 0x10-7 for the initial GWAS to identify candidate loci . A stringent cut-off is applied at the replication stage where we control the ( study-wide ) FDR at 5% based on permutation analysis [55] . This accounts for the correlation structure of individuals , SNPs and metabolites and the multiple testing issue ( for details see below section “Genome-wide association analysis and SNP replication” ) . Functional relevance of identified loci was studied in the LIFE Leipzig Heart cohort by analyzing expression quantitative traits ( eQTL ) and gene expression-metabolite associations followed by causal inference regarding discovered associations . Venous blood samples were obtained from all study participants and 40μl of native EDTA whole blood were spotted on filter paper WS 903 ( Schleicher and Schüll , Germany ) in the LIFE Leipzig Heart study . In the Sorb cohort , 40μl cell suspension obtained after plasma centrifugation ( 10 min at 3500 x g ) were spotted on filter paper . All blood spots were stored at -80°C after 3 hours of drying until mass spectrometric analysis . Sample pretreatment and measurement is described elsewhere [56–58] . In brief , 3 . 0 mm diameter dried blood spot punches ( containing 3 μL whole blood ) were extracted with methanol containing isotope labelled standards . After sample extraction and derivatization , analysis was performed on an API 2000 tandem mass spectrometer ( Applied Biosystems , Germany ) . Quantification of 26 amino acids , free carnitine and 34 acylcarnitines including related metabolites was performed using ChemoView 1 . 4 . 2 software ( Applied Biosystems , Germany ) . Samples were analysed within 23 analytical batches with two quality controls samples in each batch . Mean inter-assay coefficients of variation were below 11% for amino acids and below 19% for acylcarnitines . Further , using these 61 directly measured analytes , we derived a number of biologically relevant sums ( n = 1 , total acylcarnitine ) and ratios ( n = 34 ) to assess reaction equilibria within physiological pathways and processes ( e . g . Fischer’s ratio [59] ) . Consequently , a total of 96 quantities were analyzed as GWAS traits . A list of metabolites and quantities is presented in S1 Table . Metabolites with more than 20 percent of values below detection limit were dichotomized for analysis ( below detection limit versus above detection limit ) . This applies for the metabolites C5:1 , C6DC , C14OH , C16OH , MeGlut , C18:1OH , C18:2OH , C18OH and C20:3 . Quantities were arsinh-transformed ( area sinus hyperbolicus ) which is close to a log-transformation for large values but does not emphasize differences between small values and can operate on values of zero . Transformed quantities were approximately normal distributed . Values outside of the Interval Mean ± 5*SD were considered as outliers and were removed to stabilize subsequent regression analysis . We previously analysed a variety of factors influencing blood metabolites . Age , sex , diabetes and fasting status show pronounced effects on several metabolites while log-BMI , smoking and some blood traits showed effects on selected metabolites . Therefore , we decided to adjust our analyses for these potential confounders . Genome-wide association analyses for blood 96 metabolites was performed in the LIFE Leipzig Heart samples ( N = 2 , 107 with complete phenotypes , covariates and high-quality genotypes ) . Associations were tested by linear regression models using gene-doses of imputed SNPs . We adjusted for age , sex , log-BMI , diabetes status , smoking status , fasting status , haematocrit , platelet count , white blood cell count and the first three genetic principal components . Results revealed no signs of genomic inflation ( maximum lambda equal 1 . 018 , see S10 Table ) . To avoid reporting of redundant SNP information , the top-SNP list was ordered according to minimal p-values and pruned applying a linkage disequilibrium cut-off of r2<0 . 3 . Replication analysis was performed in the independent cohort of Sorbs ( N = 923 with complete genotype and metabolite data ) and for all combinations of SNPs and metabolites achieving a p-value of <10−7 in our first stage GWAS . Based on our unpruned GWAS top-list , we retrieved all SNPs within a ±50kB environment which were successfully imputed in the Sorbs ( IMPUTE-info score>0 . 3 in both , 500K and 6 . 0 subsample ) . Then , on the basis of the LIFE Leipzig Heart data , we assessed which of these SNPs are the best proxies of the corresponding top-SNPs to pair GWAS top-SNPs with optimal proxies of good quality within the Sorbs study . Associations between pairs of proxies and metabolites were again analyzed using linear regression analyses of gene-doses . Here , we adjusted for age , sex , log-BMI , diabetes status , smoking status , haematocrit , platelet count , white blood cell count and the relatedness structure ( [52 , 64 , 65] , function “polygenic” of the “GenABEL” package of R was used to deal with the relatedness structure [63] ) . Since test statistics are correlated due to LD between SNPs and correlations between metabolites , we decided to control the false-discovery rate ( FDR ) at 5% rather than family-wise error rates . Null-distribution for q-value calculation was determined by permutation analysis . For this purpose , 1000 random permutations of the links between SNPs and metabolites were analyzed . We compared our results with published GWAS hits on the basis of the GWAS catalogue ( http://www . genome . gov/gwastudies/ , date of download March , 4th , 2014 ) . Required LD information was derived from HapMap3 ( release 28 ) and 1000genomes project ( release 20110521 version 3 f , restricted to SNPs with a MAF ≥ 1% ) . In addition , further evidence from published mQTL studies was manually included in this analysis to assess novelty of our results . A total of 13 studies were analyzed [8 , 9 , 12–21 , 23] ( see also S4 Table ) . A locus was considered as novel if none of its SNPs were in linkage disequilibrium ( r2>0 . 3 ) with any published mQTL hit reaching study-wide significance as defined by the authors of the corresponding publication . To increase relevance , we did not match the associated metabolic phenotypes between our study and the published ones , i . e . our approach of considering loci as novel is conservative . In complete analogy to this analysis , we determined whether our top hits are associated with other traits for which results are published in the GWAS catalogue as well as those reported in two GWAS on plasma lipids [10 , 31] . These traits could point toward other causal or pleiotropic effects . If applicable , information on genetic disorders related to our loci were retrieved from OMIM ( http://omin . org ) . Peripheral blood mononuclear cells were isolated in the LIFE Leipzig Heart cohort using Cell Preparation Tubes ( CPT , Becton Dickinson ) as previously described [66] . Total RNA was extracted using TRIzol reagent ( Invitrogen ) and quantified with an UV-Vis spectrophotometer ( NanoDrop , Thermo Fisher ) . 500 ng RNA per sample were ethanol precipitated with GlycoBlue ( Invitrogen ) as carrier and dissolved at a concentration of 50–300 ng/μl prior to probe synthesis . N = 2 , 501 samples were hybridised to Illumina HT-12 v4 Expression BeadChips ( Illumina , San Diego , CA , USA ) in batches of 48 and scanned on the Illumina HiScan instrument according to the manufacturer’s specifications [60] . Documentation of sample processing included batch information at any processing step allowing adjustment in subsequent data analysis . Raw data of all 47 , 323 probes was extracted by Illumina GenomeStudio , 47 , 308 probes could be successfully imputed in all samples . Data was further processed within R/ Bioconductor R [67] . Individuals having an extreme number of expressed genes ( defined as median ± 3 interquartile ranges ( IQR ) of the cohort’s values ) were excluded . Transcripts that were not expressed according to Illumina’s internal cut-off as implemented in the “lumi” Bioconductor package ( p ≤ 0 . 05 in at least 5% of all samples ) were excluded from further analysis . Expression values were quantile-normalised and log2-transformed [68] . For further outlier detection , we calculated the Euclidian distance between all individuals and an artificial individual which was defined as the average of samples after removing 10% samples farthest away from the average of all samples . Individuals with a distance larger than median + 3 IQR were excluded . Furthermore , we defined for each individual a combined quantitative measure combining quality control features available for HT-12 v4 ( i . e . ratio of levels of perfect-match vs . mismatch control probes , mean signal of perfect-match control probes , mean of negative control probes and labelling-control probes , ratios of high-concentrated , medium-concentrated and low-concentrated control-probes , mean of house-keeping genes , Euclidian distances of expression values , number of expressed genes , mean signal strength of biotin-control-probes ) . We calculated Mahalanobis-distance between all individuals and an artificial individual having average values for these quality control features . Individuals with a distance larger than median + 3 IQR were excluded . Transcript levels were adjusted for known batch effects using an empirical Bayes method as described [69] and residualised for age , sex , monocyte counts and lymphocyte counts . Additionally , we calculated principal components of the expression data and residualised for the first five principal components of expression data to account for unmeasured batch effects [70] . Pre-processing resulted in 28 , 295 expression probes corresponding to 19 , 519 genes . Chromosomal mapping of expression probes and assignment of gene names was done using information as reported by the manufacturer ( HumanHT-12_V4_0_R2_15002873_B ) . After quality control , combined SNP and gene-expression data were available for a total of 2 , 112 individuals , from which 1 , 856 had been included in the GWAS . eQTL analysis of the pruned GWAS top-list was performed by linear regression analysis of gene-doses using the R add-on package Matrix eQTL [71] . EQTLs were considered as cis-regulated if the distance between SNP and the centre of the associated expression probe was not larger than 1 Mb , otherwise they were considered as trans-regulated . Cis- and trans- specific significance thresholds were derived by a Benjamini-Hochberg ( B-H ) procedure implemented in Matrix eQTL . For our data , cis associations with a p-value up to 0 . 0039 and trans-associations with a p-value up to 3 . 6x10-14 were considered study-wide significant at FDR<5% . B-H q-values were empirically confirmed by 100 permutation tests ( permutation of SNP and gene-expression profiles ) . Further details can be found elsewhere [72] . Association analysis of gene-expression and metabolites was performed in 1 , 957 individuals for which both information as well as covariates were available ( 1 , 856 of these individuals had been included in the GWAS ) . Again , we adjusted for age , sex , log-BMI , diabetes status , smoking status , fasting status , haematocrit , platelet count , white blood cell count . FDR was controlled at 5% . As we observed multiple relationships between genetic loci , gene-expressions , and metabolites , we visualized all associations found at FDR 5% in a network . Previously published relations were identified by mapping genetic loci , genes , and metabolites from mQTL , eQTL , and gene-expression-metabolite association analysis to QIAGEN’s Ingenuity Pathway Analysis ( IPA , QIAGEN Redwood City , www . qiagen . com/ingenuity ) , as of May , 2015 ) . This database includes , among many other information , data on genome-wide protein-protein interactions , activation / co-localization and enzymatic reactions . Significantly associated SNPs were represented by the three most proximal genes and metabolite ratios by the individual nominator and denominator . For a more detailed characterization of the observed SNP-metabolite associations , we integrated genotype , gene expression and metabolite data to construct association triangles . A triangle is defined as a SNP that is significantly associated with both , a certain expression probe and a certain metabolite . Thereby , the expression probe must be also associated with the metabolite . For this purpose , we first determined the top associated SNP per locus , its corresponding best associated metabolite and eQTLs of that SNP ( FDR = 5% , see above ) . Resulting triples of SNP , transcript level of eQTL and metabolite level were restricted to those showing a significant association between mRNA expression and metabolite level ( FDR = 5% , see above ) . These gene-expressions were considered as possible explanatory quantities of the SNP-metabolite association . We simulated the expected number of these association triangles under the null distribution by performing a comprehensive permutation analysis: We performed 100 permutations where we randomly assigned expression datasets and metabolic datasets to genetic datasets . We analysed these datasets for mQTLs , eQTLs , and gene-expression associations in accordance to our original analysis . For each of these 100 permutation-based datasets , we counted the number of pairwise associations and association triangles and compared it with the results of our original dataset . We calculated the empirical likelihood of triangles by comparing the observed number of six triangles with the number of triangles under the null assuming a Poisson distribution . In order to exclude spurious correlation between gene-expression and metabolites as a cause of the observed association , we performed a Mendelian randomization analysis using our eQTL SNPs as instrumental variables [73] . In general , it is not easy to prove that the conditions of Mendelian randomization are fulfilled . In particular , a direct SNP effect on metabolites cannot be excluded , violating one of the assumptions [74] . Therefore , we adapted the Mendelian randomization analysis by using the residuals of metabolites regarding the remaining direct SNP effects ( see also S1 Text for an extended discussion ) . Standard errors of Mendelian randomization effects were derived by Jackknife [75] . Furthermore , we tested whether gene-expressions explain at least parts of the observed mQTL associations . A subset of 1 , 856 individuals for which SNP , gene-expression , metabolite and covariate data were available , was eligible for this purpose . We analysed regression models of metabolites in dependence on SNPs , covariables and with or without gene-expression . We asked whether the absolute value of the beta-estimator of the SNP is reduced if gene-expression is added to the model . In this case , gene-expression explains a part of the observed SNP-metabolite association . The difference of these SNP beta-estimators is tested against zero by calculating Jackknife standard errors . This analysis also provides evidence for causal relations between genetic variants , gene-expression levels and metabolite concentrations . Since we observed that it is more stringent and conservative than Mendelian randomization analysis , our conclusions regarding causality are based on this type of analysis . To gain additional insights into possible functional mechanisms of our loci , we performed the same analysis for all independently associated top-SNPs .
|
Human metabolite levels differ between individuals due to environmental and genetic factors . In the present work , we analyzed whole blood levels of amino acids and acylcarnitines , reflecting disease relevant metabolic pathways , in a cohort of 2 , 107 individuals . We then performed a genome wide association analysis to discover genetic variants influencing metabolism . Thereby , we discovered six novel regions in the genome and confirmed ten regions previously found to be associated with metabolites in plasma , serum or urine . Subsequently , we analyzed whether these variants regulate gene-expression in peripheral mononuclear cells and at several loci we identified novel causal relations between SNPs , gene-expression and metabolite levels . These findings help explaining the functional mechanisms by which associated genetic variants regulate metabolism . Finally , several SNPs associated with blood metabolites in our study overlap with previously identified loci for human diseases ( e . g . kidney disease ) , suggesting a shared genetic basis or pathomechanisms involving metabolic alterations . The identified loci are strong candidates for future functional studies directed to understand human metabolism and pathogenesis of related diseases .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Integration of Genome-Wide SNP Data and Gene-Expression Profiles Reveals Six Novel Loci and Regulatory Mechanisms for Amino Acids and Acylcarnitines in Whole Blood
|
Epidemiological studies suggest that allergy risk is preferentially transmitted through mothers . This can be due to genomic imprinting , where the phenotype effect of an allele depends on its parental origin , or due to maternal effects reflecting the maternal genome's influence on the child during prenatal development . Loss-of-function mutations in the filaggrin gene ( FLG ) cause skin barrier deficiency and strongly predispose to atopic dermatitis ( AD ) . We investigated the 4 most prevalent European FLG mutations ( c . 2282del4 , p . R501X , p . R2447X , and p . S3247X ) in two samples including 759 and 450 AD families . We used the multinomial and maximum-likelihood approach implemented in the PREMIM/EMIM tool to model parent-of-origin effects . Beyond the known role of FLG inheritance in AD ( R1meta-analysis = 2 . 4 , P = 1 . 0 x 10−36 ) , we observed a strong maternal FLG genotype effect that was consistent in both independent family sets and for all 4 mutations analysed . Overall , children of FLG-carrier mothers had a 1 . 5-fold increased AD risk ( S1 = 1 . 50 , Pmeta-analysis = 8 . 4 x 10−8 ) . Our data point to two independent and additive effects of FLG mutations: i ) carrying a mutation and ii ) having a mutation carrier mother . The maternal genotype effect was independent of mutation inheritance and can be seen as a non-genetic transmission of a genetic effect . The FLG maternal effect was observed only when mothers had allergic sensitization ( elevated allergen-specific IgE antibody plasma levels ) , suggesting that FLG mutation-induced systemic immune responses in the mother may influence AD risk in the child . Notably , the maternal effect reported here was stronger than most common genetic risk factors for AD recently identified through genome-wide association studies ( GWAS ) . Our study highlights the power of family-based studies in the identification of new etiological mechanisms and reveals , for the first time , a direct influence of the maternal genotype on the offspring’s susceptibility to a common human disease .
Atopic dermatitis ( AD , eczema ) is a chronic inflammatory skin disease with 10–20% prevalence in industrialized countries . The etiology of AD is complex , with multiple genetic and environmental factors influencing disease risk . Genome-wide association studies ( GWAS ) have successfully identified common genetic variants predisposing to AD , but the effect of these risk loci is small and altogether only account for a fraction of the disease heritability . The filaggrin gene ( FLG ) encodes a structural protein playing a critical role in the terminal differentiation of the epidermis and in skin barrier function [1] . Loss-of-function mutations in FLG were identified as the cause of ichthyosis vulgaris , a common Mendelian trait characterized by dry , scaly skin and frequent AD [2] . Subsequent studies revealed that FLG mutations also strongly predispose to AD [3 , 4] . This observation has been widely replicated , rendering FLG the strongest and best characterized AD risk locus to date [1] . Overall , evidence from human and animal studies demonstrated that filaggrin deficiency results in altered skin structure , impaired barrier function and enhanced antigen penetration through the skin , leading to the production of allergen-specific IgE antibodies ( specific sensitization ) and AD [5–7] . Epidemiological studies on allergic diseases have shown that maternal allergy is a stronger risk factor for the child than paternal allergy [8 , 9] , although some conflicting results have been reported for AD [10 , 11] . The molecular basis of this preferential maternal transmission of allergy risk is currently unknown but it can potentially occur through two different biological mechanisms , genomic imprinting or direct maternal genotype effects . In genomic imprinting , either the maternally or the paternally inherited allele is expressed while the alternate allele is silenced . Thus , the effect of an allele depends on its parental origin resulting in phenotypic differences between reciprocal heterozygotes ( parent-of-origin effects ) [12 , 13] . Recent studies have demonstrated that parent-of-origin effects in complex diseases can be due to genetic variation in imprinted genes [12 , 13] . Alternatively , maternal genotype effects occur when the maternal genotype directly influences the child’s phenotype . This effect is independent of the child’s own genotype and occurs through the maternally provided environment during prenatal development . Maternal genotype effects can lead to phenotypic differences between reciprocal heterozygotes and are thus considered parent-of-origin effects [13 , 14] . We hypothesized that loss-of-function mutations in FLG may show parent-of-origin effects . Analysis of 2 large family-based cohorts strongly supports that maternal FLG mutations directly increase AD risk in the children .
To systematically identify loss-of-function variants at the FLG locus , we used data of the Exome Aggregation Consortium ( ExAC [15] ) , which includes whole exome sequencing results of 61 , 486 individuals . Filtering by frameshift or non-sense mutations revealed 254 loss-of-function mutations in the gene ( S1 Table ) . The majority of FLG mutations were very rare , 227 of 254 FLG mutations ( 89 , 4% ) had an allele frequency ( AF ) < 0 . 0001 . Further analysis revealed the presence of population-specific mutations . For example , the p . L4022X mutation was common in East Asia ( AF = 0 . 02 ) but absent from all other populations studied . This data confirms and extends previous reports of allelic heterogeneity and population-specific mutations in FLG [1] . In the European ( non-Finnish ) population ExAC reported 146 loss-of-function mutations with a combined AF of 0 . 052 . Of these , the 4 most prevalent mutations , accounting for 86% of all mutant alleles in this population , were selected for genotyping in the present study: p . 761fsX35 ( c . 2282_2285delCAGT; rs558269137 , referred to as c . 2282del4 ) , p . R501X ( c . 1501C>T; rs61816761 ) , p . R2447X ( c . 7339C>T; rs138726443 ) , and p . S3247X ( c . 9740C>A; rs150597413 ) . The 4 selected mutations were genotyped in 759 complete nuclear families from Central Europe recruited through one or more children with AD ( methods and Table 1 ) . The allele frequencies in the founders were in good agreement with previous studies ( 0 . 059 , 0 . 034 , 0 . 01 and 0 . 002 for c . 2282del4 , p . R501X , p . R2447X and p . S3247X , respectively ) [16 , 17] . As previously reported , FLG mutations showed a strong over-transmission from heterozygote parents to AD-affected children in a Transmission Disequilibrium Test ( TDT; Table 2 [16 , 17] ) . We observed no linkage disequilibrium among FLG mutations , since each mutation was on a different haplotype and 2 different mutations never occurred together in the same haplotype ( S2 Table ) . Since previous studies reported that these 4 loss-of-function FLG mutations have the same effect on AD risk , we decided to merge all variants into a combined genotype [4] . This enabled us to work with one common instead of 4 low-frequency variants . Unless stated otherwise , the results presented below refer to the combined genotype . Testing for parent-of-origin effects was performed with the PREMIM/EMIM software . This tool uses a multinomial model-based maximum-likelihood approach for flexible modelling of parent-of-origin effects ( see methods ) [18 , 19] . In order to increase power , we included FLG genotypes of 1147 unrelated AD cases and 3339 population-based controls . This set of unrelated individuals does not provide information on parent-of-origin effects , but increases the power to detect them by improving the estimation of genotype frequencies in the general population ( see methods ) [18 , 19] . We performed a step-by-step analysis starting with a basic genetic model and successively including additional risk parameters modelling parent-of-origin effects . The basic scenario ignored the available parental genotypes and tested the effect of the child´s genotype on his own phenotype . As expected , we observed large effects with relative risks of 3 . 1 for heterozygous ( R1 parameter ) and 10 . 5 for homozygous FLG mutation carriers ( R2 parameter ) ( Child Genotype or CG model; R1 = 3 . 1 , R2 = 10 . 5; PCG = 5 . 9 x 10−74; Table 3 ) . Next maternal genotype effects were modelled by including an additional parameter , S1 , to estimate the relative AD risk of children whose mother carried a FLG mutation . Children of FLG mutation-carrier mothers had a striking 1 . 55 fold increase in AD risk independently of their own genotype ( Mother-Child Genotype model or MCG; R1 = 2 . 57 , R2 = 7 . 97 , S1 = 1 . 55; PMCG = 2 . 7 x 10−77; Table 3 ) . A comparison of the Child Genotype and the Mother-Child Genotype models by a likelihood-ratio test , strongly supported the existence of a maternal genotype effect ( PMCG vs CG = 5 . 0 x 10−6 ) . In addition , we observed no evidence of interaction between the child and maternal genotypes , indicating that carrying a mutation and having a mutation carrier mother are independent risk factors with additive effect on disease risk ( S3 Table and methods ) . Thus , children with both risk factors , i . e . carrying a FLG mutation and having a mutation carrier mother , have a nearly 4-fold increased disease risk ( R1 x S1 = 3 . 6 ) . Importantly , both genomic imprinting and maternal genotype effects can lead to similar patterns of parent-of-origin effects and specific tests need to performed to distinguish them [13 , 14] . Finally , we tested an imprinting model by including the imprinting parameter , Im , which represents the relative risk of the child when inheriting a mutant allele from the mother as opposed to the father . A comparison with the Child Genotype model provided marginal support for the presence of imprinting ( PIm vs CG = 0 . 047; Table 3 ) . In order to test which parent-of-origin scenario better fits our data we performed comparisons versus a full model containing all risk parameter ( R1 , R2 , S1 and Im ) . Interestingly , adding the maternal genotype parameter S1 to a model already containing Im resulted in a significantly better model ( Table 4; p = 3 . 2 x 10−6 ) . On the contrary , adding Im to a model already containing S1 provided only a marginal improvement ( p = 0 . 03 ) . These results favour the existence of a direct maternal genotype effect of FLG . We aimed to replicate our findings by examining the same 4 FLG mutations in an independent Northern European population including 450 AD families and 1854 population-based control individuals ( methods and Table 1 ) [17 , 20–22] . Step-by-step analysis with PREMIM/EMIM again supported a maternal genotype effect . The genotypes of both children and mothers had an independent effect on AD risk , and children of FLG-carrier mothers showed a 1 . 4 fold increased risk ( R1 = 2 . 13 , R2 = 5 . 89 , S1 = 1 . 42; PMCG = 1 . 4 x 10−17; PMCG vs CG = 0 . 005; Table 3 and S3 Table ) . Importantly , the results obtained were consistent in both populations studied providing strong support to the existence of maternal genotype effects on FLG . A meta-analysis was performed using the inverse variance method as implemented in METAL [23] , which uses the effects estimates and standard errors from each risk parameter . This revealed a highly significant 1 . 5 fold increased AD risk in children of FLG-carrier mothers ( S1meta-analysis = 1 . 50; P = 8 . 4 x 10−8; Table 3 ) . Analysis of parental genotypes revealed a higher prevalence of FLG mutations in mothers than in fathers in both study populations ( Fig . 1 ) . This is consistent with the maternal genotype effect observed . Additionally the frequency of FLG mutations in the parental population ( mothers and fathers together ) was remarkably higher than in population-based controls of unknown phenotype . This is likely due to the recruitment of families with multiple affected children leading to a parental population enriched in strong genetic risk factors . At this stage we considered the potential weaknesses of our study in order to discard false positives due to methodological issues and to gain further support to the existence of a FLG maternal genotype effect . We analysed the FLG c . 2282del4 , p . R501X and p . R2247X mutations independently ( this was not possible for p . S3247X since it was too rare ) . The maternal effect and the increased frequency of mutations in mothers were found for all 3 mutations in both study populations ( S4 Table and S5 Table ) . This suggests that the maternal effect is not specific to a given variant but a general characteristic of FLG loss-of-function mutations . In the populations studied , mothers typically have a more prominent role than fathers in children’s health care [24] . We hypothesized that preferential ascertainment of AD-affected mothers carrying FLG mutations may be the cause of the observed maternal effect . Indeed , we observed a higher frequency of AD-affected mothers than fathers , which could be due to a genuine maternal effect or to ascertainment bias ( AD prevalence in Central European mothers = 0 . 23 and fathers = 0 . 12; Northern European mothers = 0 . 35 and fathers = 0 . 19 ) . In order to avoid this potential bias we repeated our analysis including only families in which both parents had a negative history of AD . Importantly , the maternal genotype effect remained strong and significant in the remaining population ( Meta-S1 = 1 . 38; P = 0 . 003; S6 Table ) . We also tested the potential effect of the paternal FLG genotype on the children . Since this option is not available in PREMIM/EMIM , we performed the analysis after exchanging the paternal and maternal genotypes on our genotype files . This analysis revealed no significant effect of the paternal FLG status ( S7 Table ) . A large proportion of the families included in the present study ( 60% ) were recruited through an affected sib pair . Aiming to maximize power , all previous analyses were performed considering all affected siblings as independent individuals which may lead to biased risk parameter estimates . We therefore repeated the analysis including only one affected child per family and found that the magnitude of the maternal effect remained constant in this set of independent trios ( S1meta-analysis = 1 . 45; P = 1 . 1 x 10−4; S8 Table ) . Filaggrin has a major role in cutaneous barrier function [1] . According to publicly available datasets [25–27] FLG expression is highest in skin and absent in tissues relevant for mother-child interactions such as uterus , placenta , or mammary gland ( S1 Fig . and S2 Fig . ) . However , recent studies demonstrated that FLG mutations result in increased antigen penetration through the skin and the production of allergen-specific antibodies ( IgE , specific sensitization ) [6 , 7 , 28] . Thus , we hypothesized that systemic inflammatory responses in FLG carrier mothers may influence AD risk in children via feto-maternal immune crosstalk . All available maternal plasma samples from the Central European study were therefore tested for the presence of allergen-specific IgE , which is a well-characterized biochemical marker of allergy [29] . We performed a stratified analysis in 253 families with and 311 families without maternal specific sensitization . In the families with maternal specific sensitization ( +Mat_sens ) we observed a strong maternal genotype ( S1 = 1 . 63; PMCG vs CG = 0 . 005 ) and a weak child genotype effect ( R1 = 1 . 37; ; Pnull = 0 . 03; Table 5 ) . This was consistent with a marginal over-transmission of FLG mutations from parents to affected offspring in a transmission disequilibrium test ( TDT , transmitted ( T ) : untransmitted ( U ) = 97:65 , P = 0 . 04; Table 6 ) . In contrast , the opposite pattern was observed in the group of families with non-allergic mothers ( −Mat_sens ) . Here , the maternal genotype effect was not significant ( S1 = 1 . 18; PMCG vs CG = 0 . 3 ) while the child genotype effect was strong ( R1 = 2 . 3; Pnull = 9 . 9 x 10−11 ) . This observation was confirmed by a striking over-transmission of FLG mutations in the TDT ( T:U = 169:68 , P = 5 . 3 x 10−8 ) . In concordance with the different rates of mutation transmission observed in both groups , the frequency of FLG mutations was significantly higher in AD-children of non-allergic mothers ( mutation frequency 0 . 16 and 0 . 11 in affected children of −Mat_sens and +Mat_sens families , respectively; OR = 1 . 48; P = 0 . 005; Table 6 ) .
We report here that maternal loss-of-function mutations in FLG directly influence AD risk in the offspring independently of the child’s own genotype . Importantly , this maternal effect was observed consistently for all different FLG mutations tested and in 2 independent populations . Given that genomic imprinting and maternal genotype effects can lead to similar patterns of parent-of-origin effects [14] , we specifically modelled both scenarios . Although we cannot completely exclude an imprinting effect , our data supports a direct maternal genotype effect of FLG mutations . This is consistent with the lack of known imprinted genes in the 1q21 . 3 genomic region containing FLG [25] . It is not obvious how maternal mutations in a skin-barrier gene can influence the child’s phenotype . However , given that filaggrin deficiency promotes specific sensitization , we hypothesized that systemic immune responses may play a role in the FLG maternal effect . Our observation that the FLG maternal effect is significant only in the group of sensitized mothers supports this hypothesis and indicates that FLG mutations in allergic mothers act as a strong environmental risk factor for AD in the child . The importance of prenatal mother-child interaction in shaping the child’s immune phenotype is highlighted by studies in pregnant mice: while the induction of Th2 immune responses increased susceptibility to allergic asthma in the offspring , a protective effect was observed upon induction of Th1 responses , emphasizing the importance of maternal immune status during pregnancy [30–32] . Supporting evidence in humans arise from epidemiological studies showing that exposure to a microbial-rich farm environment during pregnancy protects children from the development of allergic diseases [33] . It is unknown how this “immunological imprinting” may be transmitted from mother to child , but animal and human studies suggest the induction of epigenetic modifications in relevant immune cells in the offspring [31 , 34] . Future large population-based studies with parental DNA , data on maternal allergic sensitization and biological material available for epigenetic analysis will be required to further explore this interesting hypothesis . A recent mouse study found that parent-of-origin effects are widespread and account for an unexpectedly large proportion of complex trait heritability [35] . This is supported by human studies demonstrating that the parental origin of an allele inherited by the offspring can affect disease susceptibility to complex diseases [12 , 13 , 36–38] . However , evidence for the existence of maternal genotype effects , which occur without transmission of the risk allele to the offspring , comes mainly from animal studies analysing the effect of maternal gene knockouts in wild-type offspring [39] . Examples of such effects in humans are scarce and refer to rare congenital malformations [40 , 41] . The present work is , to our knowledge , the first report of a large maternal genotype effect in a common human disease . Notably , the magnitude of the maternal effect ( RR = 1 . 5 ) was consistent in both data sets and exceeded that reported for most AD genetic risk factors identified to date . Interestingly , AD is commonly the first manifestation of allergic disease and filaggrin deficiency is a risk factor for the transition from AD to other atopic diseases such as food allergy , hay fever , and asthma [5 , 16] . Thus , it is tempting to speculate that maternal FLG mutations may influence the risk of a much wider range of allergic disorders . This and other studies provide proof-of-principle that associations originally discovered by case-control analysis can arise as a consequence of parent-of-origin effects , although with an underestimation of the effect size due to inaccurate genetic modelling [12] . Family-based studies re-evaluating previously identified susceptibility loci will enable the identification of parent-of-origin effects and help characterize part of the missing heritability in complex traits .
The study was carried out in accordance to the approval of the ethics commission of the Charité—Universitätsmedizin Berlin ( ref EA2/054/10 ) and following the guidelines of the declaration of Helsinki . Informed consent was obtained from all probands or their legal guardians . We investigated samples originating from European family-based and population-based studies . All samples were divided , according to the country of origin , into a Central and a Northern European study population ( Table 1 ) . The GENUFAD study ( Genetic Analysis of Nuclear Families with Atopic Dermatitis ) recruited complete nuclear families with at least two children affected with early-onset ( <2 years of age ) and moderate to severe AD as previously described [20] . A doctor’s diagnosis of AD was made according to standard criteria [42] . The GENUFAD study contributed 522 complete German nuclear families to the Central European study and 32 Swedish families to the Northern European data set . A large proportion of these families have been reported in previous studies [16 , 43] . The MAS ( Multicenter Allergy Study ) is a previously described population-based birth cohort in which 1314 German infants were followed since 1990 to investigate the epidemiology of allergic diseases [44] . The diagnosis of AD was made as previously described [16] . 112 German MAS trios , consisting of a child with AD and both parents , were included in the Central European study . The ETAC ( Early Treatment of the Atopic Child ) is a European study which recruited infants diagnosed with AD in their first year of life into a randomized , double blind , placebo controlled trial on the efficacy of cetirizine in the prevention of asthma [45] . Children with early onset and moderate to severe disease were selected for the present study when parental DNA was available . The ETAC study contributed 21 Swedish trios to the Northern European study . Additionally , 125 ETAC trios from different European countries were included in the Central European study ( 48 from the Netherlands , 23 from Italy , 20 from the UK , 15 from France , 13 from the Czech Republic , and 6 from Germany ) . A previous study from Sweden contributed 397 families to the Northern European study group [17] . This included 272 complete affected sib pair families with AD diagnosed according to the U . K . Working Party’s Diagnostic Criteria [46] . The remaining 125 families were incomplete nuclear families including mother-child or father-child pairs . In all family-based studies , information regarding the parental history of AD was obtained by a questionnaire at the time of family recruitment . The analysis shown in S6 Table was performed in families in which both parents had a negative history of AD . Families in which one or both parents had a positive or unknown disease history were excluded . As described below , the analytical methods used allowed the incorporation of unrelated individuals to increase power . Thus , the Central European study also included previously published FLG genotypes of 772 unrelated German AD cases and 373 German controls from a previous GWAS [43] . In addition , we genotyped FLG mutations in 375 German children with AD diagnosed at a tertiary care center for pediatric allergy at Charité Universitätsmedizin Berlin . Also , previously published FLG genotype counts of 2 , 963 population-based German individuals from the International Study of Asthma and Allergies in Childhood II Study [47] were included . Likewise , in our Northern European study population we included genotypes of the 3 most prevalent FLG mutations ( c . 2282del4 , p . R501X and p . R2447X ) previously reported in the Swedish population-based BAMSE cohort [22] . The rarest FLG mutation ( p . S3247X ) was not available in the BAMSE dataset . Data on specific allergic sensitization was available in a large proportion of mothers from the GENUFAD and MAS studies . Plasma levels of specific IgE against grass and birch pollen , ribwort , cat and dog dander , mold ( Cladosporidium herbarum , Alternaria tenuis ) , hen’s egg , cow’s milk , fish , peanut , and house dust mite were determined using CAP-RAST-FEIA ( Pharmacia ) . A mother was defined as sensitized if specific IgE ≥0 . 7 kU/l ( CAP2 ) to at least one allergen was detected . Genomic DNA was prepared from whole blood by standard methods . The FLG c . 2282del4 variant was analyzed with fluorescence-based semiautomated genotyping [16] and the FLG p . R501X , p . R2447X and p . S3247X mutations with Taqman allelic discrimination ( Applied Biosystems , Foster City , California , USA ) as previously described [4] . Genotyping of p . R2447X in the Northern European families was performed using a fluorescent Kaspar assay ( KASP-By-Design genotyping assays , LGC group , Teddington , UK ) . Genotyping with Taqman and Kaspar was performed using a ViiA 7 Real-Time PCR System ( Applied Biosystems , Foster City , California , USA ) . When analyzing the rare p . R2447X and p . S3247X mutations , a sample known to be a mutation carrier was included on each genotyping plate as a positive control . The FLG mutations are named according to the nomenclature recommendations by den Dunnen and Antonarakis [48] . The positions of the mutations in the cDNA refer to the A of the ATG-translation initiation codon of NM_002016 . 1 . The PREMIM and EMIM tools were used to test for imprinting and maternal genotype effects of the FLG mutations [18 , 19] . First , the PREMIM tool was used to classify each trio according to the number of copies of FLG mutations carried by mother , father , and affected child . Incomplete nuclear families , unrelated AD cases , and population-based samples were also included in the analysis to increase the power to detect parent-of-origin effects [18 , 19] . Since the EMIM analysis is based on the assumption that the genotype frequencies in controls correspond to those in the general population , all population-based controls were included in the analysis irrespectively of disease status . Starting values for allele frequencies of the FLG mutations in the study population ( including controls ) were estimated with PREMIM ( –a option ) . The trios were analyzed using the EMIM tool , which uses a multinomial modelling approach to estimate genotype relative risk parameters on the basis of observed counts of genotype combinations in case-parent trios . The following parameters influencing the disease risk in the child were modelled with EMIM: R1 ( R2 ) , the factor by which an individual’s disease risk is multiplied if they carry one ( two ) risk alleles at a given locus . S1 ( S2 ) , the factor by which an individual’s disease risk is multiplied if the mother carries one ( two ) risk alleles at that locus . Im ( Ip ) the factor by which an individual’s disease risk is multiplied if inheriting a risk allele from the mother ( or father ) . γ11 ( interaction term ) , the factor by which an individual’s disease risk is multiplied if both mother and child have 1 copy of the risk allele . Previous data indicated that FLG mutations do not fit an additive genetic model , since the risk of AD in the homozygous carriers is too high . Thus , instead of using the default EMIM settings assuming an additive model we choose to independently estimate the R1 and R2 parameters . Modelling of maternal genotype effects was done using the default additive model . All analyses were performed under the “conditional on exchangeable parental genotype” ( CEPG ) assumption . This assumption should protect from potential biases in parameter estimation due to the inclusion of families recruited through multiple affected individuals , at the cost of reduced power to detect parent-of-origin effects compared to assuming only Hardy-Weinberg equilibrium [18 , 19] . A step-by-step analysis was performed by including additional risk parameters in the model as indicated in S9 Table . Maximum likelihood estimates were obtained from each model and a likelihood ratio test was performed to assess the significance among nested models . Note that it is not possible to directly compare the MCG and Im models in a likelihood ratio test , since they are not nested . However , they can be compared indirectly by comparison to the full model ( Table 4 ) . In order to test for genetic interaction between the child and the maternal genotypes we included and interaction term in the model ( γ11 ) . This parameter estimates the factor by which an individual’s disease risk is multiplied if both mother and child have one copy of the risk allele . A likelihood ratio test comparing the Maternal Child Genotype ( MCG ) and the MCG-Interaction model was then performed ( see S10 Table ) . A meta-analysis of the results from the Central and Northern European populations was performed using METAL [23] . The inverse variance method was used and the corresponding betas and standard errors were obtained from the EMIM summary file . The meta-analysis was performed on the single risk parameters estimates ( R1 , R2 , S1 or Im ) from each model . An analysis of heterogeneity was also performed with METAL in order to evaluate if the observed effect sizes were homogeneous across datasets . In order to increase power in the main analysis we allowed the inclusion of all affected offspring available in each family which were considered as independent trios ( -xa option in PREMIM ) . This may lead to biased results when using large pedigrees but it is unlikely to have a large effect in our study since most families had only 1 or 2 affected children . In order to exclude this potential bias we repeated the analysis including one affected child per family ( omitting the −xa option in PREMIM; S8 Table ) . The Transmission Disequilibrium Test ( TDT ) was performed with PLINK [49] . In order to account for multiple affected offspring within families , empirical p-values were calculated with the —tdt —perm option , which flips the allele transmitted from parent to offspring with 50:50 probability . Allelic effects were calculated with PLINK —assoc using the offspring of the +Mat_Sens as controls and those from the −Mat_Sens as cases . Haplotype frequencies on the central European Study were calculated with FAMHAP [50] , which computes maximum-likelihood estimates obtained with the expectation-maximization algorithm .
|
Most human diseases are caused by a combination of multiple environmental and genetic influences . The widely used case/control approach aims to identify disease risk genes by comparing the genetic constitution of affected and healthy individuals . Although successful , this approach ignores additional mechanisms influencing disease risk . Here , we studied mutations in the filaggrin gene ( FLG ) , which are strong risk factors for atopic dermatitis ( AD ) and allergies , in a large number of families with AD . We found that FLG mutations in the mother , not the father , increased the AD risk of the children , even if the child did not inherit the mutation . Thus , our study revealed , for the first time , a direct influence of a maternal mutation on the child’s risk for a common disease . The maternal FLG effect was only found when the mothers were allergic , and was absent in families of non-allergic mothers . This finding suggests that FLG-induced changes in the maternal immune response shape the child’s immune system during pregnancy and increase the child’s risk for AD . Our study indicates that maternal FLG mutations act as strong environmental risk factors for the child and highlights the potential of family-based studies in uncovering novel disease mechanisms in medical genetics .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[] |
2015
|
Maternal Filaggrin Mutations Increase the Risk of Atopic Dermatitis in Children: An Effect Independent of Mutation Inheritance
|
The two available drugs for treatment of T . cruzi infection , nifurtimox and benznidazole ( BZ ) , have potential toxic side effects and variable efficacy , contributing to their low rate of use . With scant economic resources available for antiparasitic drug discovery and development , inexpensive , high-throughput and in vivo assays to screen potential new drugs and existing compound libraries are essential . In this work , we describe the development and validation of improved methods to test anti-T . cruzi compounds in vitro and in vivo using parasite lines expressing the firefly luciferase ( luc ) or the tandem tomato fluorescent protein ( tdTomato ) . For in vitro assays , the change in fluorescence intensity of tdTomato-expressing lines was measured as an indicator of parasite replication daily for 4 days and this method was used to identify compounds with IC50 lower than that of BZ . This method was highly reproducible and had the added advantage of requiring relatively low numbers of parasites and no additional indicator reagents , enzymatic post-processes or laborious visual counting . In vivo , mice were infected in the footpads with fluorescent or bioluminescent parasites and the signal intensity was measured as a surrogate of parasite load at the site of infection before and after initiation of drug treatment . Importantly , the efficacy of various drugs as determined in this short-term ( <2 weeks ) assay mirrored that of a 40 day treatment course . These methods should make feasible broader and higher-throughput screening programs needed to identify potential new drugs for the treatment of T . cruzi infection and for their rapid validation in vivo .
Chagas disease , caused by the protozoan parasite Trypanosoma cruzi , is the leading cause of cardiac disease in many countries of Latin America . The World Health Organization has estimated that 16–18 million people are currently infected and 90 million are at risk of acquiring the infection [1]–[2] . Both of the available compounds for treatment , benznidazole ( Radanil® , Roche , Rio de Janeiro ) and nifurtimox ( Lampit® , Bayer , Leverkusen ) , have potential side effects , require long courses of treatment , and exhibit variable efficacy [3]–[6] . Therefore , there is an urgent need to discover new treatment options with reduced toxicity . Several compounds have been proposed as new therapies for T . cruzi infection , however , few have moved beyond the candidate stage . The development of in vitro and in vivo high-throughput assays for the screening of anti-T . cruzi compounds is essential . Epimastigotes of T . cruzi can be easily obtained in abundance from axenic culture and drug efficacy determined using a variety of approaches , including manual , spectrophotometric [7]–[11] or fluorometric [12]–[13] assessment of parasite growth . In addition to the fact that these epimastigote-based assays may not truly reflect the effectiveness of compounds on the life cycle stages of T . cruzi that are present in mammals ( the extracellular trypomastigotes and intracellular amastigotes ) [14] , these assays may be laborious and difficult to scale up for high-throughput screening ( HTS ) [15]–[16] . Assays to detect drug susceptibility of the more appropriate T . cruzi life cycle stage , the intracellular amastigotes , have been modified to use parasites expressing the Escherichia coli β-galactosidase gene ( lacZ ) [15] and this assay has also recently been scaled up for HTS [17] . However one downside of this method is that it is a single endpoint assay requiring post-assay processing to interpret . With respect to in vivo drug testing of anti-T . cruzi compounds , the vast majority of studies use a mouse model system where the compounds are administered early in the acute phase of the infection [18]–[20] , with the main criteria of treatment efficacy based on the suppression of acute parasitemias , the measurement of mortality rates post infection , and/or on the use of parasite detection techniques that frequently yield negative results ( i . e . fail to detecte parasies ) even in the absence of treatment [6] , [18] , [20]–[22] . In this study we describe the generation of T . cruzi parasite lines that express either the firefly luciferase ( luc ) or the tandem tomato fluorescent protein ( tdTomato ) and the use of these lines to establish accurate and simple in vitro as well as in vivo systems to screen and test anti-T . cruzi compounds . tdTomato red fluorescent parasites were detectable by microscopy and flow cytometry and their fluorescence intensity was easily quantified using a fluorescence plate reader . Moreover , the replication of epimastigotes or amastigotes could be monitored at multiple time points over the culture period rather than at endpoint , providing a more accurate assessment of parasite growth kinetics . Bioluminescent as well as fluorescent parasites were detectable via in vivo imaging after infection in mice and their expansion was used to rapidly assess the in vivo efficacy of anti-T . cruzi compounds . These results highlight the use of the methods described here as powerful new tools for the more rapid and efficient high-throughput screening of potential trypanocidal drugs in vitro and in vivo .
All animal protocols were approved by the University of Georgia Institutional Animal Care and Use Committee . Epimastigote forms of T . cruzi CL WT or tdTomato strain were cultured at 27°C in supplemented liver digest-neutralized tryptose ( LDNT ) medium as described previously [23] . After 3–4 days in LDNT , epimastigotes were submitted to a stress in triatome artificial urine ( TAU ) medium for 2 h [24] . Then , parasites were incubated in TAU3AAG medium [24] for 6–7 days , at the end of which the highest number of metacyclic trypomastigote was generally obtained . Parasites were then incubated in complement-active FBS to lyse the remaining epimastigotes ( by complement activity ) . Vero cell monolayer cultures were infected with metacyclics to generate trypomastigotes to use in the amastigote growth inhibition assays . Infected Vero cells were cultured in RPMI 1640 medium with 10% fetal bovine serum ( FBS ) in a humid atmosphere containing 5% CO2 at 37°C . To generate T . cruzi parasites expressing tandem dimeric tomato red fluorescent protein [25] we constructed a plasmid using the expression vector pTrex as a backbone [26] . The 1464 bp tdTomato gene was generated by PCR amplification using the primer set 5′-AGAATTCATGGCGCCTAGGGTGAGC-3′ ( forward ) and 5′-TACGTCGACTTAGAGCTCGATATCGACG-3′ ( reverse ) , and the pCTR2t vector as the template [27] . The forward primer has an EcoRI site and the reverse primer has a SalI site . This PCR fragment was digested and cloned downstream of the rRNA gene promoter and the HX1 fraction into EcoRI and SalI sites in the multi-cloning site of the pTREX plasmid , generating the pTREX-tdTomato plasmid ( Figure 1A ) . Transgenic parasites were generated as previously described [28] . A total of 1×107 early-log epimastigotes were centrifuged at 1 , 620 g for 15 min and suspended in 100 µl Human T Cell Nucleofector™ Solution ( Lonza , Cologne ) at room temperature . The resuspended parasites were then mixed with 10 µg DNA in a total volume of 10 µl and electroporated using the program “U-33” in an AMAXA Nucleofector Device ( Lonza ) . The electroporated parasites were then cultured in 25 cm2 culture flasks ( Corning Incorporated , Lowell , MA ) with 10 ml LDNT medium and 300 µg/ml G418 was added at 24 h post-transfection . Parasite cultures are monitored by flow cytometry frequently for loss of fluorescence , and when necessary , sortedto remove non-fluorescent contaminants using a MoFlo cytometer ( DakoCytomation , Carpinteria , CA ) with an Enterprise 631 laser tuned to 488 nm for excitation and an emission filter with a band pass of 570/40 nm . Multiple strains of T . cruzi ( CL , Tulahuen , Brazil , Colombiana and TCC ) expressing the tdTomato protein were generated and the CL tdTomato strain used in the experiment shown here unless otherwise noted . To generate bioluminescent T . cruzi parasites , we constructed a plasmid based on Multi-Gateway technology ( Invitrogen , Carlsbad , CA ) . The vector pLew90β was modified to include the Gateway cloning cassette by first inserting a new 62 bp multiple cloning site ( CCCGAG ( AvaI ) ACCGGTGTTAACGGGCC CTTCGAAGGTACCGAGCTCTTAATTAAGATATC ( SpeI ) ACTAGT ) into pLew90β at the Ava I and Spe I sites downstream of the Tc Beta Tubulin gene . The Multi-Gateway cloning cassette was digested from the pDEST-R4R3 ( Invitrogen , Carlsbad , CA ) plasmid using Pci I . A Klenow reaction was used to fill in the Pci I overhang . The fragment was then digested with BsaHI and cloned into pLew90β at HpaI/BstBI sites in the new multiple cloning site to create the pLew90β-GW-T7 destination plasmid . The combined T7 promoter and PARP splice acceptor site for the luciferase gene was cloned into the pDONR-P4P1R plasmid using a synthetic DNA construct in the pPCR-Script vector ( GeneArt , Burlingame , CA ) . This fragment was digested from the pPCR-script vector with KpnI and Hind III ( flanking sites underlined ) then recombined into the pDONR-P4P1R ( Invitrogen , Carlsbad , CA ) plasmid utilizing the Gateway ( Invitrogen , Carlsbad , CA ) Att recombination sites ( bold italics; GGTACCGGGGACAACTTTGTATAGAAAAGTTGGGAGCTCGTAATACGACTCACTATAGGGCGAATTGGATCCTGCACGCGCCTTCGAGTTTTTTTTCCTTTTCCCCATTTTTTTCAACTTGAAGACTTCAATTACACCAAAAAGTAAAATTCACAAGCAAGTTTGTACAAAAAAGCAGTCCCCAAGCTT ) was digested . The luciferase gene was amplified as an 1 . 8kb PCR product from the pPT7:Otet/Luc vector ( a gift from John E . Donelson , University of Iowa , Iowa City , IA ) . The primers used were flanked by the appropriate Gateway Att recombination sites ( bold italics ) and were as follows: luciferase sense primer; GGGACAAGTTTGTACAAAAAAGCAGGCTCAACCATGGfoAAGACGCCAAAAACATAAAGAA , luciferase antisense primer; GGGGACCACTTTGTACAAGAAAGCTGGGTCTTACACGGCGACTTTTCC GCCCTT CTTGG . The luciferase PCR fragment was then cloned into the pDONR-221 vector ( Invitrogen , Carlsbad , CA ) utilizing the Gateway Att recombination sites . A third vector was generated with the 3′UTR for the luciferase gene by amplifying the T . cruzi Aldolase 3′UTR sequence from the original pLew90β plasmid with flanking Gateway Att recombination sites ( bold italics ) : Aldolase 3′UTR sense primer , GGGGACAGCTTTCTTGTACAAAGTGGGGTCTTAAGGATCCTGCCCATT , Aldolase 3′UTR antisense primer , GGGGACAACTTTGTATAATAAAGTTGTGCCCGGGCTCGAATCCC CCC . The Aldolase 3′UTR fragment was then cloned into the pDONR-P2RP3 vector ( Invitrogen , Carlsbad , CA ) utilizing the Gateway Att recombination sites . Finally , the pLew90β-GW-T7 ( Destination ) plasmid , pDONR-P4P1R/T7/PARP SAS , pDONR-221/luciferase , and pDONR-P2RP3/Aldolase 3′UTR plasmids were combined as per the Multi-Gateway protocol ( Invitrogen Carlsbad , CA ) to yield a final plasmid construct of pLew90β-GW/T7/PARP SAS/luciferase/Aldolase 3′UTR ( Figure 1B ) . This vector was transfected into T . cruzi epimastigotes using the same protocol as described above for the tdTomato vector . tdTomato epimastigotes ( 1×104 parasites/well ) were plated in 96 well Costar black plates ( Corning Incorporated , Corning , NY ) with or without drug and the change in fluorescence intensity was measured as surrogate of growth using a fluorescence plate reader ( SpectraMax M2 , Molecular Devices , Sunnyvale , CA ) daily for 4 days . Benznidazole ( Radanil® , Roche , Rio de Janeiro ) tablets ( 100 mg ) were pulverized , dissolved in nanopure ( Milli-Q ) water , then filter sterilized and used as the reference trypanocidal drug ( positive control ) . EXO2 derivatives [29]–[32] ( EXO2-04 , EXO2-12 , EXO2-17 and EXO2-36 ) were kindly provided by Jose C . Aponte and Gerald B . Hammond ( University of Louisville , Louisville , KY ) and were dissolved in DMSO , with a final concentration containing less than 0 . 1% DMSO per well . The 50% inhibitory concentration ( IC50 ) values were determined by linear regression analysis on day 3 of culture . Vero cells were exposed to 2000 rad of gamma radiation for ten minutes [33] and 1 . 7×104 cells were plated in 96-well plates overnight at 37°C/5% CO2 . Vero cells were exposed to trypomastigotes of the CL tdTomato strain of T . cruzi for approximately 5 h at a multiplicity of infection ( MOI ) of 10 . After infection , cell cultures were washed to remove non-internalized parasites and fresh media with or without benznidazole ( as the reference trypanocidal drug ) or test compounds were added . The change in fluorescence intensity was determined as a measurement of growth over 3–4 days of culture . The 50% inhibitory concentration ( IC50 ) values were determined by linear regression analysis at day 3 . Balb/c mice were purchased from the National Cancer Institute ( Frederick , MD ) and maintained in the University of Georgia animal facility in microisolator cages under specific pathogen-free conditions . For the short term in vivo assay mice were infected in each pad of both hind feet with either 2 . 5×105 of T . cruzi trypomastigotes ( CL strain ) expressing the tdTomato protein or with 1×105 trypomastigotes expressing luciferase . Mice were then submitted to benznidazole ( BZ ) , ENH-5 and NTLA-1 ( nitrotriazole derivatives; the gift of Maria Papadopulou , Northwestern Medical Center ) , Posaconazole , an antifungal triazole derivative ( POS ) [34] , EXO2-04 , or BIS767 ( a bisphosphonate; the gift of Melina Galizzi and Roberto Docampo ) , treatment from day 6 to day 11 ( tdTomato parasites ) or from day 4 to day 10 ( luciferase parasites ) post infection . An untreated control , as well as a naïve non-infected group of mice was also monitored in each experiment . BZ was prepared by pulverization of one tablet containing 100mg of the active principle , followed by suspension in distilled water . BZ was administered orally , with daily doses of 100 mg/kg body weight . ENH-5 ( 20mg/kg/day ) and NTLA-1 ( 2mg/kg/day ) were suspended in PBS and daily injected intraperitoneally ( i . p . ) into mice while EXO2-04 ( 20mg/kg/day ) was suspended in 1% DMSO and administered to the mice i . p . POS was dissolved in an aqueous solution of 2% methylcellulose and 0 . 5% Tween 80 and administered orally , with daily doses of 20 mg/kg body weight . BIS767 was suspended in PBS and daily injected subcutaneously into mice . The fluorescence and bioluminescent intensity ( photons/cm2/sec ) was measured before and after treatment as a surrogate of parasite load at the site of infection . For the long term in vivo assay mice were infected i . p . with 1000 trypomastigotes of CL strain of T . cruzi and sacrificed by CO2 inhalation at different time points post infection . Infected mice were divided into the following groups: mice infected without specific treatment ( untreated ) ; mice treated orally with BZ , with daily doses of 100 mg/kg body weight for 40 days ( 15 to 55 dpi ) ( BZ40 ) , mice treated with daily i . p doses of 2mg/kg/day of NTLA-1 for 50 days ( 15–65 dpi ) , mice treated orally with POS for 40 days ( 15–55 dpi ) with daily doses of 20mg/kg/day ( POS ) and mice treated subcutaneously with BIS767 for 30 days ( 15–45 dpi ) with daily doses of 500µg/kg/day ( BIS767 ) . Prior to bioluminescent imaging , mice were anaesthetized with 1 . 5% isofluorane and then injected with 150 mg/kg body weight of substrate luciferin potassium salt ( Caliper , Hopkinton , MA ) dissolved in PBS was administered by a single i . p . injection . Mice were imaged in a bioluminescent imaging system ( IVIS 100; Xenogen , Alameda , CA ) . Briefly , mice were placed into the camera chamber , where a controlled flow of 1 . 5% isofluorane in air was administered through a nose cone via a gas anesthesia system . The luciferin substrate was allowed to adequately disseminate in the mice [35] for 10 min before imaging . Mice were imaged in dorsal , position by capturing a grayscale body image overlaid by a pseudocolor image representing the spatial distribution of the detected photons . Images were collected with 5 min integration time . Data acquisition and analysis were performed by using the Living-Image software ( Xenogen , Alameda , CA ) where luminescence could be quantified as the sum of all detected photon counts per second within a chosen region of interest . For in vivo fluorescent imaging , footpads of mice subcutaneously infected with 2 . 5×105 tdTomato parasites in 4 ul were imagined every other day using the Maestro2 In Vivo Imaging System ( CRi , Woburn , MA ) with the green filter set ( acquisition settings: 560 to 750 in 10 nm steps; exposure time 88 . 18 ms and 2×2 binning ) . Collected images were unmixed and analyzed with the Maestro software v2 . 8 . 0A . Statistical analysis was performed by ANOVA and unpaired T test , using the GraphPad PRISM 3 . 0 software . Differences between two groups were considered significant at p<0 . 05 .
The tdTomato gene used in these studies was generated by genetically fusing two copies of the gene encoding the Tomato Red Fluorescent protein to create a tandem dimer ( td ) named tdTomato [25] . This construct possesses many desirable properties relative to previously studied fluorescent proteins , including a faster and more complete maturation and increased brightness [25] , [35] . In addition , the tdTomato fluorescence signal can be detected at ≥620 nm ( outside the range of of the bulk of animal tissue autofluorescence ) , which minimizes autofluorescence and greatly increases the penetration of fluorescence signals into the tissues [25] , [36] . The expression vector pTrex-Neo [26] , carrying the T . cruzi ribosomal promoter , has been shown to provide strong protein expression in T . cruzi and to enable the stable integration of exogenous genes into the ribosomal locus [26] , [37]–[38] . Thus the gene encoding tdTomato was cloned into pTrex-Neo and the pTREX-Neo-tdTomato ( Figure 1A ) construct transfected into epimastigotes of various T . cruzi strains , which were then drug-selected as described in the Materials and Methods . Parasite fluorescence was monitored by microscopy and flow cytometry ( Figure 2A–B ) . Transfectant parasites expressing the tomato protein showed a bright red fluorescence distributed throughout the cell in all life cycle stages ( Figure 2A ) . Moreover , the fluorescence was stable in the absence of antibiotic pressure for >5 months ( Figure 2B ) . Extracellular epimastigotes are easily obtained and thus are often used as a first step in the screening of new potential anti-T . cruzi drugs in vitro [39] . To address whether fluorescent tdTomato parasites were useful to screen for such compounds , we plated tdTomato epimastigotes with or without the proven anti-T . cruzi compound benznidazole ( BZ ) and the fluorescence intensity as a representative of parasite growth was measured daily . A dose dependent decrease in parasite growth was evident within 2 days of addition of BZ to cultures and was clearly evident throughout the remainder of the 4 day assay period ( Figure 3A ) . The IC50 calculated for benznidazole is similar to that of other methods , including visual counting ( Figure 3A ) , and as previously reported in the literature for several T . cruzi strains [40]–[41] . This method showed low intra-assay variation and a very good consistency by the inter-assay analysis ( Figures 3B ) . The analysis of the test EXO2 compounds ( EXO2-04 , EXO2-12 , EXO2-17 and EXO2-36 ) revealed one with IC50 below than that of BZ ( Figure 3C ) . The ability of compounds to inhibit the intracellular growth of T . cruzi amastigotes is a more rigorous and relevant test of anti-T . cruzi activity , as it is applied to a stage which is the predominant form in mammals and because the killing assay requires that drug also cross the host cell membrane . Amastigote growth assays in Vero cells worked similarly to epimastigotes assays , with parasite fluorescence increasing over the 4 day culture period . BZ exhibited a dose-dependent inhibitory effect on parasite growth ( Figure 4A ) . Furthermore , this assay produced comparable results in 96 or in 384 well plates ( Figure 4B–C ) , demonstrating the potential utility for the testing of large compound libraries . Lastly , the expression of tdTomato in different parasite strains which may differ in susceptibility to various drugs provides an easy method to confirm the susceptibility of multiple parasite strains to potential anti-T . cruzi compounds ( Figure 4C ) . We next explored the utility of T . cruzi lines expressing firefly luciferase or tdTomato for assessing the activity of compounds on the in vivo growth and survival of T . cruzi . In order to optimally visualize parasite development as well as to simulate a normal route of infection , parasites were delivered subcutaneously into the footpad of mice . Mice were then left untreated or were treated with various compounds starting at 6 days post-infection for those infected with tdTomato parasites or 4 days post infection for mice infected with bioluminescent parasites . These time points were chosen as they allowed sufficient time for establishment of the infection and easy visualization using the respective imaging technique . In addition to the proven compound BZ , a number of other compounds were tested with either assay , including EXO2 which had shown excellent IC50 in the in vitro assay ( Figures 3 and 4 ) , NTLA-1 and ENH-5 , two nitrotriazole derivatives also previously demonstrated to have high in vitro activity against T . cruzi amastigote growth ( M . Papadopoulou personal communication ) , a bisphosphonate ( BIS767 ) with in vitro anti-T . cruzi activity ( Galizzi and Docampo personal communication ) and the antifungal posaconazole ( POS ) , which has been reported to have in vivo trypanocidal activity [22] , [42] . In all cases the compounds were delivered systematically , by oral gavage in the case of BZ and POS and by i . p . injection in the cases of the EXO2-04 , NTLA-1 and ENH-5 and by a subcutaneous injection in the ventral abdomen in the case of BIS767 . Both the tdTomato and the luc-based assays revealed a rapid in vivo parasite clearance activity for BZ and POS , with dramatic control of parasite load within 1 day with BZ and slightly longer in the case of POS ( Figures 5 and 6 ) . In contrast , none of the other test compounds exhibited significant effects on in vivo parasite growth , despite their previously demonstrated in vitro anti-T . cruzi activity . In previous work , we showed that BZ treatment in the acute or chronic phase of the infection can provide cure of mice infected with T . cruzi [43] . We sought to explore whether the same compounds used in the in vivo short term assays were effective to cure mice infected with T . cruzi . To address this question , C57BL/6 mice were infected with the wild-type CL strain of T . cruzi and treated with either BZ , POS , NTLA-1 or BIS767 or left untreated . All mice exhibited detectable parasitemias by day 14 post-infection; in untreated mice this acute phase parasitemia peaked at 21 dpi and became undetectable by approximately 35 dpi ( Figure 7A ) . All of the compounds evaluated in this study suppressed parasitemia , which became undetectable in all cases by 21 dpi ( Figure 7A ) . However , only the mice treated for 40 days with BZ , the majority of POS-treated ( 90% ) and a small fraction of NLTA-1-treated mice ( 20% ) were able to clear the infection and cure , as by the failure to detect parasitemias after cyclophosphamide ( cy ) immunosuppression ( Figure 7B ) . These results demonstrate that suppression of parasitemia soon after drug-treatment initiation Figure 7A ) is a poor indicator of drug efficacy and the potential for a compound to achieve parasitological cure over a long-term course of treatment . Moreover , the results of the long-term-treatment assay and the short-term in vitro screens using measurement of parasite growth in the foot-pad following the injection of luminescent or fluorescent T . cruzi ( Figures 5 and 6 ) are perfectly concordant , suggesting that the short-term in vivo assay is strong predictor of in vivo drug efficacy and clearly superior to measuring suppression of parasitemia .
There is a clear need for new compounds to treat T . cruzi infection at all stages of the infection/disease [3] , [6] , [44]–[45] . One crucial limitation in the identification of such compounds is the lack of adequate and efficient in vitro and in vitro tests of drug efficacy . Various colorimetric and fluorometric assays have been developed to screen for anti-T . cruzi compounds [7]–[12] which are more accurate and objective than the microscopic visual counting . More recently , an efficient method to quantify T . cruzi parasites in drug screening assays using genetically engineered parasites that express the Escherichia coli β-galactosidase gene , lacZ has been described [15] . Although widely and effectively used and even scaled to a 384 well format ( Ana Rodriguez , personal communication ) , this method depends on a single endpoint reading following the addition of detergent and a convertible substrate . T . cruzi lines expressing alternative reporters such as luciferase [46] , fluorescent proteins [37] , [47] have also been produced , but the use of these parasites lines for in vitro and in vivo screening of potential anti-T . cruzi compounds has not been reported . Herein , we report the generation of stable T . cruzi lines constitutively expressing the tandem tomato fluorescence protein ( tdTomato ) or the firefly luciferase protein and the use of these lines to screen for anti-T . cruzi compounds in vitro and to confirm their activity in vivo . tdTomato-expressing parasites showed strong fluorescence in all life stages including when the parasites were maintained in the absence of antibiotic selection . Among the advantages of the use of these fluorescent parasites for in vitro assays is the elimination of the need for fixation or cell permeabilization , and detection of fluorescence with minimal handling using a fluorimeter , microscope-based screener or other imaging system . The use of fluorescent parasites also allows for the continuous measurement of parasite replication during the experiment , making possible the analysis of the parasite growth and/or the initiation and cessation of growth inhibition by drugs over time . The fluorescence-based growth inhibition assays of both epimastigotes and intracellular amastigotes had high intra- and inter-assay consistency and could be easily scaled to at least a 384 well format , allowing for the development of high throughput screening of large compound libraries . The assays also work well in multiple T . cruzi strains; to date we have produced TdTomato expressing parasites from all 4 T . cruzi lines in which we have attempted . While the in vitro drug screening assay is a significant improvement in methodologies for testing of anti-T . cruzi compounds , the use of T . cruzi lines expressing luciferase and tdTomato for in vivo compound screening bridges a more significant technical gap and thus represent a much bigger step forward . The current standard for the in vivo testing of anti-T . cruzi compounds is the measurement of the suppression of parasitemia in the acute phase of the infection or the protection from acute-phase death , neither of which assess parasitological cure [18]–[22] . We have recently developed an in vivo treatment protocol that critically evaluates parasitological cure by using suppression of immune responses following drug treatment , a procedure that releases constraints on parasite growth and thus reveals drugs and treatment regimens that fail to completely cure the infection [43] . In this study , we have used this immunosuppression technique as the gold standard for drug-induced cure . However , the considerable drawback of this protocol is the long course of the experiment , including ∼40 days for compound treatment and a minimum of 75–80 days for the entire experiment . Herein we demonstrate that a short-term treatment protocol , involving 5–6 days of treatment and a total of 11–12 days for the full experiment , that assesses the suppression of parasite replication ( although not full clearance ) at subcutaneous sites . The assay is not only quick , but is also quite easy , requiring short-term compound dosing and occasional animal imaging . Most importantly , for the limited set of compounds tested to date , this short-term assay using either bioluminescent or fluorescent parasites gives results that are nearly identical to those obtained using the long-term treatment protocol followed by immunosuppression; In both assays BZ is most highly effective compound of those tested and POS is slightly less effective . A particularly notable point from the results of these different screening assays is that multiple compounds that strongly suppress epimastigote and amastigote growth in vitro , and furthermore , rapidly and significantly suppress the level of parasites in the blood of acutely infected mice , fail to cure mice in the long term treatment assays or to suppress parasite replication in the tissues in the short-term assay . Although more compounds will have to be compared using both the short- and long-term in vivo protocols used herein , the results to date are highly encouraging that the short-term assay measures a drug effect that is more like what is required for parasitogical cure and thus is a better screen for in vivo efficacy than the acute suppression of parasitemia . Certainly , these two assays are likely measuring very different parameters , with suppression of parasitemia highlighting drugs that may have anti-parasite effects largely or solely on extracellular parasites . It is especially interesting that compounds that clearly suppress intracellular parasite replication in vitro fail to control parasite growth in the short term in vivo assays , despite the fact that the parasites are largely intracellular during these short-term in vivo tests ( as they would have to be if they are increasing in number as indicated by the increase in fluorescent signal over time ) . Although the short-term in vivo assay is quick and straightforward , there may also be room for additional improvements . A common limitation of these assays is the inability of fluorescent or bioluminescent signals to easily penetrate animal tissues , including fur . For this reason , we monitored parasite growth in subcutaneous hairless sites ( footpads ( shown here ) ) and ears ( data not shown ) , although luciferase-expressing parasites can also be visualized throughout the body when infections are given at higher doses and/or at later time points in the infection ( data not shown and [46] ) . The sensitivity of detection of the luciferase-expressing parasites is greater than that of the tdTomato parasites . However this advantage is surpassed by the fact that visualization of the tdTomato parasites does not require the repeated injection of the luciferase substrate nor a requirement to control for the time post-substrate injection that the imaging is done . Given that the beneficial effects of both BZ and POS are obvious within a day or 2 post-administration , it is also possible that the identification of drugs equivalent in efficacy to BZ may require only a single administration of the compound and only one or two imagining time points . These modifications would make it possible to screen 100's of compounds in vivo in very short order . Compounds that give a positive result could then be submitted to the longer-term treatment/immunosuppression assay . The focusing of parasites largely at the site of infection was also evident in these experiments and contributed to the success of this technique for assessing drug treatments . The imaging of the luciferase or tdTomato parasites is not sufficiently sensitive to detect the movement of small number of parasites from the subcutaneous injection point to other distant sites . However it is clear that the vast majority of parasites remain at the injection site , where they infect cells and replicate . Interestingly , even in the absence of drug treatment , parasite load at the site of injection diminishes after ∼10 days post infection . Preliminary results suggest that this reduction is immune mediated , as it does not occur in mouse strains with defects in T cell responses , and suggest that whole animal imagining of parasite growth and dissemination might also be useful in the study of immune control of T . cruzi at tissue sites Padilla , et al , unpublished ) . The tools described in this paper were dependable , facile and low cost and provide new methods for the rapid screening of anti-T cruzi compounds in vitro and in vivo . These methods should make feasible the broader and higher-throughput screening programs needed to identify potential new drugs for the treatment of T . cruzi infection .
|
The treatment of Trypanosoma cruzi infection ( the cause of human Chagas disease ) remains a significant challenge . Only two drugs , both with substantial toxicity , are available and the efficacy of these dugs is often questioned – in many cases due to the limitations of the methods for assessing efficacy rather than to true lack of efficacy . For these reasons relatively few individuals infected with T . cruzi actually have their infections treated . In this study , we report on innovative methods that will facilitate the discovery of new compounds for the treatment of T . cruzi infection and Chagas disease . Utilizing fluorescent and bioluminescent parasite lines , we have developed in vitro tests that are reproducible and facile and can be scaled for high-throughput screening of large compound libraries . We also validate an in vivo screening test that monitors parasite replication at the site of infection and determines the effectiveness of drug treatment in less than two weeks . More importantly , results in this rapid in vivo test show strong correlations with those obtained in long-term ( e . g . 40 day or more ) treatment assays . The results of this study remove one of the obstacles for identification of effective and safe compounds to treat Chagas disease .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/protozoal",
"infections",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance"
] |
2010
|
In Vitro and In Vivo High-Throughput Assays for the Testing of Anti-Trypanosoma cruzi Compounds
|
Vibrio cholerae , the causative agent of the cholera disease , is commonly used as a model organism for the study of bacteria with multipartite genomes . Its two chromosomes of different sizes initiate their DNA replication at distinct time points in the cell cycle and terminate in synchrony . In this study , the time-delayed start of Chr2 was verified in a synchronized cell population . This replication pattern suggests two possible regulation mechanisms for other Vibrio species with different sized secondary chromosomes: Either all Chr2 start DNA replication with a fixed delay after Chr1 initiation , or the timepoint at which Chr2 initiates varies such that termination of chromosomal replication occurs in synchrony . We investigated these two models and revealed that the two chromosomes of various Vibrionaceae species terminate in synchrony while Chr2-initiation timing relative to Chr1 is variable . Moreover , the sequence and function of the Chr2-triggering crtS site recently discovered in V . cholerae were found to be conserved , explaining the observed timing mechanism . Our results suggest that it is beneficial for bacterial cells with multiple chromosomes to synchronize their replication termination , potentially to optimize chromosome related processes as dimer resolution or segregation .
The diversity of regulatory systems of DNA replication has been studied in multiple bacteria [1–4] . An especially interesting group of bacteria with regard to DNA replication are those with multiple chromosomes . While a single chromosome is the norm in bacteria , about 10% of species in a diverse set of phyla carry more than one chromosome [5] . The best studied system in this respect is that of V . cholerae , the causative agent of the cholera disease [6 , 7] . The genome of strain O1 El Tor N16961 is divided into two chromosomes of about 3 ( Chr1 ) and 1 Mbp ( Chr2 ) respectively [8] . Chr1 carries most of the essential genes [8 , 9] . Replication at the origin of Chr1 ( ori1 ) is initiated by the initiator protein DnaA , as is the case in almost all known bacteria [10] . Chr2 encodes its own initiator , RctB . [8 , 11] . Notably , no RctB-like proteins have yet been found outside the phylogenetic group of Vibrionales . The structure of its central two domains ( of four in total ) resembles that of several plasmid replication initiators [12 , 13] . RctB binds to a set of so-called iterons within ori2 to initiate replication [14 , 15] , which contain the sequence GATC , methylated at the adenine by the Dam methyltransferase [16] . Binding of RctB was shown to be specific for fully-methylated GATCs , which in conclusion renders Dam essential in V . cholerae , unlike in E . coli [17–19] . RctB also binds to another type of sequence , the so-called 39-mers , which are also located at ori2 [20] . However , the binding of RctB to the 39-mers does not activate replication as does its binding to the iterons; on the contrary , this suppresses initiation [20] . The balance between the activating and repressing action of RctB , in conjunction with a handcuffing mechanism , is thought to generate tight control of Chr2 replication in a cell-cycle-dependent manner [20 , 21] . It was found that the two chromosomes start replication with a time delay in between [22–25] . In the search for regulatory mechanisms of communication between the two chromosomes , it was found that Chr1 was insensitive to the blockage of Chr2 replication [26]: it was shown that Chr1 controls replication of Chr2 through a short sequence about 800 kbp downstream from ori1 [27] . This site was later named crtS , for ‘Chr2 replication triggering site’ [25] . Replication of crtS triggers the replication of Chr2 , which is initiated after a short delay [25] . Moving the crtS site to other positions on Chr1 led to a corresponding shift in Chr2 initiation . The mechanism underlying the triggering effect of crtS is not yet fully understood but might involve physical contacts that were observed to occur between crtS and ori2 [25] . Replication of Chr2 in V . cholerae starts after about two-thirds of Chr1 are replicated . This timing leads to termination of Chr2 replication at about the same time as that of Chr1 . To better understand the mechanism underlying this phenomenon , we investigate here if it is the Chr2 replication starting after two-thirds of Chr1 is replicated which is important to the cell , or if the orchestrated termination of both chromosomes is the driving force of evolutionary selection . To this end , we tested whether the V . cholerae paradigm applies to other species of the Vibrionaceae and derive general rules of replication control .
While early studies suggested a synchronous replication start of the two V . cholerae chromosomes , more recent studies support a time delay between Chr1 and Chr2 initiation [22 , 23 , 25 , 28] . In synchronized V . cholerae cell cultures , such a time delay should lead to a situation with only Chr1 replicating in all cells short after initiation and later Chr2 replication . However , to date no synchronization method for V . cholerae has been available . Here we test if a synchronization method established for Escherichia coli can be used to synchronize V . cholerae populations [29] . The method is based on the induction of the stringent response as a cellular answer to nutrient limitation . In E . coli , addition of serine hydroxamate ( SHX ) blocks re-initiation of DNA replication , while ongoing replication rounds are finished , leading to cells with fully replicated chromosomes . Transfer of the cells to SHX-free medium then leads to a synchronous re-start of DNA replication . The stringent response in V . cholerae can also be induced by SHX treatment [30] . Consequently , addition of 0 . 9 mg/ml SHX to an exponentially growing V . cholerae batch culture resulted in clear inhibition of growth ( Fig 1A ) . Flow cytometry analysis of the cellular DNA content shows asynchronous replicating cells before SHX treatment and cells with either 1+1 or 2+2 fully replicated chromosomes ( Chr1 and Chr2 ) after SHX treatment for 150 minutes ( Fig 1B ) . After transfer to growth medium without SHX , the DNA content of the cells increases gradually as would be expected for a synchronously replicating population ( Fig 1B and 1C ) . To analyze replication of the two chromosomes individually , we performed marker frequency analyses using high density custom microarrays . We used the genome sequence of strain N16961 as a reference , which is very similar to the strain A1552 used here . A1552 ( El Tor biotype , Inaba serotype ) is a pathogenic strain of V . cholerae that was isolated from a traveler to Peru in 1992 ( Strain DSM 106276 in the German Collection of Microorganisms and Cell Culture , DSMZ ) [31] . However , during initial experiments we observed patterns indicating some sort of chromosomal rearrangements within the V . cholerae A1552 in comparison to strain N16961 . For the latter , it was found that the strain used in most labs actually carries a chromosomal inversion between two operons encoding ribosomal RNAs ( [25] , Supporting S1A and S1B Fig . We used a set of PCRs to check potential additional inversions between ribosomal operons within the A1552 strain ( Supporting S2 Fig ) . Results suggested a secondary inversion between rRNA operons A and C . To support this finding , we sequenced the A1552 genome with a combination of Illumina short read and Pacific Bioscience long read sequencing ( GenBank Accession CP024867 and CP024868; see supporting S1 Text ) . Adjusting the genomic positions to the new genome sequence , we were able to follow the replication activity of the two V . cholerae chromosomes after release from the stringent response ( Fig 2A–2D ) . Indeed , Chr1 initiated replication first as seen by a higher copy number of genomic loci near the replication origin 10 . 5 minutes after shifting to SHX-free medium . At later time points , this region of higher copy number increased gradually in size , indicating bi-directional replication towards the terminus region . Chr2 replication was not detected until 28 minutes after release from stringent response . Notably , its replication was more difficult to detect due to its smaller size and the large integron not providing reliable copy number values . A stepwise function was fitted to a total of 13 copy-number plots of Chr1 to determine average replication fork positions at different time points ( Supporting S3 Fig ) [32] . Interestingly , one replication fork runs about 60 kb ahead of the other on Chr1 , similar to what was found for E . coli ( Fig 2E ) . Based on the progression of the replication forks , we calculated a replication rate of 22 kbp/min or 360 bp/s for the replication in V . cholerae ( Fig 2E ) . Clearly , the secondary chromosome in V . cholerae starts to replicate after about two-thirds of the primary chromosome are replicated , causing a synchronous termination of both chromosomes . Two different models could be derived from these observations for the DNA replication within the family of Vibrionaceae . First , it is of biological relevance to the cells to replicate two-thirds of the primary chromosome before starting replication of Chr2 . Second , it is of biological relevance to the cells to terminate the two chromosomes at approximately the same time . Here we wanted to test both hypotheses to pave the way for a general understanding of the delay in initiation timing of DNA replication of Chr2 within the Vibrionaceae . The replication start control model , in which replication of two-thirds of Chr1 is important before Chr2 replication is initiated , implies that in Vibrio species with smaller secondary chromosomes than that of V . cholerae , replication of Chr2 ends before that of Chr1 ( Fig 3A ) . In Vibrio species with larger secondary chromosomes , replication of Chr2 would terminate after Chr1 . In contrast , the replication end control model would imply that replication of a small secondary chromosome starts later than that of a larger one ( Fig 3A ) . To test both models we used a comparative genomics approach . Recently , replication of a sequence called crtS on Chr1 was found to trigger Chr2 initiation in V . cholerae [25 , 27] . If such a site also appears in other Vibrio species , its position on Chr1 could be used as proxy for the time of Chr2 initiation . Based on the sequence of the V . cholerae crtS site , we searched the database for similar sequences occurring only once per genome in Vibrionaceae and generated a multiple sequence alignment ( Fig 4A ) . The most conserved sequence parts were then used to find a set of 129 sequences and generate a corresponding sequence logo ( Fig 4B ) . To test experimentally , if initiation at secondary replication origins in Vibrio species other than Vibrio cholerae is triggered by crtS sites , we analyzed the replication of mini-chromosomes , each driven by one of eleven secondary replication origins from different species of the Vibrionaceae . For a corresponding mini-chromosome based on V . cholerae ori2 , it was shown that the copy number increases in an E . coli strain carrying a copy of the crtS site; this did not occur in a strain lacking crtS [27 , 33] . As readout for the replicon copy number , we measured how well each strain tolerated increased amounts of antibiotic ( Fig 5 , see Method section for details , [34] ) . This method is based on the logic that an increased replicon copy number correlates with an increased copy number of the resistance gene , and so correlates with a higher antibiotic tolerance [35 , 36] . A significant increase in copy number was observed for 8 out of 11 mini-chromosomes in an E . coli strain carrying the V . cholerae crtS site integrated into the chromosome in comparison to a strain without crtS ( Fig 5 , compare red and grey bars ) . Mini-chromosome copy number was similarly increased in strains carrying either the V . nigripulchritudo or the V . parahaemolyticus crtS . ( Fig 5 , green and blue bars ) . For two of the mini-chromosomes ( synVivuII and synVihaII ) , the copy number appeared to be high already in the strain without crtS and one mini-chromosome ( synPhopII ) showed no crtS-dependent copy number increase ( Fig 5 ) . In summary , the data showed that replication origins of secondary chromosomes in Vibrionaceae are triggered by crtS sites in general , suggesting this mechanism to be conserved . The data also suggested that crtS sites do not function specifically on the ori2 of their corresponding species , but appear to be interchangeable . To test the replication start and end control models , the position of the crtS sites on the primary chromosomes of 29 fully sequenced Vibrionaceae species was determined , and the relative distance to ori1 and ter1 calculated ( See method section for details ) . A correlation of the length of two-thirds of one Chr1 replichore to the distance of the crtS site to ori1 would be expected if the start control model for replication in Vibrionaceae held true ( Fig 6A , grey dots ) . However , the data from our comparative genomics approach showed no such correlation ( Fig 6A , black and red dots ) . To test the replication end control model , the distance of the crtS site to ter1 was plotted against the length of a Chr2 replichore ( Fig 6B ) . Here , the values derived from comparative genomics resembled the theoretical data quite well , where the shift between the two respective regression lines correlate with the delay between crtS replication and ori2 initiation observed in V . cholerae [25] . Our findings support the replication end control model to explain the replication timing of the two chromosomes in Vibrionaceae ( Fig 3 ) . To further test DNA replication in Vibrionaceae , we performed marker frequency analysis by next-generation sequencing of eleven different strains from the Vibrionaceae group . If the replication start control model holds true , one would expect copy numbers of ter2 to be higher than ter1 in Vibrio species where Chr2 is smaller than one-third of Chr1 . In species with Chr2 bigger than one-third of the corresponding Chr1 , the copy number of ter2 should be below that of ter1 ( Fig 3B ) . If the replication end control model was applicable , one would expect the copy numbers of ter1 and ter2 to be equal in all Vibrionaceae . Chromosomal DNA was isolated from exponentially growing cultures and from cells in stationary phase . The DNA samples were then analyzed by Illumina sequencing and copy numbers plotted according to chromosomal positions . Copy numbers of the two chromosomes were close to one in stationary phase and showed a flat distribution in most cases , as expected for non-replicating cells ( Supporting S5 Fig , supporting S4 Table ) . In all analyzed cases of exponentially growing cells , the copy number plots formed typical triangular shapes , with the replication origin at the highest point and copy numbers declining towards the termini ( Fig 7 ) . Note that the data were not normalized to the copy numbers of stationary phase cells . Two lines were fitted to each of the chromosomes and their intersection assigned as the minimal and maximal copy number as described [37] . The position of the maxima corresponded well with the positions of ori1 and ori2 with about 39 kbp deviation on average ( below 1% of genome size ) , supporting good data quality ( Supporting S2 Table ) . Also , data correlated well in biological replicates ( Supporting S6 Fig , supporting S2 and S3 Tables ) . The copy number of ori2 was lower than the copy number of ori1 in all strains , consistent with a conserved replication mode within the Vibrionaceae . In fact , the copy number of ori2 was also lower than that of the region carrying the corresponding crtS site on Chr1 in all studied strains , suggesting that crtS-based triggering of Chr2 replication is a conserved mechanism . The genomic plots showed the copy numbers of ter1 and ter2 within individual strains to be very similar , although some variation occurred ( Fig 7 ) . To test the two proposed models of replication start versus replication end control , we plotted the ter1/ter2 ratio for the analyzed strains ( Fig 7L ) . Values were around one , with some variation supporting the replication end control model . Notably , we found no good correlation between crtS position , Chr2 size and how well the ter1/ter2 ratio matches 1 . The ratio of copy numbers between the Chr1 position two-thirds of replichore size from ori1 and the ori2 copy number was higher , with a mean of 1 . 4 . This indicates that replication in Vibrionaceae does not follow the replication start model ( Fig 7L ) .
In V . cholerae , the secondary chromosome initiates its replication after about two-thirds of the primary chromosome has been replicated . As a consequence , the two chromosomes terminate at approximately the same time . It has been shown that this replication pattern can be changed by moving the crtS sites to other positions , either further away or closer to the terminus [25] . Such engineered strains have no dramatic deficiencies in cell viability . Why has evolution shaped the replication timing to be as it is found in V . cholerae and other species of Vibrionaceae ? We approached this question by asking if the selection pressure lies at the start or end of Chr2 replication timing ( Fig 3 ) . By analyzing replication rules in multiple Vibrio species , we show here that it is in fact the timing of the end of replication relative to Chr1 which is under selection , and not the start . In other words , the delay between the start of Chr1 and Chr2 replication respectively seems to be unimportant , but it appears to be more important that the two chromosomes terminate replication at approximately the same time . This begs the question of why synchronous termination of both chromosomes in Vibrionaceae is important . One reason could be the coordination of chromosome segregation and cell division . The chromosomal region opposite the replication origin is the part of the chromosome where dimer resolution occurs at the dif site [38] . In addition , chromosome segregation is coordinated with cell division through interactions of the Ter domain ( s ) with the divisome in E . coli , as well as in V . cholerae [39–41] . Interestingly , it was found that in engineered V . cholerae strains , in which Chr2 terminates long before Chr1 , the two copies of ter2 remain at the middle of the cell until cell division , and segregate approximately at the same time as ter1 , like in wildtype cells [25] . Cohesion of ter1 and ter2 with their respective sister ter sequences near the division site thus seems to be important for segregation and the synchronized termination of the two chromosomes might facilitate this mechanism . We could imagine an alternative explanation of why termination of the two chromosomes is conserved , which at present is more speculative . It could be that the replication pattern is the result of two opposing selection pressures . One driving force in Vibrio evolution could be the simultaneous replication of the two chromosomes . For cell cycle regulation and to limit the overall replication time to a minimum , it could be beneficial for the cell to not replicate one chromosome after the other . On the other hand , the secondary chromosome could be viewed as an invader which the cell needs to keep at bay as a second form of selection pressure . Indeed , Chr2 is thought to originate from a plasmid that the Vibrionaceae acquired early in evolution [20] . The cell might suppress Chr2 replication as far as it can to limit the danger of Chr2 taking over . Indeed , there is evidence that replication origins act as selfish genetic elements [42] . With overlapping replication cycles in fast growing cells , one could essentially imagine patterns in which Chr2 initiates replication before Chr1 with regards to the cell cycle . However , in different growth conditions , it is always Chr1 that initiates before Chr2 [22] . This finding might support a selective process active in Vibrio species to keep “Chr1 first” . In this context , it is also interesting that the initiator protein RctB actually has the capacity to mediate higher copy numbers of Chr2 , since many different single amino acid changes lead to copy up of Chr2 numbers [14 , 34 , 43] . However , selection obviously works against these copy-up mutations in Vibrios occurring in nature . The selection pressure resulting from this Chr2 suppression would be to keep the Chr2 copy number as low as possible . The combination with the second selection pressure for simultaneous replication of the two chromosomes would finally result in termination synchrony . It is noteworthy that the termination synchrony appears to tolerate some deviation , as seen in the variation in copy number ratios of ter1 to ter2 ( Fig 7 ) . This observation suggests the evolutionary process leading to termination synchrony to be slow compared to chromosome rearrangements leading to changed distances between relevant genetic loci ( ori1 , crtS ) within the system . One form of such chromosome rearrangement will be discussed in the next section . Besides the observed deviation in termination of the two chromosomes in Vibrionaceae , the delay between crtS replication and ori2 initiation also appears to be variable ( Fig 7 ) . We did not find any correlation between this delay and other relevant parameters , such as crtS site position , Chr2 size , the distance of ori1 to the crtS site , or between any combinations of these ( Supporting S7 Fig ) . Factors influencing the delay duration remain to be discovered . One of the longest delays between crtS and ori2 replication was observed for Photobacterium profundum ( Fig 7J ) . Interestingly , the respective ori2 mini-chromosome was not triggered by any of the three crtS sites tested in E . coli ( Fig 4 ) . The reason remains to be discovered since we found no obvious deviation of the Photobacterium crtS-site sequence from the other sequences ( Fig 4 ) . We also observed no increase of copy numbers for the mini-chromosomes with ori2 copies of V . harveyi and V . vulnificus ( Fig 5 ) . However , here the copy numbers of mini-chromosomes were high already in strains without crtS . Notably , the used assay is not able to detect an even further increase in copy number . The copy number of the secondary chromosomes in V . harveyi and V . vulnificus are not increased , suggesting that the observed copy-up phenomenon is mini-chromosome specific . Most published studies on DNA replication in V . cholerae have used the O1 El Tor strain N16961 . This is the strain for which the first V . cholerae genomic sequence was published [8] . However , this strain was later found to not be transformable via natural competence due to a frameshift mutation in the regulator hapR [44 , 45] . In contrast , the closely related O1 El Tor strain A1552 encodes a fully functional system of natural competence [44] . We thus decided to use strain A1552 for our experiments , expecting results directly comparable to studies using strain N16961 , as we have frequently used the published N16961 genome sequence for primer design for A1552 sequences and never experienced any deviation . Our probe design for DNA microarrays was therefore also based on the N16961 genomic sequence . However , the initial plots showed non-continuous slopes from origin to terminus , indicating some sort of chromosomal rearrangement [46] . Interestingly , an inversion around ori1 was also found recently in strain N1691 , in contrast to the published genome sequence ( Supporting S1B Fig and [25] ) . This inversion at ribosomal operons B-H was also detected in strain A1552 . Additionally , a second inversion was found at rRNA operons A-C by diagnostic PCR and whole genome sequencing , including long reads to cover the large rRNA encoding operons . These operons were indeed the point at which the inversions occurred . This is certainly due to the high similarity of rRNA operons to each other presenting a good target for homologous recombination as mechanism of inversion . Such inversions at homologous sequences such as rRNA operons or transposons have been observed when comparing genome arrangements of related strains or during long term evolution experiments [47 , 48] . A reconstruction of Yersinia evolution by comparative genomics suggests that as much as 79 inversions have happened to shape the genome arrangement seen today [49] . Inversions such as those found here in V . cholerae might thus happen quite frequently . This is certainly interesting in the context of chromosomal macrodomains that often rely on biased distributions of DNA motifs in the origin-to-terminus orientation [50–52] . However , it is also relevant for the crtS-based regulation of Chr2 initiation , in which chromosomal distances between crtS and origin or terminus certainly matter [25] , as also seen in our study . In fact , the distance between the crtS site and ori1 is changed by 142 kbp in the strain A1552 analyzed here compared to the N16961 strain analyzed before . This difference correlates well with an earlier termination of Chr2 relative to Chr1 in strain A1552 in comparison to N16961 ( Fig 7 , [25] ) . We cannot exclude that crtS-site positions in the different Vibrio species we analyzed are also re-localized slightly as , in some cases , we analyzed sub-strains not exactly resembling the strain with an available genome sequence . However , the analyses of many Vibrios instead of only individual strains should even out such uncertainties . Bacterial populations in batch cultures are mixtures of cells in all different cell cycle stages , ranging from newly-born cells to large cells shortly before division . Such cultures can be used in many ways to study mechanisms of DNA replication and related processes . However , often a synchronized cell culture is desirable in studies investigating temporal resolution in DNA replication . In bacteria , different methodologies have been developed to generate synchronous populations . These include differential density centrifugation for Caulobacter and related bacteria [53] , the baby machine [54] , the baby cell column [55] , and the blocking , and latter release , of DNA replication by the temperature shift of strains carrying temperature-sensitive protein mutants involved in initiation of DNA replication [56 , 57] . The later approach was first developed in E . coli based on screens for mutant strains with temperature sensitive DNA replication mechanisms . Mutations appeared either in the initiator protein DnaA , or the DnaC protein responsible for loading the helicase DnaB . It was shown that amino-acid exchanges which rendered the E . coli DnaA temperature sensitive could help to rationally design a temperature sensitive DnaA in other bacteria [58] . Since V . cholerae and E . coli are relatively closely related and their DnaA proteins are highly similar , we mutated the V . cholerae dnaA according to the temperature sensitive E . coli DnaA . Notably , DnaC could not be used because V . cholerae lacks a homolog . The constructed V . cholerae DnaA showed temperature-sensitive activity in the heterologous E . coli system , but we were not able to exchange the natural V . cholerae dnaA with the mutated allele . As alternative way of synchronization , we tested the stringent-response-based method established by Ferullo and colleagues . [29] . Here , initiation of replication is blocked by addition of serine hydroxamate ( SHX ) , which induces the stringent response . Transfer to SHX-free medium leads to synchronous initiation of DNA replication in E . coli . Ferullo and his collaborators suggested that this method should be transferable to other bacteria with a stringent response system such as that in E . coli . We demonstrated here that this assumption is true and predict that more , but not all , bacterial species could be synchronized using the stringent response . A negative example would be Bacillus subtilis , where the elongation of DNA replication , not the initiation , is blocked by inhibiting primase , an essential component of the replication machinery , during the stringent response [59] . Although we clearly showed the synchrony of the culture , including the linear increase of cellular DNA content as well as temporally-separated initiation of Chr1 and Chr2 replication , the synchronization could be further improved . In a perfectly synchronized culture , all cells would initiate DNA replication at the same point in time , which is hardly feasible with currently available synchronization methods . For example , the initiation of chromosomal replication in the widely applied DnaC-ts system in E . coli differs in the range of some minutes between cells [60 , 61] . Furthermore , in the system established here , cells initiate replication after the first cells have started replication , as can be seen by the increasing copy number of ori1 in comparison to ter1 over time after release from the stringent response ( Fig 2 ) . One possibility to limit replication initiation to a narrower window could be to add SHX for a second time which should limit initiation to the time between release and re-addition of SHX . The chromosomal replication origin in E . coli was found to be asymmetric , this being the root of an offset between the two replisomes [32] . The offset varied from strain to strain between 40 to 130 kbp . We observed a similar offset for the two replication forks on Chr1 in V . cholerae in synchronized cells ( Fig 2E ) . Regarding the sequence of the replication origins , the replication fork that runs ahead is the same in V . cholerae and E . coli . It was suggested that the asymmetry of replication is caused by the asymmetry of the replication origin itself , where the initiator protein DnaA multimerizes on the right side to melt an AT-rich region on the left side [32] . Intuitively , one could imagine the replication to start more easily in the direction where no initiator complex sits in the way . In the context of Chr2 regulation , the observed offset is interesting because the exact time point of crtS site replication , and with it the time of Chr2 initiation , depends on ori1 orientation . The frequent chromosomal inversions around ori1 discussed above might consequently lead to frequent changes in Chr2 replication timing and might also explain the observed deviation from exact termination synchrony in other Vibrios ( Fig 7 ) .
All strains , plasmids and oligonucleotides used in this study are listed in supporting S5–S7 Tables . Unless indicated otherwise , cells were grown in LB medium , Marine broth , or AB medium supplemented with 25 μg/ml uridine , 10 μg/ml thiamine and 0 . 2% glucose with 0 . 5% casamino acids ( AB Glu CAA ) or 0 . 4% sodium-acetate ( AB So-Ac ) [62] . Antibiotic selection for E . coli was used at the following concentrations if not indicated otherwise: Ampicillin 100 μg/ml , kanamycin 35 μg/ml , spectinomycin 100 μg/ml . For growth curves , cells were grown in a 96-well plate at 37°C in a microplate reader ( Victor X3 Multilabel Plate Reader , PerkinElmer ) . OD450 was measured every 6 min for 18 h . ori2-based mini-chromosomes synVihaII and synVitaII were constructed as described in [34]: ori2s with parAB and rctB were amplified from gDNA of the respective strain . For synVihaII , the ori2 region was amplified with primers 1515/1517 and 1410/1516 from gDNA of V . harveyi . For synVitaII , the ori2 region was amplified with primers 1164/1557 from gDNA of V . tasmaniensis . Both ori2 regions were assembled with AscI-digested synVicII-1 . 351 per Gibson assembly [63] and transformed in E . coli XL1Blue cells . For integration of crtS in E . coli MG1655 , integration cassettes were constructed by MoClo assembly [64] . For pMA161 , the crtS was amplified with primers 1474/1475 from gDNA of V . nigripulchritudo and for pMA892 with primers 1443/1444 from gDNA of V . parahaemolyticus . All PCR products were assembled in pMA349 by MoClo assembly as described in [65] and transformed into E . coli TOP10 cells . For pMA451 , the backbone pMA327 was assembled with pICH50900 , pMA709 , pMA710 , pMA431 and pMA161 . For pMA454 , the backbone pMA327 was assembled with pICH50900 , pMA709 , pMA710 , pMA431 and pMA892 . All assemblies were transformed in E . coli DH5α λpir cells . Integration cassettes were cut out with BsaI , integrated in E . coli AB330 , transferred in E . coli MG1655 per P1-transduction and recombined to remove the resistance as described in [65] . oriII-based mini-chromosomes were added to wild type and crtS strains by conjugation or transformation . V . cholerae A1552 grown in AB Glu CAA was treated with 0 . 9 mg/ml serine hydroxamate ( SHX ) at an OD450 of around 0 . 15 ( exponential phase ) . After an incubation of 150 min , the cells were harvested by centrifugation and re-suspended in fresh medium without SHX . Samples for flow cytometry and CGH were taken every 3 . 5 min if not indicated otherwise . Unless described otherwise , the cells were harvested and washed twice in TBS ( 0 . 1 M Tris-HCl pH 7 . 5 , 0 . 75 M NaCl ) . They were fixed in 100 μl TBS and 1 ml 77% ethanol and stored at least overnight at 4°C . The samples were washed in 0 . 5 M sodium-citrate and treated with 5 ng/ml RNase A in 0 . 5 M sodium-citrate for 4 hr at 50°C . They were stained with 250 nM SYTOX Green Nucleic Acid Stain ( Thermo Fisher Scientific ) and analyzed on Fortessa Flow Cytometer ( BD Biosciences ) . The SYTOX Green fluorescence was measured through a 530/30 nm bandpass filter . V . cholerae A1552 cells grown in AB So-Ac were fixed with ethanol and stained as described above . These cells served as the standard and were measured alternatingly with the samples . Data was processed with the software FlowJo ( Treestar , Ashland , USA ) . For display as density maps , the sample data was aligned to the corresponding standard and converted into density maps by R . CGH was performed as described [37] . For hybridizing , Agilent SurePrint G3 Custom CGH Microarrays , 8x60K ( Design ID: 074887 ) were used . They were designed on V . cholerae N16961 ( NC_002505 . 1 and NC_002506 . 1 ) with a probe length of 60 bp and a probe distance of 7 bp . Probes with multiple hybridization sites were excluded . As a reference , DNA from stationary V . cholerae A1552 grown in AB Glu CAA was used . Probe signal ratio values were merged in 1000 bp windows . A Lowess fitting was applied to the microarray data to get a locally weighted average ( shown as green line in CGH plots ) . For Chr1 , a stepwise function was fitted on the data . The stepwise function divides the plot into five parts: two flat parts at the edges ( not yet replicated ) , one flat risen part in the middle ( already replicated ) and an increasing and decreasing part ( replicating at that moment ) . The four points at the transition from one of these lines to the next were defined by chromosomal position ( x1 to x4 ) and the heights ( h1 to h4 ) . These values were estimated based on plots of the raw data and then used for fitting in conjunction with the stepwise function using nls ( nonlinear least-squares ) of the R statistics software . Mean positions of the replication forks were calculated as the middle of the increasing ( left fork ) or decreasing ( right fork ) part . Progression and asymmetry of the forks was calculated in Excel . For Chr2 , copy number values of 10 . 000 bp windows were compared to the mean copy number of the corresponding plot and displayed as above or below the mean . crtS sequences of Vibrionaceae in Fig 4A were found with BLAST [66] . The alignment was done with ClustalOmega [67] and the layout with MView [68] . Additionally , crtS sites in 114 Vibrionaceae from NCBI were found with fuzznuc ( http://www . hpa-bioinfotools . org . uk/pise/fuzznuc . html ) by using the sequence CAGnATATGTAACTnATGCTTTCGG with a maximum of three mismatches . This search resulted in only one hit per genome . The consensus was visualized with WebLogo 2 . 8 . 2 [69] . For comparative genomics of Vibrionaceae , data of 29 fully sequenced and annotated strains from NCBI were used . Positions of ori1 and ori2 were either found at dOriC [70] or by assigning the intergenic region between gidA/mioC ( ori1 ) or rctB/parAB ( ori2 ) . The genes were either found by annotation or with BLAST [66] . One half of each chromosome was defined as a replichore , and ter1 was then calculated as the opposite position to ori1 on Chr1 ( ori1 + 1/2 Chr1 ) . The position of crtS was found using the consensus sequence with fuzznuc . Expected values of both parameters were calculated by using the same data on both X and Y axis ( two-thirds Chr1 replichore and Chr2 replichore , respectively ) . Analysis of mini-chromosome copy numbers was as described [34] . Cells were grown in LB medium containing either 100 or 500 μg/ml ampicillin at 37°C in 96-well plates in a microplate reader ( infinite M200pro multimode microplate reader , Tecan ) . The 150 μl of main culture was inoculated 1:1000 and growth curves recorded for 15 hr . Statistical significance of differences between wt and crtS strains was calculated by a two sample t-test . For better visualization , 1 divided by the time needed to reach an OD of 0 . 1 was defined as a measure of the copy number . For Illumina sequencing of the Vibrionaceae , cells were grown in marine broth at either 28°C or 10°C ( Photobacterium profundum ) to either exponentially or stationary phase . Genomic DNA was prepared by incubating resuspended frozen cell pellets in 300 μl TE-buffer with 1 . 2% SDS and 4 mM EDTA for 5 min at 65°C . After adding 750 μl isopropanol , the precipitate was incubated in 500 μl TE with 50 μg RNase A for 90 min at 65°C and additional 15 min at 37°C with 50 μg proteinase K . DNA was isolated with phenol/chloroform . Final DNA was resuspended in deionized sterile water and quantified using a NanoDrop ( ThermoFisher Scientific ) . Genomic DNA was sequenced by applying the Nextera XT library kit and a MiSeq v3 reagent kit with 150 cycles on an Illumina MiSeq ( Illumina , USA ) . For PacBio sequencing , V . cholerae A1552 [31] was grown in AB Glu CAA at 37°C to stationary phase . Genomic DNA was prepared as above . Final DNA was resuspended in TE-buffer and quantified using a NanoDrop ( ThermoFisher Scientific ) and a Qubit Fluorometer ( Life Technologies ) . The SMRTbell template library was prepared according to the instructions from PacificBiosciences , Menlo Park , CA , USA , following the Procedure & Checklist—20 kb Template Preparation Using BluePippin Size-Selection System . Briefly , for preparation of 15kb libraries ~8μg genomic DNA libraries was sheared using g-tubes from Covaris , Woburn , MA , USA according to the manufacturer´s instructions . DNA was end-repaired and ligated overnight to hairpin adapters applying components from the DNA/Polymerase Binding Kit P6 from Pacific Biosciences , Menlo Park , CA , USA . Reactions were carried out according to the manufacturer´s instructions . BluePippin Size-Selection was performed according to the manufacturer´s instructions with a size selection cutoff of 4 kb ( Sage Science , Beverly , MA , USA ) . Conditions for annealing of sequencing primers and binding of polymerase to purified SMRTbell template were assessed with the Calculator in RS Remote , PacificBiosciences , Menlo Park , CA , USA . SMRT sequencing was carried out on the PacBio RSII ( PacificBiosciences , Menlo Park , CA , USA ) taking one 240-minutes movie . Genome assembly was performed with the RS_HGAP_Assembly . 3 protocol included in SMRT Portal version 2 . 3 . 0 . Both chromosomal contigs were successfully assembled and trimmed , circularized , as well as adjusted to dnaA ( Chr1 ) and rctB ( Chr2 ) as the first gene . Quality improvement of the PacBio HGAP assembly was performed by a mapping of all corresponding Illumina short reads using the Burrows-Wheeler Aligner ( BWA ) using bwa aln and bwa sampe [71] . Illumina reads were mapped onto the obtained chromosome and plasmid sequences with subsequent variant and consensus calling using Varscan2 [72] and GATK [73] . A final quality score of QV60 was attained . Automated genome annotation was carried out using Prokka [74] . The genome sequence was submitted to GenBank ( Accession Number: CP024867; CP024868 ) . In the first step , reads from the exponential and stationary phase were mapped on the respective Vibrio replicons using qalign from the QuasR R package . Subsequently , replicon-wide coverage was calculated by bedtools genomecov using the 5' ends of the reads . Single base coverage was smoothed by a 5 kbp sliding window averaging with a shift of 1 kbp . Windows with an internal standard deviation that exceeded three times the difference between the median and the third quartile of standard deviations of windows within 500 kbp were removed . These are windows , with an average coverage that does not properly reflect the coverage of individual bases . Furthermore , windows with an internal standard deviation below three times the difference between the median and the 1st quartile of standard deviations of windows within 500 kbp were removed . These are windows with many bases of low or mostly zero coverage indicating deviations of reads from the template sequence . The procedure removes unreliable window averages ( data points ) taking the noisiness of the data and regional specificities into account . Sequence bias was removed as follows: firstly , the coverage of exponential and stationary phase samples was normalized to the total amount of mapped reads to remove the bias of total read counts in the samples . Then , ratios of exponential and stationary phase coverage were determined . Ratios were subsequently corrected for a systematic sequence-dependent local bias [25] , using the second exponential phase sample .
|
Most bacteria encode their genetic information on a single chromosome . The pathogenic bacterium Vibrio cholerae is an exception to this rule and carries two chromosomes of different sizes , each having one origin of replication . A very basic research question is how the replication of the two chromosomes is timed starting from their replication origins . If they start simultaneously , the smaller chromosome would finish replication earlier than the larger chromosome . Interestingly , the timing in V . cholerae is such that the smaller chromosome starts replication after a time delay , resulting in synchronous replication termination of the two chromosomes . Here we answer the question whether it is the termination synchrony which is under evolutionary pressure , or whether a certain duration of the delay between the two chromosomes to start replication is of biological importance . To this end , we analyzed replication in different species of the Vibrionaceae phylogenetic group with differently sized chromosome pairs . Our results indicate that a synchronous termination of the two chromosomes in this group of bacterial species is under evolutionary selection , suggesting it to be potentially important for the process of cell division .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"bacteriology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"vibrio",
"cell",
"processes",
"cell",
"cycle",
"and",
"cell",
"division",
"microbiology",
"operons",
"vibrio",
"cholerae",
"rrna",
"operons",
"dna",
"replication",
"bacterial",
"genetics",
"dna",
"microbial",
"genetics",
"bacteria",
"bacterial",
"pathogens",
"microbial",
"genomics",
"research",
"and",
"analysis",
"methods",
"bacterial",
"genomics",
"sequence",
"analysis",
"genomics",
"sequence",
"alignment",
"bioinformatics",
"medical",
"microbiology",
"microbial",
"pathogens",
"comparative",
"genomics",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"organisms"
] |
2018
|
Synchronous termination of replication of the two chromosomes is an evolutionary selected feature in Vibrionaceae
|
Bacterial extracellular polysaccharides are a key constituent of the extracellular matrix material of biofilms . Pseudomonas aeruginosa is a model organism for biofilm studies and produces three extracellular polysaccharides that have been implicated in biofilm development , alginate , Psl and Pel . Significant work has been conducted on the roles of alginate and Psl in biofilm development , however we know little regarding Pel . In this study , we demonstrate that Pel can serve two functions in biofilms . Using a novel assay involving optical tweezers , we demonstrate that Pel is crucial for maintaining cell-to-cell interactions in a PA14 biofilm , serving as a primary structural scaffold for the community . Deletion of pelB resulted in a severe biofilm deficiency . Interestingly , this effect is strain-specific . Loss of Pel production in the laboratory strain PAO1 resulted in no difference in attachment or biofilm development; instead Psl proved to be the primary structural polysaccharide for biofilm maturity . Furthermore , we demonstrate that Pel plays a second role by enhancing resistance to aminoglycoside antibiotics . This protection occurs only in biofilm populations . We show that expression of the pel gene cluster and PelF protein levels are enhanced during biofilm growth compared to liquid cultures . Thus , we propose that Pel is capable of playing both a structural and a protective role in P . aeruginosa biofilms .
Biofilms are surface associated communities embedded within an extracellular matrix [1] , [2] , [3] . Biofilm communities exhibit enhanced antibiotic tolerance [4] , [5] , [6] . As a result , biofilm infections tend to be chronic and difficult to eradicate [2] , [7] . This enhanced tolerance is thought to be multi-factorial , owing to biofilm-associated patterns of gene expression , slow growth rate , and reduced antimicrobial diffusion within the biofilm [4] . A focus of research has been to identify biofilm-associated factors that contribute to their antibiotic tolerance . The opportunistic pathogen , Pseudomonas aeruginosa , is a model organism in biofilm research . P . aeruginosa is well known for the chronic infections it causes in individuals with the genetic disease , cystic fibrosis ( CF ) [7] . Biofilm formation within the CF airways is believed to facilitate the infection , helping the bacteria to withstand aggressive antimicrobial treatment and host defenses [8] , [9] . The extracellular matrix is a distinguishing feature of biofilms , capable of functioning as both a structural scaffold and protective barrier to antimicrobials [1] , [2] , [10] , [11] , [12] . A key component of the matrix is extracellular polysaccharides [13] . Exopolysaccharides carry out a wide range of functions involving surface and cell-cell interactions , as well as protecting against antimicrobials and host defenses [3] , [10] , [14] , [15] , [16] . P . aeruginosa produces three exopolysaccharides , alginate , Pel and Psl , all of which have been implicated in biofilm development under different circumstances [11] . Pel's composition has yet to be fully elucidated . Initial carbohydrate analysis suggests Pel is a glucose-rich polysaccharide polymer although the exact structure remains unknown [17] . Pel synthesis machinery is encoded by a seven gene operon ( pelA-F ) originally identified in a mutagenesis screen for the loss of pellicle formation in PA14 [17] . Pel also appears to be important in static microtiter dish biofilm assays . A pel mutant strain had a defect in biofilm biomass accumulation in comparison to wild-type PA14 [17] , [18] . The mechanism behind this observation remains unclear . Other studies have demonstrated that in the absence of type IV pili , Pel can play a role in attachment suggesting it can compensate as an attachment factor in the absence of other adhesins [18] . In this study , we conducted an analysis of Pel function . We focused our study on two common laboratory strains , PAO1 and PA14 . PAO1 is capable of making both the Pel and Psl exopolysaccharides , while PA14 is only capable of producing Pel since three genes of the psl operon are deleted in this strain . We show that Pel is critical for maintaining cell-cell interactions in developing PA14 biofilms as well as providing protection against aminoglycoside antibiotics during biofilm growth . We also show that Pel does not appear to play any critical role in PAO1 biofilm development , where Psl appears to be the primary biofilm polysaccharide . Finally , we demonstrate that the pel operon is transcriptionally induced and PelF protein levels increase during biofilm growth . Thus , we propose that Pel can serve both as a structural and protective factor within a biofilm community .
To initiate our study , we constructed a Pel overexpression strain . The native promoter region of pelA was replaced with the araC-PBAD promoter on the chromosome in two common laboratory strains , PAO1 and PA14 , allowing arabinose-dependent expression of the pel operon ( Figure S1A ) . PA14 is a clinical strain obtained from a burn patient that has a multi-gene truncation in the N-terminal region of the psl operon and is incapable of synthesizing the Psl polysaccharide [19] . Accordingly , PA14 serves as a useful strain to study the contribution of the Pel polysaccharide independently of Psl . In contrast , PAO1 has the necessary genes to produce both polysaccharides . The inducible strains will be referred to as PAO1PBADpel and PA14PBADpel . Quantitative RT-PCR was used to quantify pelA transcript level from log phase cells . The level of pelA transcript increased with increasing concentrations of the inducer , arabinose ( Figure S1C ) . The dose-dependent increase in transcription level was similar between PA14 and PAO1 . Addition of 0 . 2% arabinose led to 51- and 61-fold increase in expression levels in PA14PBADpel and PAO1PBADpel , respectively . pelA transcript is expressed 1 . 8 times higher in wild-type PA14 compared to wild-type PAO1 relative to the internal control transcript , ampR ( Figure S1C ) . We evaluated PAO1PBADpel and PA14PBADpel for the ability to conditionally produce more Pel polysaccharide with increasing pel transcription . Previous work has demonstrated that Pel synthesis is controlled at multiple levels , transcriptionally and allosterically [20] , [21] . Congo red binding and liquid culture aggregation are two phenotypes associated with increased polysaccharide production in multiple bacterial species [20] , [22] , [23] . Addition of 1% arabinose to both PA14PBADpel and PAO1PBADpel leads to bacterial aggregation in liquid culture relative to the uninduced strain and these bacterial aggregates hyperbind Congo red ( Figure S1B ) . Pel expression was previously demonstrated to impact pellicle formation and colony morphology [17] . Wild-type PA14 forms a distinct pellicle after about two days of incubation at room temperature which becomes more pronounced over time . When induced , PA14PBADpel rapidly produces a thicker pellicle compared to wild-type . A top-down view reveals that the PA14PBADpel stain produces a pellicle with a highly defined wrinkly architecture that is resistant to extensive vortexing ( Figure S1D , bottom panel ) . Consistent with previously published data , a mutation in pelB leads to a dramatic reduction in pellicle formation compared to the parental strain [17] , [24] . Pellicles produced in wild-type PAO1 grown under the same conditions are less distinct than PA14 pellicles . No discernable difference in pellicle formation is seen between PAO1 and PAO1ΔpelB , but overexpressing Pel enhances pellicle formation similar to PA14 ( Figure S1D ) . Overexpression of pel produces enhanced wrinkly colony morphology in PA14 , whereas the pelB mutant grows as a smooth colony with little Congo red binding ( Figure S1D , top panel ) . Unexpectedly , PAO1 did not produce a wrinkly colony morphology for any of the tested strains despite many attempts with varying temperature and media conditions ( Figure S1D , top panel ) . We investigated the function Pel plays in biofilm development using two biofilm culturing methods , a microtiter dish assay and a flow-cell reactor . A microtiter dish assay quantifies biofilm formation on plastic during static incubation . In contrast , a flow cell bioreactor allows a microscopic analysis of live biofilms growing in dilute medium under conditions of continuous flow . The influence of Pel on initial attachment to a plastic surface was examined . A pelB mutation in PA14 did not impact bacterial attachment ( Figure 1A ) . However , overexpressing pel in either PA14 or PAO1 increased surface attachment ( Figure 1A and 1B ) . Similar to PA14 , no phenotype were seen in a PAO1ΔpelB mutant compared to wild-type PAO1 ( Figure 1B ) . In contrast , but consistent with previously published work , a polar deletion in psl had a strong attachment defect in PAO1 , indicating that Psl , and not Pel , is an important adhesin for surface attachment under these conditions [25] , [26] . Crystal violet staining is an indirect measurement of bacterial attachment and thus we took a complementary , microscopic approach to evaluate Pel's role in attachment to a glass surface in a flow-cell reactor . Images were acquired by scanning confocal laser microscopy ( SCLM ) and analyzed by COMSTAT 1 software for surface coverage [27] . No statistical differences between PA14 , PA14ΔpelB , uninduced and induced PA14PBADpel for attachment are observed ( Figure 1C and S2A ) . Similar to PA14 , no difference is observed for any of the PAO1 strains tested under non-inducing and inducing conditions ( Figure 1D and S2A ) . These results are slightly inconsistent with our microtiter dish assay , which demonstrate a modest but clear increase in attachment for the Pel overexpression strains . However , in both PAO1 and PA14 , a pel mutation did not affect attachment in either biofilm culturing method . Unlike surface attachment , we found that Pel has a significant impact on later stages of biofilm development and this impact was found to be strain dependent . To assess effects of Pel on later stages of biofilm development , we grew strains for 24 h in a microtiter dish assay and found that the pelB mutant strain of PA14 has a significant reduction in biofilm biomass compared to the parental PA14 strain , similar to previous findings ( Figure 2A ) [17] , [18] . The PA14ΔpelB biofilm defect was complemented by supplying Ppel in trans . Ppel contains the entire pel operon cloned into an arabinose-controlled expression plasmid , pMJT-1 ( Figure 2A ) . Overexpressing Pel in PA14 increased biofilm biomass almost two-fold ( Figure 2A ) . In contrast to PA14 , no difference is seen between PAO1 and PAO1ΔpelB , while overexpressing Pel results in a modest increase of biofilm biomass ( Figure 2B ) . Conversely , PAO1ΔpslBCD has a pronounced defect , suggesting that Psl is the dominant polysaccharide in PAO1 for both attachment and biofilm maintenance , as reported by Ma , et al . [26] . Biofilm formation by these strains was monitored in a flow-cell bioreactor to allow for live imaging and structural analysis . Under these conditions , P . aeruginosa forms biofilms that contain mushroom-shaped multicellular structures . PA14 forms small microcolonies by day two that further develop into a structurally complex biofilm with large multicellular aggregates of bacteria by day four ( Figure 2C ) . In stark contrast , PA14ΔpelB fails to form cellular aggregates . After four days of growth , PA14ΔpelB remains as a dense monolayer of cells attached to the glass surface , incapable of developing the complex three-dimensional structures typical of the wild-type strain ( Figure 2C ) . The absence of cell aggregates in PA14ΔpelB indicates Pel may be responsible for the cell-to-cell adhesion necessary for aggregate formation . In support of this , overexpressing Pel results in larger cellular aggregates and enhanced biofilm biomass compared to wild-type PA14 . Flow cell images were quantified for four properties of biofilm development using COMSTAT 1 , average thickness , roughness coefficient , surface-to-volume ratio and maximum thickness ( Figure S2B ) [27] . Pel overexpression in PA14 affected each property by increasing the average thickness , decreasing the roughness coefficient , decreasing the surface to volume ratio and increasing the maximum thickness . In contrast to PA14 , no major visual or quantifiable difference is seen in biofilm structure after four days of growth for PAO1 , PAO1ΔpelB and PAO1PBADpel ( Figures 2D and S2B ) . However , a modest , but not statistically significant , increase in average biofilm thickness is detected for PAO1PBADpel . We subsequently assessed whether a Pel-dependent phenotype might manifest itself in older biofilms . Yet , even after nine days no significant differences were observed for PAO1 , PAO1ΔpelB and PAO1PBADpel ( Figure S3 ) . Continuous production of the Psl polysaccharide was recently shown to be required for both the addition of new biofilm biomass to a growing biofilm and for the maintenance of existing biofilm structure [26] . Conditional loss of Psl expression resulted in a halt of biofilm growth and an eventual erosion of the existing biofilm structure [26] . Using our conditional expression system we grew PA14PBADpel biofilms for two days in the presence of arabinose and either continued providing arabinose to the biofilm culture for an additional two days or we removed it from the growth medium ( Figure 3 ) . Interestingly , halting Pel expression by removing arabinose resulted in a biofilm that failed to increase in size , but retained the original shape and mass after two days as calculated by COMSTAT ( Figure S4 ) . The biofilm that was supplied arabinose continued to grow in size . These results suggest that continuous Pel production is important over the course of biofilm development . However , unlike Psl , continuous Pel production is not required to maintain existing biofilm structure . We hypothesized that the absence of cell aggregates in the PA14ΔpelB mutant biofilms is due to a defect in the cell-to-cell interactions necessary to hold an aggregate together . To initially test this hypothesis , we used time-lapse microscopy to analyze the behavior of biofilm cells at an early point in biofilm development . Dividing cells on the glass surface were monitored and the fate of daughter cells were separated into two categories [28] . Daughter cells that remained closely associated with the mother cell were termed “aggregate builders” . Cells that did not remain closely associated with the mother cells were designated “flyers” . We predicted if Pel were important in cell-to-cell interactions , cells incapable of Pel production would show a larger percentage of daughter cells exhibiting “flyer” behavior . Our analysis determined that Pel is a crucial determinant in daughter cell behavior in PA14 ( Figure 4 ) . As predicted , expression of Pel is related to daughter cell association with the parental cell . A pelB mutant displayed increased “flyer” behavior ( 88 . 3% ) in comparison to wild-type PA14 ( 40 . 2% ) and a reduced proportion of “aggregate builders” were observed ( 11 . 7% ) than in PA14 ( 59 . 8% ) . Overexpressing pel resulted in an increased proportion of aggregate builders ( 82% ) and relatively few flyers ( 18% ) . Conversely , in PAO1 Psl appears to be the primary polysaccharide involved in aggregate building . PAO1 and the PAO1Δpel mutant display indistinguishable daughter cell behavior profiles , with 33 . 9% and 32 . 6% of “flyers” , respectively . Conversely , PAO1ΔpslBCD exhibits a much larger proportion of “flyers” ( 85 . 3% ) compared to PAO1 and PAO1ΔpelB ( Figure 4 ) . These data support our hypothesis that Pel contributes to aggregate formation in a PA14 biofilm by promoting retention of daughter cells within a growing aggregate , while Psl appears to be the critical polysaccharide for aggregate building in PAO1 . To complement the time-lapse microscopy study , we developed a novel assay involving infra-red laser [29] . This assay involves maintaining an optical trap in a liquid suspension of bacteria . Once bacteria enter the trap , they remain there . Initial experiments determined that continuous trapping of PA14 cells in liquid culture promoted the formation of stable aggregates . Using this technique we are able to study the effects of the Pel polysaccharide for the ability to form and maintain bacterial clusters . Since Pel was required for maintaining cell-to-cell interactions in flow cell biofilms , we predicted that Pel would be required for maintaining stable aggregates in this assay . Wild-type PA14 forms aggregates after 20 min of trapping in all visualized fields ( Figure 5A ) . In contrast , PA14ΔpelB did not form aggregates , even though a significant amount of free-floating bacteria entered and remained in the trap ( Figure 5B ) . Rather , the mutant strain requires a minimum of 45 min of trapping to form aggregates ( Figure 5C ) . Even with the extended incubation in the trap , 16% of the fields of view are absent of aggregates . Based on the differences in time for bacterial clustering to be observed , these data conclude Pel is an important component in the initiation of cellular clustering . Subsequently , we tested to see if Pel was important in maintaining clustering aggregates after an aggregate was formed . These experiments were set up similarly by allowing bacterial aggregation to occur for 20 min for wild-type PA14 and 45 min for the pelB mutant . After the designated incubation time with the laser , cluster stability was monitored by microscopy five minutes after the laser trap was disengaged . More than six aggregates in each strain were visually assessed for stability and separated into three categories as described in the figure legend ( Figure 5 ) . 85% of wild-type PA14 cell aggregates are stable five min after the release of the trap . In contrast , only 16% of the PA14ΔpelB aggregates remained after the laser trap is removed . These results further support that Pel is critical for both initiating and maintaining cell-to-cell interactions . A primary function attributed to the extracellular matrix is protection [12] . Several well-studied polysaccharides are known to confer resistance to a range of antibiotics . In P . aeruginosa , alginate and cyclic glucans have been demonstrated to protect biofilms from aminoglycosides by directly binding these cationic antibiotics [15] , [30] , [31] . In addition , rugose small colony variants ( RSCVs ) , which produce elevated levels of Pel and Psl , show increased tolerance to tobramycin , an aminoglycoside [23] , [32] . Thus , we hypothesized that Pel may provide protection from antimicrobials . Therefore we tested the sensitivity of our strains to several clinically relevant antibiotics: tobramycin , gentamicin , ciprofloxacin , kanamycin , meropenem , ceftazidime , tetracycline , and carbenicillin . Planktonic cultures of PAO1 , PAO1ΔpelB , PAO1Δpsl , PAO1PBADpel and WFPA801 ( arabinose-inducible psl strain [26] ) were initially tested for antibiotic susceptibility by determining the minimum inhibitory concentration ( MIC ) of each strain . No difference is detected between PAO1 , PAO1ΔpelB , PAO1Δpsl and WFPA801 for any of the antibiotics tested ( Figure S5 ) . However , overexpressing Pel in PAO1 slightly increases the MIC in comparison to wild-type PAO1 to gentamicin and tobramycin , two aminoglycoside antibiotics . No difference is seen in MICs between PA14 , PA14ΔpelB and PA14PBADpel for any of the antibiotics tested ( Figure S5 ) . We then assessed Pel's involvement in planktonic survival by treating log-phase cultures of our strains with both tobramycin and ciprofloxacin . Ciprofloxacin was chosen as a representative antibiotic that has the same MIC for all three strains PAO1 , PAO1ΔpelB and PAO1PBADpel . Equal susceptibility is seen between the wild-type and ΔpelB mutants in both PAO1 and PA14 ( Figure S6 ) . Like the MIC experiments , overexpression of Pel in PAO1 affords a small degree of protection to killing by tobramycin and gentamycin , while Pel overexpression in PA14 does not ( Figure S6 ) . Similar killing curves are observed between PA14 and PAO1 strains during ciprofloxacin treatment ( Figure S6 ) . We subsequently assessed Pel's role in antibiotic resistance in a biofilm model . For a valid comparison of antibiotic tolerance , the same number of cells must be challenged with the antibiotic of interest . In order to satisfy this criterion , we used a 48-h colony biofilm technique that has been demonstrated previously to capture a biofilm-specific model of antibiotic susceptibility [33] . Bacterial strains were grown on polycarbonate filters for two days , allowing complete coverage of the filter and equal colony forming units ( CFUs ) for all strains . The filter was then transferred to solid medium containing antibiotic and incubated for 24 hours . After treatment , the viable CFUs were quantified . In PA14 , a pel mutation rendered biofilms more susceptible to the aminoglycosides tobramycin and gentamicin , while not impacting the susceptibility to ciprofloxacin ( Figures 6 and S7 ) . However , a pel mutation in PAO1 did not influence susceptibility to any antimicrobial tested . Overexpressing pel in both PAO1 and PA14 led to an elevated tolerance to tobramycin and gentamicin compared to the corresponding parental strain ( Figures 6 and S7 ) . To test whether Psl overexpression might provide similar aminoglycoside protection to PAO1 , we used an arabinose-inducible Psl expression strain , WFPA801 . Psl overexpression strain was found not to confer protection from tobramycin ( Figure 6 ) . To make a more direct comparison between planktonic and biofilm cultures , we compared 24-h-old stationary phase cells with 24-h-old biofilm cells for tobramycin sensitivity . The tobramycin sensitivity profiles of stationary phase liquid cultures are identical for PAO1 and PA14 wild-type and the corresponding pelB-mutant strains ( Figure 7 ) . Interestingly , overexpression of Pel in PA14 provides protection in stationary phase cells that is not observed in log phase cells ( compare Figures 7 and S6 ) . Similar to log-phase treated cells , PAO1PBADpel provides protection to stationary phase cells . 24-h-biofilms reveal the same susceptibility profiles as the 48-h-treated biofilms shown in Figure 6 , with enhanced sensitivity of PA14ΔpelB compared to wild-type ( Figure 7 ) . To complement our analysis of colony biofilms , we determined the spatial distribution of tobramycin killing in flow cell biofilms . As expected , PA14 and its derivatives display a similar tobramycin resistance pattern as the filter biofilm . The pelB mutant strain produced a monolayer that is easily killed , while the PA14PBADpel strain biofilm is the least susceptible , probably in part due to the production of greater amounts of biofilm biomass than PA14 ( Figure S8 ) . Our antimicrobial tolerance data suggest that in PA14 , Pel plays a more important role in biofilm communities as compared to planktonic cultures . One explanation for this observation is that pel expression may be enhanced during biofilm growth as compared to planktonic growth . To test this , we analyzed the expression of the pel operon in planktonic and biofilm cells using quantitative RT-PCR . To generate enough biofilm biomass for RT-PCR we grew the strains on the surface of silicon tubing under constant flow . We observed that pelA transcript in PA14 is 7 . 2-fold higher ( +/−2 . 0 ) when grown as a biofilm for 48 h than in planktonic conditions ( for either logarithmic or stationary phase cells ) , while in PAO1 it is 5 . 11-fold higher ( +/−3 . 49 ) . The control transcripts pslA , lasR and sadC did not exhibit biofilm-specific induction ( Figure 8A ) . The pslA transcript was chosen as a control because the Psl polysaccharide is an important structural component in biofilm development in PAO1 . The lasR transcript was chosen because LasR responds to an increase in biomass and the sadC transcript was chosen because the product SadC is a diguanylate cyclase important in regulating biofilm advancement [34] , [35] . Therefore , an increase in pelA transcript is specific to the pel operon and not all genes involved in biofilm formation . Even by 24 h of biofilm growth , PAO1 shows nearly a 15-fold increase in pelA transcript ( Figure S9 ) . To corroborate our transcriptional analysis , we also demonstrate PelF protein levels are elevated during biofilm growth but went undetected in stationary phase liquid cultures ( Figure 8B ) . These data suggest a biofilm-associated role for Pel .
In this study , we have identified two key biofilm-associated functions of the Pel polysaccharide . Pel is critical for initiating and maintaining cell-cell interactions . These functions have been implicated in polysaccharides in other species , such as the MDX polysaccharide of Shewanella oneidensis and colanic acid of E . coli K-12 [36] , [37] . This appears to be a crucial mechanism by which parent cells retain their daughter cells in the biofilm community . In the absence of Pel , biofilm formation does not progress beyond the monolayer stage in PA14 . In addition , Pel appears to provide a measure of protection from aminoglycoside antibiotics . The antibiotic susceptibility experiments suggest that Pel is capable of providing protection to planktonic cells when artificially overexpressed , although there is no phenotype for the pel mutant strain in liquid culture for any of the tested conditions . However , in biofilms of PA14 , both pel overexpression and a pel mutation impacted aminoglycoside sensitivities . This suggests that Pel may play an important protective role in biofilms of this strain . The mechanism responsible for protection is not clear , but if Pel behaves similarly to other polysaccharides leading to elevated aminoglycoside resistance like alginate and ndvB-encoded glucans , it may bind or sequester the antibiotic . Both alginate and ndvB-encoded glucans have a high negative charge that is consistent with their ability to bind positively charged aminoglycosides . If this model proves to be true , Pel may be an acidic polysaccharide capable of interacting with cationic antibiotics . This hypothesis helps explain why no differences were seen in killing of planktonic or biofilm cells by ciprofloxacin , an anionic antibiotic , and why no protection was afforded by overexpressing the neutral polysaccharide , Psl ( Figure 6 ) . Another possibility is that Pel production can influence biofilm structure , which in turn may influence antimicrobial susceptibility . However , we feel this is unlikely since the structure of colony biofilms tends to be uniform . Using PAO1 and PA14 as representative P . aeruginosa laboratory strains , we see that the role Pel plays in biofilm formation can vary drastically . In PAO1 , it appears that Psl is the predominate polysaccharide of the biofilm EPS matrix , while in PA14 Pel is required . However , it appears in other strains , MJK8 , PAO1ΔwspF and ZK2870 , both polysaccharides contribute to biofilm and/or autoaggregation phenotypes [20] , [24] , [32] . Whether each polysaccharide has a distinct role in biofilm formation and/or protection , or if their functions are redundant remain to be determined . Although purely in terms of surface attachment , it appears that Psl is more important , while Pel is less so . The enhanced expression of Pel in biofilms is noteworthy . pelA transcript levels were minimally expressed under the planktonic culturing conditions we used . Yet despite low pelA transcript in planktonic conditions , only in biofilms do we detect PelF protein expression ( Figure 8 ) . Therefore , the protection afforded to P . aeruginosa by Pel from aminoglycosides appears to be a biofilm-associated mechanism of antimicrobial tolerance . To date , only the cyclic glucans encoded by the ndvB locus has been shown to be a biofilm-specific mechanism of antimicrobial tolerance in this species [30] . Characterizing the structure of Pel and the specific mechanism behind aminoglycoside protection is underway . Finally , the ability to prohibit PA14 biofilms from growing larger by arresting Pel expression is exciting . The biofilm does not dissipate indicating that continuous Pel is not necessary for biofilm maintenance . This result is contrary to PAO1 biofilms that require Psl to be continuously produced for biofilm maintenance [26] . Thus , manipulating Pel and Psl expression may be a central strategy for disrupting biofilms and targeting them for antibiotic therapy .
Strains and primers used in this study are listed in Table S1 [38] , [39] , [40] , [41] , [42] , [43] . Plasmid and strain construction are described in Text S1 . Unless otherwise noted , strains were grown at 37°C in LB medium . For plasmid selection , 300 µg/ml carbenicillin or 100 µg/ml gentamicin was used with P . aeruginosa , and 100 µg/ml ampicillin or 10 µg/ml gentamicin was used with Escherichia coli . RNA was extracted using an RNeasy kit ( QIAGEN ) according to manufactures instructions . Contaminating DNA was removed with an on-column RNase-free DNase I treatment ( QIAGEN ) and remaining DNA was removed by an off-column DNase I treatment ( Promega ) as recommended . The RNA prep was confirmed to be free of DNA by PCR . cDNA was generated by SuperScript III First-Strand Synthesis System for RT-PCR using random hexamers ( Invitrogen ) . cDNA synthesis was verified by PCR and quantitated by RT-PCR using the SYBR Green PCR Master Mix ( Applied Biosystems ) as the fluorescent dye . Fluorescence was measured using ABI Prism 7000 Sequence Detection and pelA transcript levels were normalized to ampR . 96-well microtiter dish experiments were performed as described previously [44] . For rapid attachment assays , 100 µl of log-phase cells were incubated at 37°C for one hour . For analysis of biofilm development , log-phase cells were incubated at room temperature for 20 h . Standing cultures containing 3-ml LB broth were grown at room temperature in a glass tube . Pellicles were monitored by visual inspection between four and ten d . Complete coverage at the air-liquid interface of an opaque layer of cells is considered to be indicative of pellicle formation [17] . LB liquid cultures supplemented with 40 µg/ml Congo red ( Sigma-Aldrich ) and incubated shaking overnight 37°C . The supernatants were measured at OD495 to assess Congo red binding . Congo red plates contained LB without NaCl , 1% agar , 40 µg/ml Congo red and 15 µg/ml brilliant blue R ( Sigma-Aldrich ) . Cells were diluted 1/100 , 10 µl spotted and incubated at room temperature for five d . The flow cell system and tube biofilm system was assembled as described previously [32] , [45] , [46] . Additional information is found in Text S1 . Biofilms were grown in flow cells in 1% TSB as described in Text S1 for 4 d and subsequently treated with 1 µg/ml of tobramycin for 24 h . The MIC for tobramycin when the cells are grown in 1% TSB is 0 . 03 µg/ml for PAO1 and 0 . 06 µg/ml for PA14 . This is in contrast to the MIC of 1 µg/ml seen for PA14 and PAO1 grown in full strength LB . After treatment , flow was stopped and biofilms were stained with 500 µl of propidium iodide and SYTO 9 ( Invitrogen ) for 10 m according to manufactures instructions . Flow resumed and images were captured after 15 m of washing . The movement and behavior of individual daughter cells of P . aeruginosa monocultures were monitored in young ( <2 days old ) flow cell biofilms . The flow cell setup was the same as described . Fluorescent images were taken every 45 seconds for 2 hours using the time-lapse feature of the Zeiss Axiophot microscope ( Carl Zeiss ) . The fate of new daughter cells were visually tracked and each cell was classified as an “aggregate builder” or “flyer” [28] . From one cell division to the next , “aggregate builders” remained within a 15 µm diameter circle centered at the point of cell division , while “flyers” moved further away on the surface or were dissociated from the surface by media flow . A minimum of 75 cell divisions of each strain were tracked and classified . An open-top chamber for microscopy was constructed by layering five Secure-Seal Imaging Spacers 13mm in diameter ( Grace Bio-Labs ) onto a microscope cover slip and this chamber placed on an inverted Olympus microscope above a 40× or 60× long-working-distance objective . Bacteria were grown shaking at 37°C to OD600 1 . Samples were incubated statically at room temperature for 1–7 hours . Static incubation of the cells prior to the experiment allowed for a slight increase ( ∼2-fold , data not shown ) in pel expression , presumably for the same reasons pel expression is required for pellicle formation in static liquid cultures [17] . A 100 µL bacterial suspension was placed into the chamber . Laser trapping was done by focusing a 1064 nm laser through the microscope objective at the top of the sample at a transmitted power of ∼50mW . For PA14 wild-type bacteria , the first trap-induced clustering was seen in samples that had been in the open-top chamber for 20 min; for ΔpelB bacteria the first trap-induced clustering , if any , was seen in samples that had been in the open-top chamber for 45 min . Cluster stability was evaluated by monitoring formed bacterial aggregates for cluster dispersal five min after the laser beam was blocked . Immunoblots were performed with whole-cell lysates as described with equal amounts of total protein in each lane [47] . Protein concentration was measured using the Pierce 660 nm protein assay ( Thermo Scientific ) . Lysates were probed for PelF expression levels with a specific PelF antibody . Additional information on antisera production and immunoblot analysis are described in Text S1 . MIC growth curves were completed in a 96-well microtiter dish grown in LB broth at 37°C . Log phase bacteria were diluted to 105 CFU in each well . A range of concentrations was assessed for each antibiotic . Bacterial growth was measured after 24 hours of incubation using a microplate reader ( OD590 ) . For planktonic killing curves , cells were either grown to log phase in LB or grown for 24 h to assess stationary phase susceptibility . Log phase cultures were split and one culture was treated with either tobramycin ( Sigma ) , gentamicin ( Sigma ) or ciprofloxacin ( Bayer Healthcare ) , while the control culture was untreated . Stationary phase cells were resuspending in fresh media containing antibiotics . The cultures were incubated shaking at 37°C . Bacterial survival was assessed over time by viable plate counts . For biofilm killing , overnight liquid cultures were diluted 1/100 and five µl were spotted onto a UV-sterilized 25mm polycarbonate filter ( GE Osmonics ) . Biofilms were grown for two d unless otherwise stated at 37°C and moved to fresh solid media each day . Biofilms were exposed to with 5 µg/ml tobramycin or 1 µg/ml ciprofloxacin for 24 h . Bacterial viability was obtained by resuspending the filter in 1 ml of PBS and serially diluting to obtain viability counts .
|
Most bacteria live within biofilm communities , which are a complex population of microorganisms that attach to surfaces and produce copious amounts of extracellular matrix material . Exopolysaccharides are a key feature of the extracellular matrix and are found in many forms , ranging from structurally simple linear homopolymers to structurally complex branched heteropolymers . Exopolysaccharides carry out a wide range of functions involving adherence to surfaces and other cells , structural support and protection against host and environmental stress . The goal of our study was to examine the functional importance of polysaccharide production in the model biofilm organism , Pseudomonas aeruginosa . Using a deletion and over expression strategy , we characterized the function of one polysaccharide , Pel , and demonstrated that this polysaccharide has two roles , a structural role and a protective role , against an important class of antibiotics , aminioglycosides .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"microbiology/microbial",
"growth",
"and",
"development",
"microbiology/medical",
"microbiology"
] |
2011
|
The Pel Polysaccharide Can Serve a Structural and Protective Role in the Biofilm Matrix of Pseudomonas aeruginosa
|
There is considerable debate as to the nature of the primary parasite-derived moieties that activate innate pro-inflammatory responses during malaria infection . Microparticles ( MPs ) , which are produced by numerous cell types following vesiculation of the cellular membrane as a consequence of cell death or immune-activation , exert strong pro-inflammatory activity in other disease states . Here we demonstrate that MPs , derived from the plasma of malaria infected mice , but not naive mice , induce potent activation of macrophages in vitro as measured by CD40 up-regulation and TNF production . In vitro , these MPs induced significantly higher levels of macrophage activation than intact infected red blood cells . Immunofluorescence staining revealed that MPs contained significant amounts of parasite material indicating that they are derived primarily from infected red blood cells rather than platelets or endothelial cells . MP driven macrophage activation was completely abolished in the absence of MyD88 and TLR-4 signalling . Similar levels of immunogenic MPs were produced in WT and in TNF−/− , IFN-γ−/− , IL-12−/− and RAG-1−/− malaria-infected mice , but were not produced in mice injected with LPS , showing that inflammation is not required for the production of MPs during malaria infection . This study therefore establishes parasitized red blood cell-derived MPs as a major inducer of systemic inflammation during malaria infection , raising important questions about their role in severe disease and in the generation of adaptive immune responses .
Severe malaria in humans is a leading cause of morbidity and mortality , especially in sub-Saharan Africa [1] . The clinical manifestations of severe malaria are directly correlated with the induction of strong pro-inflammatory type-1 immune responses . Thus , whilst it is clear that early innate and T cell pro-inflammatory immune responses are essential for the control of malaria infection [2] , [3] , excessive production of pro-inflammatory cytokines , including IL-6 , TNF and IFN-γ , may also directly contribute to severe disease , such as severe anaemia , cerebral malaria ( CM ) and organ damage [3] , [4] . It is therefore crucial that the most potent parasite dependent and independent pro-inflammatory triggers are identified , and their signalling pathways unravelled , before targeted , successful therapeutic treatments can be developed for malaria infection . Activation of macrophages is a key event in the pathogenesis of severe malaria in both humans [5] and in experimental models of malaria [6] . P . berghei ANKA ( PbA ) infection of C57BL/6 mice , which is the best available model of CM , is characterized by the development of strong pro-inflammatory immune responses , including macrophage activation and the production of TNF , IL-12 , IL-1β , IL-6 , ROI and NO [2] , [3] . Activation of brain resident and brain-homing monocytic cells , leading to activation of brain vascular endothelial cells and consequent sequestration of pRBC and leucocytes , is believed to be a key stage in the development of the neuropathology associated with experimental cerebral malaria ( ECM ) during PbA infection . In addition , although splenic and liver macrophage populations have been shown to be required for optimal parasite control [7] , [8] , excessive macrophage responses in these organs has been directly correlated with malarial anaemia and liver damage [9] . At present there is considerable debate about the pathways driving inflammation during malaria infection . Interaction of malaria parasite-derived moieties with cells of the innate system , such as macrophages and dendritic cells , is likely to be the initial step in induction of the inflammatory response; however , despite intense research , there is no agreement regarding the identity of the primary parasite products that initiate the pro-inflammatory cascade [4] and the importance of Toll like receptor signalling and scavenger receptors ( such as CD36 ) in the recognition of parasite products and subsequent production of inflammatory cytokines remains unclear [10]–[21] . Parasitized red blood cells ( pRBC ) have been shown , depending on the model and the duration of stimulation , to either induce or suppress macrophage and dendritic cell function , including induction or suppression of pro-inflammatory cytokine production [21]–[26] . The malaria pigment hemozoin has been proposed as a novel TLR-9 ligand , inducing TNF , IL-6 and IL-12 p40 production [12] , but it has subsequently been suggested that contaminating malarial DNA , which binds to hemozoin , is responsible for TLR-9 activation [27] . However , hemozoin has also been shown to directly suppress macrophage and dendritic cell function [28]–[30] . In recent years attention has focussed on the potential role of parasite glycosylphosphatidyl-inositol ( GPI ) [19] , [20] , [31] , [32] , which is capable of inducing TNF secretion by macrophages via signalling through TLR-2 [20] and the scavenger receptor CD36 [19] and which , in one report , induced cachexia when injected into mice [31] . Most recently it has been shown that plasma-derived microparticles ( MPs ) from malaria-infected mice can induce TNF production by macrophages [33] suggesting that MPs may also contribute to the systemic inflammation that is characteristic of malaria infection . MPs are sub-micron particles ( 0 . 1–1µm diameter ) produced by vesiculation or ‘blebbing’ of the plasma membrane of cells as a result of loss of asymmetry of the phospholipid bilayer ( reviewed [34] , [35] ) . In healthy animals , circulating MPs are predominantly derived from platelets , but , depending on the situation , MPs may also be produced by leucocytes , endothelial cells and erythrocytes ( reviewed [34] , [35] ) . Vesiculation of the phospholipid bilayer and MP development is a tightly-regulated homeostatic process that occurs at an increased rate during cell activation and during apoptotic or necrotic cell death . MP formation is directly correlated with TNF and IL-1β production [34] , [35] . Thus , although basal levels of MP are found in the blood of healthy donors , elevated levels have been detected in many pathological conditions [34] , [35] including cerebral malaria [36] . Consistent with results obtained in humans , significantly higher numbers of circulating MPs are found in PbA-infected mice than in uninfected controls [33] , [35] . In this study we have investigated the ability of MPs induced during acute malaria infections in mice to stimulate macrophage pro-inflammatory responses and we have assessed the potential relevance of this pathway in the development of severe malarial disease . We find that malaria infection-induced MPs are much more potent inducers of macrophage activation than are intact , live pRBC and that pRBC-derived MPs , rather than endothelial- , leukocyte- or platelet-derived MPs , are the primary inducers of macrophage activation . Furthermore , we have defined a TLR-4- and MyD88-dependent pathway of MP-induced macrophage activation . This study establishes a major new pathway of innate inflammation during malaria infection which implicates pRBC-derived MPs as major contributors to the development of severe malarial disease .
To investigate the ability of Plasmodium berghei ANKA ( PbA ) -induced MPs to activate macrophages in vitro , bone marrow-derived macrophages were challenged for 24hrs with purified PbA infection-induced MPs , and uninfected MPs derived from naive mice; macrophage activation was assessed by up-regulation of CD40 expression and by the production of TNF . The size and granularity of the purified MP preparations relative to 1µm beads ( gated population: upper right hand side ) is shown in representative plots in Fig . 1A . As can be observed , PbA infection-induced MPs and uninfected MPs were sub-cellular in size and approximately 98% of all flow cytometric events within the MP preparation were <1µm in size . A number of flow cytometric events were observed in the PBS control , but these events were in general smaller than those observed in the MP preparations and were due to minor contaminations within the solution . PbA infection-induced MPs were homogenous in size and the majority of MPs were approximately 150–250nm in diameter when examined by scanning electron microscopy ( Fig . 1B ) . Uninfected MPs were heterogeneous in size compared with PbA-infection induced MPs , varying from approximately 75µm to 450µm in diameter . Irrespective of size , uninfected MPs and PbA infection-induced MPs were comparable in morphology and appeared spherical in appearance ( Fig . 1B ) . As expected , incubation with PBS ( no stimulation ) failed to induce macrophage activation , as measured by CD40 expression and TNF production ( Fig . 1C–E ) . Similarly , stimulation with control MPs from uninfected mice also failed to induce up-regulation of CD40 expression or the production of TNF , indicating that MPs derived from uninfected mice are non-inflammatory ( Fig . 1C–E ) . In contrast , strong macrophage activation was observed following stimulation with PbA-induced MPs ( using a comparable volume of the MP preparation as used for control MPs ) , with significantly elevated expression of CD40 and increased production of TNF compared with non-stimulated and uninfected MP stimulated controls ( Fig . 1C–E ) . These data confirm the results obtained by Combes et al [33] showing that PbA-induced MPs can stimulate TNF production by macrophages . To ensure that macrophage activation was not an artefact of LPS contamination of the MPs , endotoxin concentrations were tested in all the MP preparations and were found to be less than 0 . 24 IU/ml in all cases; the minimum concentration of LPS required to activate macrophages is approx 0 . 6 IU/ml ( data not shown but provided for review ) . Significantly increased numbers of flow cytometric events were found within PbA infection-induced MP preparations compared with uninfected MP preparations ( Fig . 1F ) : thus , it was foreseeable that the ability of PbA induced MPs - but not uninfected MPs - to stimulate macrophages was related to a quantitative difference in the number of inflammatory MP particles , over an unspecified threshold level , rather than an intrinsic qualitative difference in the immunogenicity of the MP preparations . To examine this we performed a dose response experiment using different volumes of PbA induced and uninfected MP preparations and examined the ability of the MPs to activate macrophages: uninfected MPs derived from naive mice failed to stimulate CD40 up-regulation or TNF production at any of the tested concentrations , whereas PbA induced MPs promoted up-regulation of CD40 expression and production of TNF in a dose dependent manner at 5 , 10 and 50µl volumes , equating to approximately 5×104 , 1×105 and 5×105 PbA MPs/well respectively ( Fig . 1G–I ) . In addition , uninfected MPs failed to stimulate macrophage activation when the number of flow cytometric events of the preparation was normalised to PbA MPs numbers ( results not shown ) . The kinetics of macrophage activation by PbA-induced MPs or LPS were compared . Although both LPS and PbA-induced MPs induced a maximal TNF response within 6 hrs , maximal CD40 induction was slower for PbA-induced MPs ( 24–48 hrs ) than for LPS ( 12hrs ) ( Fig . S1 ) . Taken together , these results demonstrate that MPs derived from malaria infected mice induce potent macrophage activation and are significantly more inflammatory on a particle to particle basis than MPs from normal , uninfected mice . Our results in Figure 1 demonstrated the ability of PbA infection induced MPs to stimulate pro-inflammatory responses in vitro , but these results did not specifically address the potential relevance of this pathway in the generation of inflammatory responses during malaria infection . Consequently , to examine the importance of MP-induced macrophage activation during malaria infection , and how this may relate to other parasite-specific pathways of macrophage activation , we compared the ability of infection-induced MPs and live intact PbA pRBC to promote macrophage activation . As expected , stimulation with uninfected red blood cells failed to induce up-regulation of CD40 expression or the production of TNF by macrophages ( Fig . 2A–C ) . Interestingly , stimulation with live intact parasitized red blood cells ( >80% purity , mainly late trophozoites and schizonts ) at 1∶1 , 10∶1 and 100∶1 ratios of parasites to macrophages also failed to induce up-regulation of CD40 expression or the production of TNF ( Fig . 2A–C ) . These results are surprising as strong macrophage activation , including pro-inflammatory cytokine production , has been reported following in vitro stimulation with P . falciparum schizont infected RBC [22] . In contrast , although a number of studies have clearly demonstrated that phagocytosis of murine pRBC by macrophages occurs in vitro [36]–[38] , there is very little evidence to suggest that this leads to up-regulation of co-stimulatory receptor expression on macrophages , or the production of pro-inflammatory cytokines [39] . Nonetheless , stimulation with infection induced MPs promoted significant up-regulation of CD40 expression and production of TNF by macrophages ( Fig . 2A–C ) . These results show that plasma derived MPs may exert a dominant pathway in driving macrophage activation during malaria infection , either causing much of the inflammation and pathology of infection , or initiating anti-malaria immune responses . MPs derived from malaria infected mice were considerably more inflammatory than uninfected MPs derived from naive mice ( Fig . 1 ) , even when the numbers of particles in each preparation was normalised , suggesting that MPs derived from malaria infected mice are more immunogenic than uninfected MPs . MPs can be produced by the vesiculation of the membrane of many different cell populations , including platelets , leukocytes , endothelial cells and red blood cells , a process that is modulated during inflammatory episodes [35] . Consequently , the predominant cellular source of the MPs may change during malaria infection , and this alteration in cellular source could explain the difference in ability to promote macrophage activation . To address this likelihood , we performed a phenotypic characterisation of the PbA infection derived and uninfected MP populations . We first assessed the expression of Annexin V , a marker of cellular apoptosis , which is the standard marker of classical inflammation-driven microparticles [35] . As expected based on previous reports [33] , [40] , we observed a significant and marked increase in the expression of Annexin V on PbA induced MPs compared with uninfected MPs , both in terms of frequency and total numbers of positive events ( Fig . 3A , B ) . Very low Annexin V staining was observed on the PBS control FACS events , demonstrating the specificity of the flow cytometric staining ( Fig . 3A , B ) . Importantly , not all infection derived ( or uninfected ) MPs expressed AnnexinV , indicating that a large proportion of particles within the MP preparation are not classically defined or produced MPs . To examine the cellular source of AnnexinV+ ( and AnnexinV− ) microparticles we employed a panel of antibodies to cover all the major potential cellular sources of MPs ( Fig . 3C ) . Platelets ( CD41 ) were the major source of classical AnnexinV+ microparticles within the uninfected MP population , with a number of red blood cell ( TER119 ) derived Annexin V+ MPs also found ( Fig . 3C , D ) . Few AnnexinV+ microparticles within the uninfected MP population were produced from endothelium ( CD144 ) , leukocytes ( CD45 ) and macrophages ( F4-80 ) and only a small number co-expressed CD107a , meaning most were not exosomes , or VCAM-1 , indicating that they did not emanate from activated endothelium ( Fig . 3C , D ) . In contrast the majority of AnnexinV+ MPs derived from PbA infected mice did not co-express CD41 , suggesting that they were not platelet derived . Approximately 20% of PbA infection induced AnnexinV+ microparticles appeared to be derived from red blood cells , but the majority of the AnnexinV+ microparticles failed to co-stain with any tested antibody , meaning that their cellular origin is undefined ( Fig 3C , D ) . Nevertheless , when breaking down the cellular sources of all the flow cytometric events within the PbA infection induced and uninfected MP preparations , irrespective of AnnexinV expression , a clear increase in the frequency and numbers of TER119+ ( RBC derived ) and CD45+ ( leukocyte derived ) flow cytometric events was observed within the PbA infection induced MP population ( Fig . 3C , E and F ) . The phenotypic characterisation of PbA infection-induced MPs and uninfected MPs demonstrated clear differences in the expression level of AnnexinV , TER119 and CD45 , suggesting that the predominant cellular sources of PbA infection induced MPs and uninfected MPs varied , potentially explaining the inflammatory nature of the malaria infection induced MPs . As the frequency and numbers of RBC-derived particles was increased in the PbA infection induced MP preparation , we hypothesised that infection-induced MPs may either be formed by vesiculation of the pRBC membrane during intra-erythrocytic parasite maturation and/or pRBC rupture at schizogeny , in which case MPs would be expected to contain malaria parasite-derived components , or that they may be derived from uninfected RBCs , which are known to undergo bystander lysis during acute malaria infection , contributing to the rapid onset of anaemia [3] . To distinguish between these two possibilities , we analysed PbA infection-induced MPs for the presence of PbA-specific antigens by immunofluorescence using purified anti-PbA IgG . As expected , no parasite-derived material was detected in uninfected MP preparations ( Fig . 4A ) . In contrast a large proportion of MPs from PbA-infected mice bound the anti-PbA IgG ( Fig . 4A ) , indicating the presence of significant quantities of parasite-derived material in the infection-derived MPs . Although a number of parasite moieties are likely to be incorporated within the malaria infection induced MP preparation , we failed to detect hemozoin in the plasma derived preparation by beta-hematin formation assay ( results not shown ) . The observation that a large proportion of PbA-infection induced MPs were derived from pRBC and were predominantly Annexin V− suggested that they were not classical inflammation-induced MPs . Nevertheless , PbA-infection induced MPs also displayed heterogenous expression of Annexin V and the frequency and numbers of Annexin V+ MPs increased during malaria infection , showing that inflammatory MPs were also produced during infection . Since it is not feasible to efficiently separate the two populations of MPs from the plasma of PbA-infected mice , to determine whether macrophage activation was induced by the pRBC-derived MPs or by the more classical Annexin V+ MPs , we generated a pure population of pRBC-derived MPs in vitro from purified and extensively washed pRBC and compared their macrophage activating properties with MPs from the plasma of PbA-infected or uninfected mice , with in vitro generated MPs from uRBC ( Fig . 4 ) and with MPs purified from the plasma of mice treated with LPS to induce inflammation ( Fig . 5 ) . The numbers of MPs in each preparation were counted by flow cytometry and were normalised to the number of MPs from uninfected mice prior to culture with BMDM . The in vitro-derived MPs were of similar size to those prepared from plasma ( data not shown ) . In vitro-derived MPs were predominantly ( >80% ) TER119+ , and although a proportion of the in vitro-derived MPs ( <20% ) expressed Annexin V+ , the levels of Annexin V expression were substantially lower ( MFI of 40 . 0 for in vitro-derived MPs vs MFI of 98 . 5 for infection-derived MPs ) than for infection-derived MPs ( Fig . 4B ) . In support of our hypothesis that parasite material bound to RBC membranes within the MPs is responsible for macrophage activation , we observed significant up-regulation of CD40 expression ( Fig . 4C , D ) and production of TNF ( Fig . 4E ) by macrophages cultured with in vitro-derived pRBC MPs , which was comparable to the activation observed with the PbA infection-derived plasma MPs ( Fig . 4C–E ) . In contrast , MPs derived in vitro from uRBC did not induce macrophage activation . As the vast majority of pRBC derived MP particles expressed TER119 ( Fig . 4B ) , it is unlikely that soluble non-membrane associated parasite materials were purified during the generation of the MP preparation or that non-membrane bound parasite materials were responsible for the macrophage activation . The plasma of mice injected 3 days previously with LPS - to induce inflammation - contained MPs that expressed high levels of Annexin V+ ( Fig . 5A ) and did not stain with the anti-PbA antiserum ( data not shown ) . LPS was a more potent stimulus of classical Annexin V+ MP generation than PbA infection , leading to an increase in both the frequency ( Fig . 5B ) and total numbers ( Fig . 5C ) of Annexin V+ MPs . Although fewer Annexin V+ MPs were present in the plasma of PbA-infected mice , the phenotype of these Annexin V+ MPs was very similar to that of the LPS-induced MPs , based upon TER119 and CD41 co-staining ( Fig . 5D ) , indicating that classical inflammation-induced MPs - derived from comparable cellular sources - were present in both MP preparations . Despite this , the LPS-induced MPs failed to induce macrophage activation whereas , as shown previously , PbA-infection induced MPs induced up-regulation of CD40 expression and production of TNF ( Fig . 5E–G ) . Taken together , these data demonstrate that classical , inflammation-driven MPs do not directly induce pro-inflammatory immune responses in macrophages but that the atypical MPs generated from pRBC have potent pro-inflammatory activity . To determine whether the timing of inflammatory MP generation correlated with the onset of clinical signs , the numbers and the inflammatory potential of MPs isolated from mice on days 3 , 5 and 7 of infection were compared ( Fig . 6 ) . Although the total number of plasma MPs increased only slightly over the course of infection ( Fig . 6A ) , numbers of TER119+ erythrocyte-derived MPs increased steadily over the course of infection ( Fig . 6B ) in line with the steadily increasing parasitaemia ( Fig . 6C ) . Importantly , however , day 3 and day 5 MPs were only poorly pro-inflammatory ( Fig . 6D , E ) and only day 7 MPs were able to induce significant macrophage CD40 expression and TNF production , suggesting either that there is a threshold concentration of MPs required for macrophage activation or that day 7 MPs are qualitatively different from day 3 or day 5 MPs . In either event , there is clearly a close temporal association between the accumulation of highly pro-inflammatory MPs and the onset of severe malarial disease . We have shown that pRBC-derived microparticles are phenotypically distinct from classical , inflammation-induced microparticles and , in contrast to LPS-induced MPs , possess potent pro-inflammatory activity . Nonetheless , AnnexinV+ inflammation-driven MPs have been reported to induce pro-inflammatory immune responses in other models [35] . Thus , to further explore the role of inflammation and inflammation-induced MPs in the response of macrophages to PbA-induced MPs , we prepared MPs from the plasma of PbA-infected ( day 7 p . i . ) TNF−/− , IL-12p40−/− , IFN-γ−/− and RAG-1−/− mice , all of which have major defects in their innate inflammatory response , and compared them with MPs from WT PbA-infected mice ( Fig . 7 ) . Numbers of MPs in each preparation were counted ( by flow cytometry ) and adjusted to uninfected MP numbers . Similar levels of immunofluorescence were seen when MPs were labelled with anti-PbA antiserum ( Fig 7A ) , indicating that comparable amounts of parasite material was found in each preparation . Importantly , MPs derived from PbA-infected TNF−/− , IL-12p40−/− , IFN-γ−/− , RAG-1−/− and WT mice all induced very similar levels of macrophage activation , as shown by up-regulation of CD40 expression and TNF production ( Fig . 7 B–D ) . Taken together , these data strongly suggest that parasite moieties within the membranes of these atypical , pRBC-derived MPs are responsible for macrophage activation and that these MPs can be generated in the absence of inflammation . We next investigated the pathways required for macrophage activation by malaria infection induced MPs . As our data indicated that parasite material , bound to RBC membrane within the MP preparation was responsible for driving macrophage activation , we hypothesised that TLRs and the adaptor molecule MyD88 may be required for macrophage stimulation . TLR molecules and MyD88 have previously been shown to be required for optimal pro-inflammatory cytokine production during malaria infection [10] , [12] , [17] , [20] . Our results clearly show that PbA infection-induced MP activation of macrophages is MyD88 dependent , as macrophage activation , as measured by CD40 up-regulation and TNF production , was completely ablated in MyD88−/− macrophages ( Fig . 8A–C ) . MP-induced macrophage activation was totally TLR-4 dependent; up-regulation of CD40 expression and induction of TNF production were both completely absent in TLR-4−/− macrophages ( Fig . 8D–F ) . Interestingly , however , PbA infection-induced MP stimulation of TNF production ( but not CD40 expression ) was also significantly lower in TLR-2−/− and TLR-9−/− macrophages than in WT macrophages . These data would be consistent with the presence of low levels of TLR-2 and TLR-9 ligands such as GPI and hemozoin within the MPs . As the response was completely ablated in TLR-4−/− macrophages - demonstrating that TLR-4 is essential for TLR-4 responsiveness - our results suggest that a primary TLR-4 ligand within the MPs activates the cells and that TLR-2 and TLR-9 ligands synergise with the TLR-4 stimulus to induce maximal macrophage activation .
Understanding the pathways leading to inflammation during malaria infections should allow the development of new approaches to therapy and more immunogenic vaccines . In this study we have identified an entirely novel pathway of inflammation during malaria infection , namely TLR-4/MyD88-mediated activation of macrophages by membrane microparticles emanating from parasitized red blood cells . Importantly , we have shown that malaria-infection induced MPs promote significantly stronger macrophage activation than live infected red blood cells , underlining the potential significance of MP-induced inflammation during malaria infection . The primary TLR-4 dependence of this pathway sets it apart from the previously described glycoslylphosphatidylinositol ( GPI ) -TLR-2/CD36 pathway [19]–[20] and the hemozoin/parasite DNA/TLR-9 pathway [12] . Previous studies on MPs during malaria infection have largely been in the context of their role in the pathogenesis of severe disease: circumstantial evidence supports a role for MPs in severe P . falciparum infection [36] . In addition , ABCA1 KO mice that are defective in the ability to produce MPs are protected against ECM during P . berghei ANKA infection [33] . Endothelial and platelet derived MPs have been shown to “bridge” endothelial cell and pRBC and leukocyte interactions , allowing sequestration within brain microvessels , which is a key factor in initiation of cerebral pathology [41] , [42] . However , other than the studies in ABCA1 KO mice , which have numerous defects in lipid metablism and macrophage function that might influence their susceptibility to ECM [43] , there is no causal evidence to link MPs with pathogenesis in vivo . The results of this current study have added to the complexity of the potential roles for MPs during malaria infection . We have shown that malaria infection induced MPs are capable of promoting the up-regulation of CD40 expression and TNF secretion from bone marrow derived macrophages in vitro; thus , malaria infection induced MPs promote potent activation of innate and adaptive immune responses , which is likely to have major significance in the development of inflammation during infection . Our data importantly differs from earlier studies on MPs during malaria infection , as we have defined the cellular source of immunogenic MPs as infected red blood cells , containing large quantities of parasite-material , rather than platelets or endothelial cells [36] , [40] , [40]–[42] . pRBC-derived MPs were phenotypically and functionally distinct from the classical , Annexin V+ microparticles that emanate primarily from platelets but also from leucocytes and endothelial cells during systemic inflammation [34] , [35] . Not only were pRBC-derived MPs produced in the absence of key inflammatory mediators such as TNF , IL-2 and IFN-γ but also , in our hands , classical inflammation-derived MPs had minimal macrophage activating capacity and are thus clearly products of - rather than drivers of - inflammation . One particularly striking observation was that pRBC-derived MPs are much , much more potent macrophage activators than are live , intact pRBC . This suggests that MPs may be released from pRBC only at certain very specific stages of parasite development , possibly associated with the membrane disintegration that is seen immediately prior to schizont rupture [44] . Although mature pRBC are expected to rupture and produce MPs during the course of the 24hr culture , the concentration of MPs generated may have been too low to activate the macrophages . Alternatively , phagocytosis of intact pRBC by macrophages via CD36-dependent pathways [39] may have prevented generation of free microparticles or suppressed the subsequent TLR-4 mediated signalling promoted by MPs . In agreement with the latter hypothesis , macrophage activation through anti-CD40 or TLR stimulation is suppressed following the phagocytosis of non-activating latex beads [45] and TLR-tolerance has been shown to occur during malaria infection [46] . The importance of parasite material within the MPs for stimulating macrophage activation raises intriguing questions regarding the nature and identity of the pro-inflammatory parasite molecules within the MPs . GPI , the membrane anchor for MSP-1 and MSP-2 , and hemozoin , the product of haemoglobin breakdown by the parasite , have been shown to promote [12] , [19] , [20] , [31] or suppress innate activation [28]–[30] . While hemozoin is thought to stimulate macrophage activation through ligation with TLR-9 [12] , [27] , activation by GPI is mediated by TLR-2 and CD36 through downstream activation of ERK , p38 , MAPK , JNK and NFκb signalling pathways [19] , [20] , [47] , [48] . The absolute dependence of MP-mediated macrophage activation on TLR-4 signalling makes it most unlikely that MPs are simply vehicles for GPI and hemozoin . Furthermore , plasma-derived PbA infection-induced MPs did not contain measurable levels of hemozoin and we have found that scavenger receptor A and B family members ( including CD36 ) are not required for MP-induced macrophage activation ( results not shown ) . These data suggest that the main inflammatory parasite materials within the MP preparation are unlikely to be GPI or hemozoin . Nevertheless , TLR-2 KO and TLR-9 KO macrophages produced significantly lower levels of TNF following stimulation with PbA infection-induced MPs , indicating that TLR-2 and TLR-9 signalling is required for optimal TNF production . These observations are highly consistent with a scenario in which PbA infection-induced MP recognition by TLR-4 initiates the macrophage response and macrophage activation is then amplified by GPI/hemozoin signalling through TLR-2 and TLR-9 . Co-operation and synergy of distinct TLR signalling pathways in response to complex TLR ligands is becomingly increasingly well recognised in a number of different systems [49] , [50] . The essential role for MyD88 signalling for macrophage activation by malaria-infection derived MPs is entirely consistent with a number of studies demonstrating a role for MyD88 in malarial inflammation and pathology . For example TLR/MyD88 mediated IL-12 production is responsible for liver injury during P . berghei NK65 infection [9] and TNF production induced via MyD88 signalling promotes weight loss and fever during P . chabaudi AS infection [17] . Moreover , although the role of MyD88 dependent signalling in P . berghei ANKA induced ECM remains unclear with conflicting findings [10] , [11] , [13] , [14] , MyD88 signalling is required for optimal macrophage TNF , IL-6 and IL-1α production [10] . Our data suggest that pRBC-derived MPs may be a significant inducer of all these effects . In conclusion , we have identified a novel and very potent , MyD88- and TLR-4-dependent pathway of inflammation during malaria infection that is mediated by pRBC-derived membrane microparticles . Interestingly , a recent study has shown a link between polymorphisms in the TLR-4 locus and susceptibility to severe malaria in humans [51]; our data offer a plausible biological explanation for this observation . We expect that pRBC microparticles will synergise with GPI and hemozoin to induce the extraordinarily high levels of circulating inflammatory mediators that are seen in many patients with acute malaria . However , the ability of MPs to induce expression of molecules such as CD40 on antigen presenting cells suggests that they might also play a role in T cell priming and T effector cell function . Future studies will need to identify the parasite ligands presented by these microparticles and explore their potential role as adjuvants for malaria vaccines .
Animal experimentation was approved under UK Home Office Regulations and was subject to LSHTM ethical review . Female , 8–12 week old C57BL/6 wild type , IFN-γ−/− , TNF−/− , IL-12−/− , and RAG-1−/− , were obtained from Harlan and maintained under barrier conditions . Cryopreserved Plasmodium berghei ANKA parasites were passaged once in vivo for a maximum of 4 days before being used to infect experimental animals . Mice were infected intraveneously with 104 parasitised red blood cells and parasitaemia was determined daily by examination of Giemsa-stained thin blood smears . MPs were prepared as described before [33] . Briefly , blood was collected aseptically in 0 . 124M sodium citrate solution from naive or malaria-infected mice ( day 6 or 7 of P . berghei ANKA infection ) , and centrifuged at 1 , 500g for 15 minutes at room temperature . The platelet poor plasma supernatant ( PPP ) was collected and further centrifuged at 13 , 000g for 3mins to obtain platelet free plasma ( PFP ) . This was diluted 1∶3 with citrated PBS containing heparin and centrifuged at 14 , 000g for 90 minutes at 15°C to produce the MP pellet which was then resuspended in sterile PBS . MPs were quantified by flow cytometry as numbers of events relative to a standardised number of 1µm beads ( BS Partikels , GMBH , Germany ) . Unless otherwise stated , blood from two mice was pooled to generate each MP preparation . All microparticle preparations and negative control samples were tested for LPS contamination using the Limulus Amebocyte Lysate ( LAL ) gel formation test , performed according to manufacturer's standard operating procedures ( Health Protection Agency , UK ) . P . berghei ANKA pRBC were collected on day 6 or day 7 of infection and enriched using LD column magnetic cell sorting ( Miltenyi Biotec ) . pRBC were routinely >80% pure and were mainly mid- to late-stage trophozoites and schizonts . pRBC were washed three times in PBS to remove all plasma constituents and the final pellet was resuspended to the original volume . To make microparticles , pRBC or equivalent numbers of RBC from uninfected mice ( uRBC ) were subject to repeated ( 3X ) combinations of freeze-thaw and ultra-sonication ( 10 sec/pulse ) cycles . pRBC or uRBC lysates were then centrifuged at 13 , 000g for 3min to remove particulate material and the supernatant was diluted 1∶3 with citrated PBS containing heparin and centrifuged at 14 , 000g for 90 minutes at 15°C to produce the MP pellet . All microparticle preparations and negative control samples were tested for LPS contamination using the Limulus Amebocyte Lysate ( LAL ) gel formation test , performed according to manufacturer's standard operating procedures ( Health Protection Agency , UK ) . Bone marrow derived macrophages harvested from femurs of wild type and knockout mice were prepared as described previously [52] . Briefly , bone marrow cells were washed and suspended in DMEM supplemented with 10% heat-inactivated FCS , 20% L-cell supernatant ( a source of CSF-1 ) , 5% horse serum ( SIGMA ) , L-glutamine ( GIBCO ) and penicillin and streptomycin ( GIBCO ) and cultured in tissue culture Petri Dishes ( Sterilin , UK ) . After 7 days the supernatant containing fibroblasts and mature macrophages was removed . Adherent cells were scraped off gently , washed , diluted 1∶3 and cultured to maturity for a further 4 days . Macrophages were cryo-preserved until required . BMDM were cultured at 106/ml in duplicate or triplicate in 96-well plates ( NUNC ) either in DMEM alone or with LPS ( 200ng/nl ) , pRBC/uRBC ( at various BMDM∶RBC ratios ) or MPs ( normalised numbers in each experiment relative to uninfected MP levels: average 3–5×105/well , depending on the experiment ) for 24 hours at 37°C in 5% CO2 ) . Supernatants were collected and assayed for TNF and the macrophage monolayer was harvested for flow cytometric analysis . Supernatants were assayed by standard capture ELISA using Immunolon 4 HBX plates ( ThermoLabsystems UK ) coated with monoclonal hamster antibody against murine TNF ( TN3 ) ( gift from Celltech , Slough , UK ) . Bound TNF was detected with a biotinylated goat anti-mouse TNF-α antibody ( R&D , UK ) , streptavidin peroxidase ( Sigma , UK ) and 0-phenylenediamine ( Sigma , UK ) . Recombinant murine TNF ( R&D , UK ) was used as a standard . MPs were quantified relative to a standardised number of 1µm reference beads ( BS Partikels , GMBH , Germany ) : 5µl of MPs were diluted in 200µl of sterile PBS , with 5 drops of 1µm beads added into 4mls PBS . Phenotypic characterisation of MPs was performed by incubating 10µl MPs , diluted 1∶4 in sterile PBS , with anti-mouse TER119 ( clone: TER119 ) , anti-mouse CD41 ( clone: MWReg30; BD Pharmingen ) , anti-mouse VCAM ( clone: 429 ) , anti-mouse CD45 ( clone: 30F11 ) , anti-mouse CD144 ( clone: eBioBV13 ) , anti-mouse CD107a ( clone: eBio D4B ) or anti-mouse F4-80 ( clone: BM8 ) . All antibodies , unless otherwise stated , were from E-Bioscience ( distributed by Insight Biotechnology , UK ) . MPs were then suspended in Annexin V ( BD Pharmingen ) binding buffer before being co-stained with Annexin V . MPs were then diluted to a final volume of 250µl by the addition of sterile PBS containing 1µm reference beads . Macrophage activation was examined by surface staining using anti-mouse CD40 ( clone: 1C10 ) , anti-mouse MHC-II ( clone: M5/114 . 15 . 2 ) and anti-mouse F4-80 ( clone: BM8 ) . Note , significantly altered acquisition settings were applied for flow cytometric detection of MPs compared with those usually utilised for detection of leucocytes; FSc voltage on E02 was adjusted to position 1µm reference beads in the upper right hand area of the plot . Antiserum to P . berghei ANKA-infected RBC was prepared from mice that had undergone 3 rounds of infection and drug cure and IgG was purified on Protein-G ( HiTrap , Amersham , UK ) . Ten microlitres of purified MPs were air dried and acetone fixed on gelatin-coated glass slides . Slides were blocked with rat serum prior to incubation with anti-PbA IgG for 1 hr at room temperature . Following incubation with anti-PbA IgG , slides were visualised using FITC rat anti-mouse antibody ( clone 11-4011-85: E-Bioscience ) by fluoresescence microscopy ( Zeiss , Axioplan 2 ) using Volocity software ( Improvision ) . Microparticles were purified from naïve or day 7 P . berghei ANKA infected mice as described above . After the final centrifugation at 14 , 000g for 90min the pellet was resuspended in 30 µl of PBS and MPs were seeded on to pre-coated Poly-L-Lysine-coated glass coverslips ( Poly-L-Lysine , Sigma ) and allowed to adhere overnight at 4°C in a moist chamber . The adhered MPs were fixed with 30 µl of 2 . 5% glutaraldehyde and 30 µl of 4% paraformaldehyde for 20 min at room temperature . The 60 µl of fixatives were removed and replaced with fresh solutions of 2 . 5% glutaraldehyde and 4% paraformaldehyde and left overnight at 4°C in the moist chamber . The coverslips were then washed twice in PBS 0 . 2 M and left at 4°C until processing . Post-fixation was carried out with osmium for 1h at room temperature and the preparations were dehydrated in solutions of ethanol of increasing strength from 70% , 80% and 90% for 10 min in each solution . Coverslips were finally dehydrated twice for 10min in 100% ethanol and then twice for 5min with 100% ultra-pure ethanol . After rapidly immersing for 3 min in hexamethyldisilazane ( Sigma ) , the preparations were air dried and transferred to a desiccator overnight . Samples were mounted onto , gold coated stubs for 2 min for imaging on a Zeiss scanning electron microscope ( Zeiss ULTRA plus ) . Statistical significance was determined using two tailed Student's T test , unless otherwise stated , with P<0 . 05 considered as significant .
|
Although parasite materials are responsible for the activation of the immune system during malaria infection , exactly how the immune response is initiated during infection is extremely unclear . In this study we demonstrate that sub micron particles ( microparticles ) are produced by malaria infected red blood cells during malaria infection , and we show that these microparticles can promote strong inflammatory responses by activating macrophages . We show that infected red blood cell-derived microparticles are produced in higher numbers as infection progresses , and that the host's own pro-inflammatory immune response is not required for the generation of these microparticles . We have also examined the receptors and signalling pathways required for macrophage activation by microparticles , and we show that the pathway of microparticle-induced activation is distinct from other previously reported pathways . In summary , we have defined a novel pathway of immune response activation during malaria infection , which may be important for promoting parasite control and/or causing pathology .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/protozoal",
"infections",
"pathology/immunology",
"immunology/immunity",
"to",
"infections",
"immunology/innate",
"immunity"
] |
2010
|
Parasite-Derived Plasma Microparticles Contribute Significantly to Malaria Infection-Induced Inflammation through Potent Macrophage Stimulation
|
Constitutive heterochromatin is enriched in repetitive sequences and histone H3-methylated-at-lysine 9 . Both components contribute to heterochromatin's ability to silence euchromatic genes . However , heterochromatin also harbors hundreds of expressed genes in organisms such as Drosophila . Recent studies have provided a detailed picture of sequence organization of D . melanogaster heterochromatin , but how histone modifications are associated with heterochromatic sequences at high resolution has not been described . Here , distributions of modified histones in the vicinity of heterochromatic genes of normal embryos and embryos homozygous for a chromosome rearrangement were characterized using chromatin immunoprecipitation and genome tiling arrays . We found that H3-di-methylated-at-lysine 9 ( H3K9me2 ) was depleted at the 5′ ends but enriched throughout transcribed regions of heterochromatic genes . The profile was distinct from that of euchromatic genes and suggests that heterochromatic genes are integrated into , rather than insulated from , the H3K9me2-enriched domain . Moreover , the profile was only subtly affected by a Su ( var ) 3–9 null mutation , implicating a histone methyltransferase other than SU ( VAR ) 3–9 as responsible for most H3K9me2 associated with heterochromatic genes in embryos . On a chromosomal scale , we observed a sharp transition to the H3K9me2 domain , which coincided with increased retrotransposon density in the euchromatin-heterochromatin ( eu-het ) transition zones on the long chromosome arms . Thus , a certain density of retrotransposons , rather than specific boundary elements , may demarcate Drosophila pericentric heterochromatin . We also demonstrate that a chromosome rearrangement that created a new eu-het junction altered H3K9me2 distribution and induced new euchromatic sites of enrichment as far as several megabases away from the breakpoint . Taken together , the findings argue against simple classification of H3K9me as the definitive signature of silenced genes , and clarify roles of histone modifications and repetitive DNAs in heterochromatin . The results are also relevant for understanding the effects of chromosome aberrations and the megabase scale over which epigenetic position effects can operate in multicellular organisms .
Constitutive heterochromatin is a nearly universal component of eukaryotic genomes . It was first defined in the 1920′s as distinct from euchromatin by its densely stained cytological appearance [1] . It was also associated with modulation of gene expression in Drosophila chromosome rearrangements that created new euchromatin-heterochromatin ( eu-het ) junctions [2 , 3] . The variable silencing of euchromatic genes located near the eu-het breakpoint was described as position effect variegation ( PEV ) . Striking aspects of PEV included its generality and long-range effects on genes located as far as several megabases away from the breakpoint . Decades later , it was recognized that constitutive heterochromatin is mostly composed of repetitive sequences , including transposable elements ( TEs ) [4] and satellite sequences [5] , and enriched for certain chromosomal proteins , particularly the chromodomain protein HP1 and specific modified forms of histones , such as H3K9me2 or H3K9me3 ( histone H3-di , or tri-methylated-at-lysine 9 ) [6] . Mutations that disrupt the function of these and other heterochromatin-enriched proteins or histone modifying enzymes have been identified as Suppressors of PEV ( Su ( var ) s ) [7] in Drosophila because of their effects on heterochromatin-induced PEV of euchromatic genes . Many SU ( VAR ) proteins are evolutionarily conserved among eukaryotes , including yeast and humans . The fission yeast , Schizosaccharomyces pombe , lacks cytologically defined heterochromatin but its pericentromeric and silent mating-type regions share genetic and biochemical properties , including an enrichment for H3K9me , with the heterochromatin of multicellular eukaryotes . Elegant studies using S . pombe reported two key observations regarding the nucleation and delimitation of H3K9me-enriched domains . First , distinct repetitive DNA elements produce double-stranded RNAs , which are processed into small RNAs and mediate recruitment and assembly of complexes containing the RNAi machinery , H3K9 methyl-transferase , and HP1 homologs [6 , 8–11] . Second , distinct boundary elements exist that limit spreading of the complexes from nucleation sites [11 , 12] . In the silent mating-type region , for example , there is a dramatic transition between H3K4me ( histone H3 methylated-at-lysine 4 ) -domains that contain active genes and H3K9me-domains that contain silent loci , precisely at boundary elements . Deletion of these boundary elements results in leakage of H3K9me from silent domains into neighboring domains over distances of several kilobases [12] . The discovery of small RNAs derived from TE DNAs [13] and effects of mutations in the RNAi machinery on PEV [14] has led some investigators to propose a role for RNAi in the formation of Drosophila heterochromatin [14] . Whether other properties observed for S . pombe may extend to the much larger and more complex heterochromatin of multicellular organisms remains unclear . Heterochromatin can encompass many megabases of DNA in plants and animals , and its influence can extend several megabases into euchromatin [2 , 3 , 7 , 15] . Moreover , heterochromatin contains a complex repertoire and distribution of repetitive elements in organisms such as Drosophila . Little is known about the relative contributions of various types of repetitive elements to heterochromatin formation , although a few studies in Drosophila have singled out the 1360 element as a prime candidate for Chromosome 4 heterochromatin [16] . In addition , the molecular nature of eu-het transition zones remains largely unexplored . The association of heterochromatin , heterochromatin-enriched proteins , and gene silencing is widely accepted [7] , yet hundreds of genes are embedded within heterochromatin in diverse organisms [17 , 18] . At least some of these genes in Drosophila depend on a heterochromatic location and associated chromosomal proteins for normal expression [18] . Compelling evidence is provided by studies of chromosome rearrangements that create new eu-het junctures and result in repressive effects on both euchromatic genes and heterochromatic genes on either side of the breakpoint [3 , 19–21] . In addition , genetic studies have shown that a large number of modifiers of PEV ( Su ( var ) s and E ( var ) s ) act in opposite fashions on variegating euchromatic and heterochromatic genes [22] . Several models have been proposed to account for the expression of heterochromatic genes [18] . The simplest is an “insulation” model which proposes that genes located within cytologically defined heterochromatin actually exist in isolated “islands” of euchromatin and are protected from the surrounding repressive chromatin . This model cannot account for genes that depend on heterochromatin for expression . An “exploitation” model does so , by postulating that heterochromatic genes possess specific sequences , for example specialized promoters or TE-derived regulatory sequences , which require factors enriched in or limited to heterochromatin . In previous study , we eliminated the possibility of novel types of promoter for seven of the best studied Drosophila heterochromatic genes [23] . A third model , “integration , ” proposes that heterochromatin-enriched proteins act more broadly to set up an environment that promotes heterochromatic gene expression . One mode of action may be to facilitate long-range enhancer-promoter interactions through protein-protein interactions . Implicit in this model is the view that heterochromatin-enriched proteins and sequences are multifunctional , with positive , negative , or neutral activities being context dependent . Thus heterochromatic genes may be integrated into a domain , which while incompatible for euchromatic genes , is nonetheless required for efficient expression of heterochromatic genes . To evaluate models of heterochromatic gene expression , we investigated the distribution of H3K9me and other modified histones in and around heterochromatic genes in Drosophila embryos . We also investigated whether SU ( VAR ) 3–9 , which has been characterized as the histone methyltransferase ( HMTase ) primarily responsible for H3K9me in pericentric heterochromatin [24] , affects chromatin modifications of heterochromatic genes . Our analysis provides the first large-scale view of the molecular landscape of modified histones in Drosophila heterochromatic genes and naturally occurring and rearrangement-induced euchromatin-heterochromatin transition zones . The findings are relevant for considering how modified histones and repetitive DNAs contribute to the specification and influence of heterochromatic domains .
Specific lysines in the tail of histone H3 , including Lys 4 , 9 , 27 , and 36 , can be mono- , di- , or tri-methylated . Of the many known H3 methylation states , methylation of Lys 9 ( H3K9me ) has been most extensively studied and has been strongly correlated with gene repression and establishment of heterochromatin [25] . Previous studies using specific antibodies to detect modified H3 on polytene chromosomes showed that H3K9-di-methylation ( H3K9me2 ) is the major H3K9 modification in Drosophila pericentric heterochromatin [24 , 26 , 27] . Thus , we focused on H3K9me2 . We also assayed H3 di-methylated at Lys 4 ( H3K4me2 ) and H3 acetylated at Lys 9 and 14 ( H3K9/14acet ) because they are believed to mark gene expression in euchromatin [28–30] but their presence and distribution in Drosophila heterochromatin have not been reported . We first determined modified histone distributions in three protein-encoding genes , light ( lt ) , concertina ( cta ) , and Chitinase , which are located in Chromosome 2L heterochromatin ( 2Lh ) . We assayed chromatin of postblastoderm embryos ( 4–14 h ) using chromatin immunoprecipitation ( ChIP ) , a technique that reports the average chromatin profile for the population of nuclei assayed . Although these embryos contained multiple tissue types , previous studies showed that lt and cta are widely expressed throughout development , including during embryogenesis [31–33] , and ubiquitous expression has been documented in larval polytene and diploid nuclei by RNA in situ hybridization [34] . Because maternal loading of transcript obscured detection of zygotic transcription in earlier studies [32 , 33] , we verified transcription of lt and cta in 4–14 h embryos using a nuclear run-on assay ( Figure S1 ) . This assay also verified little or no expression of the Chitinase gene in 4–14 h embryos , a result consistent with previous conclusions that this gene functions mainly in larval molting [35] . The exons of the 2Lh genes are unique sequence , but the bulk of introns and flanking sequences consist of repetitive TE-like sequences ( Figure 1 ) . This structure required careful design of PCR primer pairs for the ChIP assay since detection of recovered fragments is hybridization-based . PCR primer pairs were designed so: ( i ) at least one primer in the pair was unique sequence , or ( ii ) both primers in the pair were different repetitive sequences but the juxtaposition of the two sequences was a unique occurrence in the genome . There were three instances in the case of the lt gene in which both primers in a pair were anchored in repetitive sequences . We verified specificity of these primer sets by showing that PCR fragments were amplified from DNA of normal embryos , but not from embryos deleted for the lt gene region ( Figure S2 ) . This scheme allowed specific detection of 2Lh gene sequences , despite the fact that most regions are repeated many times in the genome . Chromatin was immunoprecipitated with antibodies specific for H3K9me2 , H3K4me2 , or H3K9/14acet , then assayed for heterochromatic gene sequences and a control euchromatic gene sequence ( Pdi ) used for normalization . Enrichment was expressed relative to input DNA . These ChIP assays showed that H3K9me2 was enriched in the lt and cta genes and distributed across most of the transcribed regions except for the 5′-ends ( Figure 1A and 1B ) . Depletion of H3K9me2 in the transcription start-regions was complemented by enrichment of H3K4me2 and H3K9/14acet . A peak of H3K4me2 and H3K9/14acet in the start-region seems to be a universal feature of expressed genes in higher eukaryotes [28–30] , and in vitro studies indicate that these modifications and H3K9me may be inhibitory to each other in individual histone molecules [36 , 37] . However , depletion of H3K9me2 in the start-region may not necessarily be due to exclusion by the antagonistic histone modifications . This is indicated by the modified histone distribution in the Chitinase gene , which lacks H3K9me2 as well as H3K4me2 and H3K9/14acet at the start-region ( Figure 1C ) . This result is consistent with minimal Chitinase transcription in embryos ( Figure S1 ) . We also examined three euchromatic genes , Moca-cyp , CG1646 , and CG5514 , which previous in situ RNA hybridization and microarray studies showed are highly and ubiquitously expressed in 4–14 h embryos [38] . All three genes were enriched for H3K9/14acet at the 5′ end , but lacked significant levels of H3K9me2 ( Figure 1D and 1E; CG5514 data not shown ) . The general lack of H3K9me2 in euchromatic genes was also evident in a larger scale ChIP-chip analysis described below . Thus , heterochromatic and euchromatic genes carried similar H3K9/14acet signatures of expression , but differed in H3K9me2 association . Distribution of H3K9me2 throughout heterochromatic genes indicated integration into , rather than insulation from , the surrounding H3K9me2-enriched environment . To obtain global perspectives of H3K9me2 and H3K9/14acet profiles in heterochromatin , we used a ChIP-chip approach . The genome tiling array contained probes for heterochromatic regions of the genome , as assembled and annotated in the D . melanogaster genome project Release 5 . 1 ( R5 . 1 ) [39 , 40] . Specifically , we focused on the proximal ends of the assembled genome sequence corresponding to the euchromatin-heterochromatin ( eu-het ) transition zones and extending into distal heterochromatin of the long 2L , 2R and 3L chromosome arms ( Figure 2A ) . We also included a segment of 3R distal euchromatin and the entire sequenced region of the small Chromosome 4 for comparisons among chromosomal regions . To circumvent the cross-hybridization problem of repetitive sequences , genome sequence was masked for annotated TEs and remaining sequences were used to design overlapping 50-mer oligonucleotides probes with 40 bp resolution . This probe set was subject to stringent filtering to eliminate repetitive sequences , and remaining probes were used for the microarray ( Materials and Methods ) . Probe density necessarily varied across heterochromatin due to repeat filtering . Reproducible H3K9/14acet and H3K9me2 profiles were obtained in the ChIP-chip studies ( Figure 3 ) , permitting a comparison with the standard ChIP results ( Figure 1 ) . Specifically , the ChIP-chip assay showed strong H3K9/14acet enrichment in the 5′ regions of heterochromatic genes , but did not indicate as high an enrichment of H3K9me2 as anticipated based on standard ChIP . Nonetheless , the observed pattern was consistent with earlier results . As shown for 2Lh in Figure 3 , H3K9me2 was enriched in intergenic regions and throughout these heterochromatic genes but not near the transcription start-regions . This was also true for genes with little or no association with H3K9/14acet ( e . g . , Chitinase , CG17018 , and CG40006 genes in Figures 3 and 4A ) . To obtain an average profile of the heterochromatic genes represented on our tiling array , we defined a set of genes as heterochromatic if they were located proximal to the CG3635 , nrm , or CG11665 genes at the base of 2L , 3L , or 2R respectively . Of the 111 annotated genes that met this criterion , 63 had cDNA evidence to support the annotation of the 5′ end . As shown in Figure 4B , the average modified histone profiles of these 63 genes differed from those for two other categories of genes included in our microarray: protein-encoding genes located on Chromosome 4 ( n = 72 ) and euchromatic genes located on distal 3R ( n = 105 ) . All three classes showed a peak of H3K9/K14acet at the 5′ start region with differences in peak size likely reflecting the different proportion of genes expressed in embryos in each gene set . Based on recovery of embryo ESTs reported in FlyBase [41] , 81% of the genes in the heterochromatic gene set , 71% of the Chromosome 4 genes , and 55% of the 3R-euchromatic genes are transcribed in embryos . Notably , heterochromatic genes showed a significantly higher average level of H3K9me2 compared to Chromosome 4 and 3R euchromatic genes ( Figure 4B ) . Overall , Chromosome 4 genic and intergenic regions exhibited slightly elevated average levels of H3K9me2 relative to those typically observed within euchromatin ( Figure S3 ) . To determine if the SU ( VAR ) 3–9 HMTase was responsible for the H3K9me2 enrichment in and around heterochromatic genes , we examined the effects of a null mutation Su ( var ) 3–906 [24] . The ChIP assay revealed small but reproducible differences between the wild type and homozygous Su ( var ) 3–906 in the relative levels of H3K9me2 across the lt gene ( Figure 5A ) . To further investigate this effect , we performed two separate ChIP-chip hybridization comparisons of wild type and Su ( var ) 3–906 embryos . We found that the landscape of H3K9/K14acet and H3K9me2 across 2Lh ( Figure 3 ) and other heterochromatic regions ( data not shown ) appeared overall similar in the two genotypes . However , in the mutant embryos , the average profile of heterochromatic genes showed a slightly elevated level of H3K9me2 just upstream of the start site relative to the level observed within the transcribed region ( Figure 5B ) . The average profile across euchromatic genes was not detectably altered in the mutant . A striking difference between the two genotypes was evident from examining H3K9me2 distribution on Chromosome 4 . Relative to wild type embryos , Su ( var ) 3–906 embryos showed a higher average level of H3K9me2 within Chromosome 4 genes and more pronounced regional increases along the length of the chromosome ( Figure S3 ) . This unexpected observation of localized increases , rather than reductions of H3K9me2 associated with Su ( var ) 3–906 indicates that in embryos , SU ( VAR ) 3–9 is not responsible for the majority of H3K9me2 in the regions we assayed here , but it apparently influences one or more HMTases that play more prominent roles . The H3K9me2 profile characteristic of the long autosomal arms differed markedly from that observed for Chromosome 4 . In euchromatic regions , the profile was typically flat ( Figures S3 and 6 ) . However , within the eu-het transition zones on 2L , 2R , and 3L , a sharp transition was observed between low levels of H3K9me2 in distal regions to enriched domains more proximally ( Figure 6 ) . Sequence inspection did not reveal common underlying sequences at the transition point , or conspicuous features such as large inverted repeats or tRNA genes that characterize the boundary elements of yeast silent chromatin domains [12 , 42] . However , the transitions coincided with increased density of retrotransposons including LTR-type and LINE-like TEs ( Figure 6D–6F ) . Within the limited coverage of ∼15 Mb of euchromatin on the tiling array , there was no significant association of H3K9me2 noted in sequences adjacent to the many solitary TEs in euchromatin . However , we observed several prominent peaks that corresponded to the coding sequences of unique genes ( Figures 6A–6C and S4 ) . A few less prominent sites resided adjacent to regions with a high TE density ( Figure 6A–6C , asterisks ) and while these were reproducible , they showed overall lower and more local H3K9me2 enrichment compared to the broad peaks evident in heterochromatin . Included on the tiling array were probes for sequences contained in 106 “canonical” TEs . These sequences represented full-length consensus sequences of each TE type [43] and were included to ask whether certain classes of TEs might exhibit overall higher associations with H3K9me2 . The interpretation of the ChIP-chip data for these sequences may be complicated by differences in TE copy numbers in the genome and the proportion of complete vs . incomplete versions which vary with different elements . Nonetheless , each TE generated a distinctive H3K9me2 profile that was reproducible in two different genotypes ( Figure S5; Table S1 ) . When TEs were ranked by the average H3K9me2-enrichment across their lengths , it was clear that the higher-ranked TEs belonged to retrotransposon class . Forty-eight TEs that showed the highest average H3K9me2-enrichment were all retrotransposons . DNA transposons , such as the 1360 element , show moderate to low H3K9me2-enrichment in this assay . Another notable observation was that TEs with especially high copy number in euchromatic regions [43] , such as the roo element , tended to show lower H3K9me2-enrichment . We designed the genomic tiling array to investigate whether a chromosome rearrangement which created a new eu-het junction might be associated with detectable changes in H3K9me2 distribution . We compared embryos homozygous for T ( 2;3 ) ltx13 , a reciprocal translocation which displaces most of 2Lh distally to 3R [19] ( Figure 2B ) , to two wild-type strains: Canton-S , which was the parent strain for T ( 2;3 ) ltx13 and y; cn bw sp , the strain used for the genome sequencing and annotation projects . Both halves of the translocation are of interest to assay for H3K9me2 profiles . Flies heterozygous for T ( 2;3 ) ltx13 and chromosomes that carry null alleles of the 2Lh genes have pronounced variegated phenotypes for lt and other 2Lh genes . However , flies homozygous for T ( 2;3 ) ltx13 are viable and show very weak variegation of the lt gene in adult tissues [19] . As shown in Figure 3 , the ChIP-chip assays did not reveal differences in the H3K9me2 profile of the distally displaced 2Lh genes in embryos homozygous for T ( 2;3 ) ltx13 compared to wild-type embryos . However , significant changes were observed in the 3R euchromatic segment that was placed adjacent to pericentric heterochromatin ( Figure 7 ) . Euchromatic sequences adjacent to the breakpoint showed an abrupt enrichment of H3K9me2 , with continued overall enrichment extending at least 200 kb ( from coordinate 22 . 4–22 . 6 Mb ) , reflecting spreading of the H3K9me2 domain across the breakpoint . Also striking were many discrete H3K9me2 peaks scattered over a ∼3 Mb region . The most prominent peaks were found over coding sequences of genes ( Figure 7D–7G ) . Some , but not all , of the peaks may have resulted from enhancement of H3K9me2 affinity sites present on normal chromosomes . For example , a roo element is present at coordinate 22 . 8–22 . 9 Mb in the y; cn bw sp strain and adjacent single copy sequences on either side have a slightly higher than average H3K9me2 association ( Figure 7A ) . It is possible that the prominent peak in T ( 2;3 ) ltx13 is formed around a roo element in this same location ( Figure 7C and 7D ) . Perhaps more relevant is that this roo element is nested in a cluster of seven duplicated genes ( Figure 7D ) . Another example of possible enhancement of a site showing H3K9me2 affinity is located at coordinates 24 . 3–24 . 4 Mb which corresponds to the coding sequence of the Papilin gene ( Figure 7F ) . Other peaks in T ( 2;3 ) ltx13 had no apparent presages in the normal chromosomes but notably , the site at coordinate 24 . 7 Mb encompassed cluster of five duplicated genes ( Figure 7G ) . Additional peaks near the telomere were also observed in T ( 2;3 ) ltx13 but not normal chromosome ( data not shown ) . These results demonstrate that a chromosome rearrangement can exert an epigenetic effect as far as 5 Mb distance from the breakpoint . In addition , the H3K9me2 profile in T ( 2;3 ) ltx13 was noticeably “bumpy” in the displaced 3R compared to the flat profile in the corresponding regions in normal chromosomes ( Figure 7 ) , suggesting a pervasive H3K9me2 association on this short chromosome arm .
Drosophila has traditionally provided a powerful , genetically tractable system for studies of heterochromatin . Here we combined genetic tools , improved heterochromatin sequence coverage , and ChIP approaches to obtain a high-resolution and large-scale molecular view of modified histone distribution in Drosophila heterochromatin . We used embryos , rather than cultured cells , for chromatin profiling to examine the effects of mutations or chromosome rearrangements whose consequences in the whole organism have been well documented . Analysis of the distribution of modified histones showed that 5′ regions of expressed heterochromatic genes were enriched for H3K4me2 and H3K9/14acet , consistent with these modifications serving as markers of active genes in both heterochromatin and euchromatin [28–30] . The complementary distributions of H3K4me2 , H3K9/14acet , and H3K9me2 support previous findings that these are mutually exclusive H3 modifications [36 , 37] . Most significantly , we found that H3K9me2 enrichment throughout transcribed regions is a general property of heterochromatic genes . Hence , heterochromatic genes constitute a category of genes that are an integral part of the H3K9me-enriched domain and are not insulated from it . This contrasts with the situation for Arabidopsis genes that are located in a heterochromatic “knob” , since the knob region is highly enriched for H3K9me and DNA-methylation as a whole , but expressed genes are largely free of these molecular marks [44] . Our results showing H3K9me2 association in and around heterochromatic genes are consistent with those reported for HP1 by de Wit et al . [45 , 46] . These investigators used the DamID chromatin profiling technique and showed HP1 enrichment across the lt , cta , and rl heterochromatic genes in cultured cells . A direct association of HP1 and H3K9me with expressed heterochromatic gene sequences is consistent with models that evoke a facilitating role of heterochromatin-enriched proteins for heterochromatic gene expression [18] . Multifunctional roles for HP1 , including in gene activation , are now widely acknowledged [47] . Our data show that H3K9me2 has , at minimum , a permissive role for heterochromatic genes . Other investigators reported recruitment of H3K9me3 , which is generally associated with “silent heterochromatin” in mammals , as well as H3K9me2 , to expressed euchromatic genes in mammalian cells [48 , 49] . The number of such examples is increasing with more comprehensive surveys of cell types [50 , 51] . Vakoc et al . [48 , 49] described complementary profiles of H3K9me3 and H3K9acet in active euchromatic genes which resembled those that we describe here , although H3K9me3 levels were far lower relative to levels of H3K9acet in these mammalian genes . H3K9me3 association with these genes is also highly dynamic and appears to correlate well with induced transcriptional activity and/or cell differentiation state , rather than repetitive DNA sequence content [48–51] . Our observations suggest pervasive H3K9me2 associated with active and inactive heterochromatic genes in embryos , except at the start regions , and the relatively high levels of H3K9me2 may be maintained by the high density of nearby repetitive DNAs that characterize heterochromatic but not euchromatic genes in flies . Notably , within the limited euchromatic regions surveyed by ChIP-chip , we also identified several euchromatic genes with enriched H3K9me2 profiles . These genes lacked prominent H3K9/14acet peaks at their 5′ ends and are poorly represented in available embryo EST and cDNA collections , indicating that they are not abundantly expressed in embryos . In one case , the H3K9me-enriched regions corresponded to the exon cassettes of Dscam , a gene which is known for its extraordinarily complex alternative splicing [52] . It is conceivable that H3K9me2 is involved in the regulation of Dscam alternative splicing , since there are other genes in which chromatin has been implicated in alternative splicing control [53] . Another possibility is that H3K9me2 association may simply be due to repetitiveness of similar exonic sequences within the gene . More extensive surveys are required to determine if H3K9me may be occurring more generally in Drosophila euchromatic genes , perhaps at levels below the sensitivity of our ChIP-chip assay , in order to understand the possible significance , if any , of H3K9me in euchromatin . An important question to address is which trans-acting factor ( s ) is responsible for enrichment of H3K9me2 in heterochromatic genes and the eu-het transition zones . SU ( VAR ) 3–9 was a reasonable candidate since it is the major HMTase responsible for H3K9me2 and HP1 recruitment in constitutive heterochromatin [24] . This role was based on the observation that the Su ( var ) 3–906 null mutation results in strong reduction of H3K9me2 and HP1 immunostaining in the chromocenter , but not the euchromatin or Chromosome 4 , in larval polytene nuclei [24 , 26] . Furthermore , Su ( var ) 3–906 mutants show reduced levels of H3K9me2 relative to wild type levels in assays of bulk chromatin of 0–12 h embryos [24] or larvae [26] , and reduced recovery of certain highly repetitive sequences in ChIP assays of larval [54] or adult [55] chromatin . Two features of the H3K9me2 profiles of Su ( var ) 3–906 embryos reported here are remarkable in light of previous observations . First , the H3K9me2 landscapes in eu-het transition zones and in heterochromatic genes was overall similar in mutant and wild type embryos . Second , differences that were observed were relatively small local increases , rather than decreases of H3K9me2 in the mutant . These changes were evident as altered shapes of the H3K9me2 profiles across heterochromatic genes and Chromosome 4 . This effect of Su ( var ) 3–906 may be limited to embryos and/or the specific heterochromatic regions that we assayed . Our assays probed the single copy sequences interspersed within much larger expanses of repetitive DNA . The ChIP-chip probes were distributed across 1 . 35 Mb of Chromosome 4 and across 15 Mb of DNA located at the base of 2L , 2R , and 3L , including the eu-het transition zones and extending into the sequenced heterochromatin of these chromosome arms [17] . Thus , while we sampled across a significant proportion of the sequenced portions of Chromosome 4 and autosomal heterochromatin , the single copy probes themselves totaled 0 . 8 Mb , representing roughly 1%–2% of the estimated 59 Mb of heterochromatin present in diploid cells . Thus , the subtle effect of Su ( var ) 3–906 detected in our ChIP assays would be easily missed in assays of bulk chromatin from embryos and masked by the pronounced effect of the mutation on H3K9me2 levels in tandemly repeated DNAs [24 , 56] . At minimum , our findings indicate that SU ( VAR ) 3–9 is not responsible for the majority of H3K9me2 in heterochromatic genes and the eu-het transition zones in embryos . They also suggest that SU ( VAR ) 3–9 influences one or more HMTases that act on heterochromatic genes and Chromosome 4 in embryos . The influence of SU ( VAR ) 3–9 may be to limit the activities of other HMTases , perhaps through direct binding or by competing for shared factors or histone substrates . We do not know if the altered H3K9me2 profiles associated with Su ( var ) 3–906 have functional consequences . We note that strong effects of Su ( var ) 3–9 mutations on heterochromatic gene expression have not yet been reported and are not expected since Su ( var ) 3–906 flies are homozygous viable and fertile while many heterochromatic genes , including lt and cta , are essential for viability and/or fertility . Recent studies have identified DmSETDB1 as an HMTase which is primarily responsible for Chromosome 4 H3K9me2 in larvae and adults , although it appears to be acting independently of SU ( VAR ) 3–9 at these stages [57] . dG9a has been characterized as a H3K9 HMTase with a partially overlapping function with SU ( VAR ) 3–9 at multiple developmental stages [56] . Hence , these and perhaps other , as yet unidentified HMTase activities , may contribute more prominently than SU ( VAR ) 3–9 to H3K9me2 dynamics in the heterochromatic regions that we assayed in embryos . The characterization of the molecular landscape of modified histones along the long chromosome arms provided new insights into the contributions of TEs in influencing H3K9me distribution . We were surprised to see the abrupt increase of H3K9me2 enrichment in the euchromatin-heterochromatin transition zones . The striking coincidence of the transitions with marked increase in the density of retrotransposons in all three of the chromosome arms ( 2L , 2R and 3L ) examined in this study suggests a mechanistic link . This is consistent with the finding that the TEs with highest average H3K9me2-enrichment in the ChIP-chip assay belong to the retrotransposon class ( Figure S6 ) . We also note an apparent overrepresentation of retrotransposon sequences in the list of cloned repeat-derived small RNAs [13] which have been implicated in heterochromatin formation [6] . Taken together , the results implicate clusters of retrotransposons as major determinants in the demarcation of H3K9me2-containing domains , and what may correspond to the distal edges of pericentric constitutive heterochromatin in Drosophila embryos . An emerging concept from this and other studies is that it is not repetitive sequences per se that specifies heterochromatin but the physical proximity of multiple repetitive sequences of a certain type [45 , 58 , 59] . This contrasts with the idea of “cryptic heterochromatin , ” suggested by Lippman et al . [44] , in which isolated copies of TEs in euchromatin are viewed as heterochromatin because they bear the molecular markers generally associated with heterochromatin , including H3K9me2 . In this study , numerous isolated euchromatic copies of retrotransposons did not bear marks of H3K9me2 and the TEs with high copy-numbers in euchromatin generally show low average H3K9me2-enrichment . If , for example , every copy of roo element in euchromatin was associated with H3K9me2 , we would expect highly elevated ChIP-to-input signal ratios on the roo element microarray probes . Instead , the results of our studies are consistent with the notion that majority of the isolated copies of TEs in euchromatin are nucleating little , if any , H3K9me2 association , at least in embryonic chromatin . The H3K9me2 landscape in the T ( 2;3 ) ltx13 chromosome revealed complex redistribution of H3K9me2 featuring several peaks at discrete sites as far as a few megabases from the breakpoint . This provides molecular evidence for the impressive distances over which epigenetic effects of chromosome rearrangements can act in multicellular organisms . It is of interest to know why certain sites might attract the complexes that permit H3K9me2-enrichment . Inspection of the newly induced euchromatic sites in T ( 2;3 ) ltx13 and euchromatic sites in normal chromosomes indicate that the two types of H3K9me2 peaks might be formed via similar mechanisms . In both cases , peaks are associated with gene exons or with clusters of duplicated genes , although sequence similarity among these duplicated segments are not necessarily extensive . Discovering additional examples should be informative to better understand the underlying basis for H3K9me enrichment . Comparisons of profiles in different chromosome sources should reveal if euchromatic sites are reproducible in different stages or tissues . Whatever determines the affinity for H3K9me2 at euchromatic sites , the analysis of the T ( 2;3 ) ltx13 rearrangement has established that discrete sites in euchromatin can nucleate high-levels of H3K9me2 if situated in the appropriate physical context . In summary , these studies of Drosophila heterochromatic genes and the euchromatin-heterochromatin transition zones challenge several traditional notions of the epigenetic signatures thought to distinguish heterochromatin from other chromatin domains . Instead , the results contribute to a growing body of evidence that suggests versatility in the activities of different types of heterochromatin-enriched repetitive DNA sequences and modified histones and emphasize the importance of chromosomal context . Knowledge of the modified histone distribution in and around heterochromatin and redistribution by chromosome rearrangement represents an important step toward understanding the function of epigenetic modifications in complex eukaryotic genomes .
D . melanogaster strains used in this study included y; cn bw sp ( the strain used for whole genome sequencing ) , T ( 2;3 ) ltx13 , and a wild-type Canton-S line which was the T ( 2;3 ) ltx13 parental strain [19] . We also used the wm4; Su ( var ) 3–906 strain obtained from G . Reuter and verified that it carried the Su ( var ) 3–906 insertion by PCR [26] . We used the following antibodies which have been previously used to detect histone modifications in Drosophila: Upstate ( http://www . upstate . com/ ) antibodies #07–441 ( anti-H3K9-di-methylation ) [24 , 26] , #07–030 ( anti-H3K4-di-methylation ) [30] , #06–599 ( anti-H3K9/K14-acetylation ) [30] , and Abcam ( http://www . abcam . com/ ) #ab1791 ( anti-H3 C-terminus ) [60] . Quality controls verified that that the Upstate anti-H3K9me2 antibody acted as previously described in immunostaining assays of polytene chromosomes [24] and also in ChIP assays . For ChIP , we obtained enriched recovery of the 359bp-satellite repeat ( 4-fold and 8 . 9-fold in two separate trials ) from immunoprecipitated relative to input embryonic chromatin using the PCR primers described by Rudolph et al . [55] Detailed ChIP protocol is described in the Protocol S1 . Briefly , 4–14 h embryos were subjected to cross-linking with 1% formaldehyde and chromatin was sheared with sonication . After immunoprecipitation with appropriate antibodies , DNA was purified from the chromatin and used for quantitative PCR or microarray hybridization . ChIP enrichment factors shown in Figure 1 were determined by comparing intensities of radiolabeled PCR fragments . Quantitative duplex PCRs were carried out as described by Noma et al . [12] and normalized using an arbitrarily chosen euchromatic control gene Pdi . The Pdi PCR fragment ( bottom band ) was 368 bp and the heterochromatin PCR fragments ( top band ) were 500–600 bp . Two or more independent immunoprecipitations were performed for each antibody and the error bars represent standard deviations . For the ChIP-chip assays shown in Figures 3–7 and S3–S6 , immunoprecipitated DNA was amplified with two rounds of ligation-mediated PCR ( 23 and 6 cycles , respectively ) essentially as described by Li et al . [61] , or by GenomePlex Whole Genome Amplification Kit ( Sigma , http://www . sigmaaldrich . com/ ) , and used for microarray hybridization . The tiling array contained sequences located in euchromatin-heterochromatin transition zones on chromosome arms 2L , 2R , and 3L and extending into the heterochromatin as assembled and annotated in R5 . 1 of the D . melanogaster genome [39] . It also included the sequenced portion of Chromosome 4 ( 1 . 35 Mb ) , a portion of distal 3R euchromatin ( 5 . 96 Mb , coordinates 21 , 932 , 592 to 27 , 897 , 863 ) , and canonical sequences from a collection of 106 TEs . The TE sequences were from the archived set DMEL . TRANSPOSABLE . ELEMENTS . 9 . 4 . 1 . ( downloaded from http://www . fruitfly . org/p_disrupt/datasets/ASHBURNER/D_mel_transposon_sequence_set . fasta [43 , 62] and are described in Table S1 ) . The contigs of interest were extracted from TE-masked R5 . 1 Drosophila genomic sequence ( downloaded from http://www . fruitfly . org/sequence/release5genomic . shtml ) . The probe set and microarray were custom-manufactured at NimbleGen ( http://www . nimblegen . com/ ) . Overlapping 50mer probes were designed with 40 bp resolution . In selecting the final set of probes for the array , only those probes with the 15mer-test score of 2 . 55556 or less were retained , meaning that 15mers contained in each 50mer probe appeared on average no more than 2 . 55556 times in the unmasked genome sequence . This is a highly stringent repeat-filtering compared to the customary cutoff score of up to 100 . When an even more stringent cutoff ( 15mer-test score of 1 . 25 or less ) was applied in the post-hybridization analyses , the overall profiles of H3K9me2 were not affected ( Figure S6 ) . The canonical TE collection was exempt from the repeat filtering and probes were designed to cover each TE . Sample labeling and hybridizations were carried out at NimbleGen using standard procedures . The ChIP/reference raw signal ratios were normalized by the medians of the entire array . Enrichment is expressed as fold enrichment relative to one of two reference controls , either total genomic DNA or anti-H3 C-terminus ChIP , which produced similar results ( Figure S6 ) .
The data discussed in this manuscript have been deposited in the National Center for Biotechnology Information's ( NCBI ) Gene Expression Omnibus ( GEO , http://www . ncbi . nlm . nih . gov/geo/ ) and are accessible through GEO Series accession number GSE7839 .
|
The chromosomal domain “heterochromatin” was first defined at the cytological level by its deeply staining appearance compared to more lightly stained domains called “euchromatin . ” Abnormal juxtaposition of these two domains by chromosome rearrangements results in silencing of the nearby euchromatic genes . This effect is mediated by heterochromatin-enriched chromosomal proteins and led to the prevalent view of heterochromatin as incompatible with gene expression . Paradoxically , some expressed genes reside within heterochromatin . In this study , we examined how heterochromatic genes fit into a genomic context known for silencing effects . We found that Drosophila heterochromatic genes are integrated into the domain enriched in the modified histone H3K9me2 , suggesting that the effect of this protein on gene expression is context-dependent . We also investigated the molecular nature of euchromatin-heterochromatin transition zones in the normal and rearranged chromosomes . The results provide insights into the functions of repetitive DNAs and H3K9me2 in heterochromatin and document the long distance over which a heterochromatic breakpoint can affect the molecular landscape of a chromosomal region . These findings have implications for understanding the consequences of chromosome abnormalities in organisms , including humans .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"drosophila",
"genetics",
"and",
"genomics"
] |
2008
|
Molecular Landscape of Modified Histones in Drosophila Heterochromatic Genes and Euchromatin-Heterochromatin Transition Zones
|
Neuronal information processing is regulated by fast and localized fluctuations of brain states . Brain states reliably switch between distinct spatiotemporal signatures at a network scale even though they are composed of heterogeneous and variable rhythms at a cellular scale . We investigated the mechanisms of this network control in a conductance-based population model that reliably switches between active and oscillatory mean-fields . Robust control of the mean-field properties relies critically on a switchable negative intrinsic conductance at the cellular level . This conductance endows circuits with a shared cellular positive feedback that can switch population rhythms on and off at a cellular resolution . The switch is largely independent from other intrinsic neuronal properties , network size and synaptic connectivity . It is therefore compatible with the temporal variability and spatial heterogeneity induced by slower regulatory functions such as neuromodulation , synaptic plasticity and homeostasis . Strikingly , the required cellular mechanism is available in all cell types that possess T-type calcium channels but unavailable in computational models that neglect the slow kinetics of their activation .
Neuronal processing is constantly shaped by fluctuations in population rhythmic activities , each defining distinctive brain states [1–4] . Neuromodulators organize the switch between different brain states [5 , 6] , changing the way networks process neural signals [7 , 8] . Precise temporal and spatial control of brain states is required for changes associated with movement , attention , perception , motivation , or expectation [9–18] . Fast acting neurotransmitter pathways allow for the rapid kinetics required for fast network and signal processing states changes [17] . Rapid control of network states has been reported to affect spatial attention in cortical circuits [6 , 7] , attention and arousal in the thalamus [7 , 18] , and movement initiation in the subthalamic nucleus [14] . The most studied example is probably the thalamocortical circuitry . The thalamus acts as a plastic relay between sensory systems , different subcortical areas and the cerebral cortex , by gating and modulating neuronal signal flow under the modulatory effect of cortical feedback [19–22] . Experimentally , brain states are identified via specific spatiotemporal signatures of the mean-field electrical activity of large neuronal populations . Shifts in the rhythmic activity occur during transitions to slow-wave sleep and sleep spindles [1 , 7 , 8 , 23–27] . These shifts correlate with strong changes in the processing of afferent neuronal signals [8 , 17 , 28] . An extreme situation is when highly synchronized sleep oscillations develop into absence epilepsy , a behavioral state that can be viewed as a brain disconnection from the external world [29–31] . In the waking state as well , transient network state switches are observed and correlate with modulations of sensory-motor signals processing [16] . What are the mechanisms that enable fast and robust mean-field switches in heterogeneous neuronal populations that exhibit rhythms over a broad range of temporal and spatial scales , from single cells to networks ? At a cellular level , the rhythms are determined by specific balances of specific ionic currents . Specific synaptic connections determine specific circuit topologies that define new and different rhythms at a circuit scale . At a network level , the circuit topologies interconnect large and heterogeneous neuronal populations . Collectively , the populations shape a mean-field activity that defines yet another rhythm for the brain state . At every scale , the rhythms are continuously changing under the action of neuromodulators that modulate cellular and synaptic conductances over time . Neuromodulatory systems act at a cellular scale but their broad projections can simultaneously affect large populations [6 , 7 , 17 , 27] . Global neuromodulators control the switch of brain states [16 , 32–34] . The question we wish to investigate in the paper is how this global control can cope with—and in fact benefit from—the heterogeneity and variability of rhythms at a cellular scale . A similar question has received considerable attention over the last two decades in the study of neuromodulation of small rhythmic circuits controlling the pyloric and gastric mill rhythms of the crab [35–38] . This work has elucidated to a great extent how heterogeneity and variability at the cellular scale is not only compatible with homogeneity and stability at the circuit scale but in fact an essential source of robustness and tunability in circuit rhythms . The present paper is inspired by this line of work: we developed a simple conductance-based computational model to investigate how heterogeneity and variability at the cellular and circuit scales contribute to the robust control and tunability of brain states . Previous computational models of brain states have focused on the role of connectivity changes in network rhythm modulation [39–41] . To account for fast fluctuations , our model instead studies network switches that do not require changes in network connectivity . We propose that the mean-field switch results from a cellular switch that is shared by a sufficient fraction of the population . This mechanism is largely independent of the network topology and the network connectivity is always assumed to be weak . This is what allows for rhythmic heterogeneity within the population . The cellular switches only control a discrete transition between two distinct modes of excitability , classically referred to as tonic firing and bursting [42] . The weak connectivity makes this discrete transition compatible with a continuum tuning of each discrete state . The homogeneous control of brain states at a network level only relies on the shared cellular switch . It is compatible with heterogeneous and variable rhythms at a cellular and circuit scales . Our paper aims at showing the computational and physiological relevance of this novel mechanism both for network studies and cellular physiology . Regarding the network computations , we show that heterogeneity and variability at the cellular and circuit scale promote robustness and tunability at the network scale . The cellular switch decouples the control of network states , which is fast and global , from the tuning of the spatiotemporal rhythm , which is ensured by modulation of intrinsic and synaptic properties at slower temporal and finer spatial scales . The decoupling of tasks between switching and tuning allows for a fine regulation of both the oscillatory and active states of the network . This tuning of states has essential functional relevance , such as modulating the transmission properties of the network in its active state . At a physiological level , our results stress the role at a network level of specific ionic mechanisms that have long been studied in single cell neurophysiology but are often neglected in network studies for the sake of numerical or mathematical tractability . The cellular switch of our model entirely relies on a tunable slow negative conductance . The switching role of this slow conductance has been studied in a series of recent papers by the authors . It regulates the cellular modulation of excitability types [43–46] and in particular the transition between tonic firing and bursting [47 , 48] . It is critical for the robustness and tunability of cellular bursting [42] . And it is critical to the robustness of rhythmic circuits such as half-center oscillators [49] . In the continuation of this research , the present paper shows how slow negative cellular conductances contribute to network neuromodulation , highlighting the physiological importance of a specific cellular mechanism at a network level .
Our computational model reproduces a generic example of network switch between an active and an oscillatory state in a population of neurons . At the cellular level , we used a conductance-based model that includes the typical fast and slow ionic currents of bursting cells . Each cell can be robustly modulated by a hyperpolarizing input between two distinct modes of excitability: a fast depolarized mode , prone to spiking and tonic firing , and a slow hyperpolarized mode , prone to bursting . At the network level , we included AMPA , GABAA and GABAB connections to model the asymmetric coupling between a subpopulation of excitatory ( E ) cells and a subpopulation of inhibitory ( I ) cells . This topology is typical of brain areas involved in state regulation , such as e . g . the thalamus [7] , the cortex [2] , and the subthalamic nucleus/globus pallidus [39] . Our model neglects intra-population connectivity , which maximizes the heterogeneity of cellular rhythms within each subpopulation . In contrast , it assumes all-to-all connectivity between the two populations , but with weak and randomly distributed synaptic weights ( see methods for details ) . We explored rhythmic properties of the network that allow for a broad heterogeneity of both intrinsic and synaptic maximal conductance parameters . This parametric heterogeneity generates a broad range of rhythms at cellular and circuit levels . The circuit and network switches are illustrated in Fig 1 . They are controlled by transient hyperpolarizations of the inhibitory neurons , which for instance mimic activation of GABA_B receptors . At the cellular level , the hyperpolarization induces a switch from the fast excitability mode to the slow excitability mode . The cellular switch in turn induces a switch at the circuit level because the slow mode of excitability induces rebound mechanisms between the excitatory and inhibitory cell . Fig 1A and Fig 1B illustrate the robustness and modulation capabilities of the rhythms in isolated E-I circuits of two cells . Fig 1A shows that a transient hyperpolarization can reliably induce a switch from asynchronous spiking to synchronous bursting , and that this switch is robust to changes in neuron intrinsic and synaptic properties induced by a persistent , global neuromodulation ( "NMD" ) and synaptic plasticity ( "Syn . Plast . " ) . Thanks to the robustness of the cellular switch to variability , the circuit rhythm is robustly maintained over broad intrinsic and synaptic parameter ranges . As a result , a persistent neuromodulator affecting both intrinsic and synaptic parameters can modulate the cellular and circuit rhythms between successive occurrences of the transient hyperpolarization ( Fig 1A and 1B ) . At the network level , the mean field of the population defines a network switch between an oscillatory state , corresponding to the slow mode of cellular excitability , and an active state , corresponding to the fast mode of cellular excitability . The network switch is robust to temporal variability and spatial heterogeneity of the population because the cellular switch exists over a broad range of intrinsic and synaptic parameters . This robustness makes the external control of the network largely independent of the network size and connectivity . Fig 1C and 1D illustrate that property in a heterogeneous network of 200 cells . The local field potential ( LFP ) activity illustrates that the transient hyperpolarization can turn on and off the mean-field rhythmic activity ( defined by a marked high power LFP frequency band ) of the entire controlled population . Fig 1D further shows that this population control is compatible with heterogeneous rhythms at a cellular and circuit resolution . In other words , reliable control of the network state is compatible with rhythmic heterogeneity and variability at a cellular resolution . The mean-field network rhythm does not result from the synchronization of the cellular or circuit rhythms . Instead , the rhythmic network state arises from a shared cellular switch that is robust to the variability of tunes at a cellular resolution . Our computational model exhibits a robust mean-field switch at the network level in spite of heterogeneity and variability at the cellular level . Such a property is not granted in a computational model that depends on thousands of uncertain and variable parameters . In our computational model , it critically relies on a switchable slow negative conductance at a cellular level . The slow negative conductance of a neuron determines its slow excitability in the same way as the fast negative conductance determines its fast excitability . In the same way as sodium channel activation enables the fast switch from rest to spike , a slow negative conductance enables the slow switch from rest to burst . Ion channels that can contribute to the slow negative conductance of a neuron are called slow regenerative [45 , 46 , 48] . A channel is slow regenerative if it activates an inward current or inactivates an outward current in a time scale that is slow relative to the fast time scale of sodium activation . In our model , only the T-type calcium channels are slow regenerative . Their activation is slow relative to sodium channel activation [50] . Moreover , because of their inactivating , low threshold nature , T-type calcium channels equip the neuron with a slow negative conductance that is switchable by an external current: it is turned on by hyperpolarization and turned off by depolarization . This switching mechanism is distinct from the classical rebound mechanism ( called post-inhibitory rebound , or PIR ) associated to T-type calcium channels and other transient inward currents such as hyperpolarization-activation cation currents [7 , 49 , 51 , 52] . It endows the cell with slow excitability . The experimental manifestation of this slow excitability notably includes rebound bursting ( RB ) ( Fig 2A , left ) and hyperpolarization-induced bursting ( HIB ) ( S1 Fig ) . Such behaviors have been widely observed experimentally [7 , 53–58] . We have studied in detail in [42] why a slow negative conductance is critical to a cellular activity that allows for a robust and controlled switch between fast and slow tunable rhythms , and how this mechanism relates to bursting models of the literature , such as square wave bursting or parabolic bursting . A similar observation applies to the two-cell excitatory-inhibitory circuit of Fig 1A . The circuit rhythm results from a well-known rebound mechanism [7 , 51] , but it does do so only when the slow negative conductance is turned on . In order to assess the specific role of the slow negative conductance , we proceeded with the same protocol as in [42]: we compared a nominal model in which the activation kinetics of T-type calcium channels is physiologically slow , ( about ten time slower than the activation kinetics of sodium channels ) to a perturbed model in which the activation kinetics of T-type calcium channels is instantaneous , that is , undifferentiated from the activation kinetics of sodium channels . Instantaneous activation of calcium channels is a frequent modeling assumption ( see Discussion ) . In both the nominal and perturbed models , T-type calcium channels provide an inactivating inward current necessary for the rebound mechanism ( Fig 2A ) . However , the slow excitability mode is lost in the perturbed model because both sodium and calcium channels only contribute to the fast negative conductance . This change has a clear signature in a voltage-clamp experiment ( Fig 2B ) . In the nominal model , a voltage step from hyperpolarized potential ( -90mV ) deinactivates T-type calcium channels , which results in two temporally distinct phases of negative conductance: a fast one ( depicted in yellow in Fig 2B , left ) and a slow one ( depicted in green in Fig 2B , left ) . The specific signature of the slow negative conductance is not present in the perturbed model . Instead , the two negative conductances add up in the fast time scale . The chosen perturbation has the advantage that it does not affect the model properties at equilibrium: the nominal and perturbed models have the same I/V curve and the same balance of currents at steady-state . Fig 2C illustrates that the slow kinetics of calcium channel activation is essential to make the switch of a 2-cell E-I circuit robust to parameter variability . A thousand 2-cell networks were simulated by randomly generating 1000 different parameter sets for maximal intrinsic and synaptic conductances ( see Methods ) . Fig 2C , left shows that the circuit switch occurs in the vast majority of parameter configurations when T-type calcium channel activation is physiological ( >90% for Iapp = -2 . 6μA/cm2 ) . This is the robustness that allows for variability , modulation ( Fig 1A and 1B ) , and heterogeneity ( Fig 1C and 1D ) at the cellular resolution . In contrast , robustness and tuning properties of the cellular and circuit rhythms are totally lost when the calcium activation is fast ( Fig 2C , right ) . The rhythm of the E-I circuit is extremely fragile to the loss of slow regenerative ion channels . It requires a very precise tuning of both the ionic ( intrinsic ) and synaptic ( extrinsic ) parameters . No circuit rhythmic activity could be found out of the 1000 random parameter sets simulated in this configuration . In the absence of slow regenerativity , the switch of the circuit cannot be decoupled from the rhythm of the circuit: both exist only for very specific parameter sets . Fig 3A and 3B illustrate the continuous tunability of rhythms at the circuit level in the two discrete excitability modes . Fig 3A shows that the frequency of tonic firing and the intraburst frequency of the bursting mode can be modulated over a broad range ( several fold ) . All cells verify the physiological property that the intraburst frequency is significantly higher than the tonic firing frequency , a key feature of bursting signaling and a physiological signature of hyperpolarization-induced bursting [7 , 53–55] . Fig 3B shows that two properties of the slow rhythm ( bursting frequency , top , and burst duty cycle , bottom ) can also be modulated over a broad range ( more than 5 fold ) without affecting the hyperpolarization-induced switch . Fig 3C illustrates the mean-field property of our network: in spite of the heterogeneity of rhythms at a cellular and circuit scales , the variability of the rhythms in the population progressively shrinks as the network size increases and eventually vanishes for very large neuronal populations ( mean-field limit ) . The existence of two discrete mean-fields is a consequence of the cellular switch . Heterogeneity and variability in the population contribute to the tunability of the two discrete mean-fields . Because of its cellular nature , the network switch described in this paper allows for precise spatio-temporal control of the population: its temporal resolution is only limited by the kinetics of the slow regenerative channels; its spatial resolution is only limited by the spatial resolution of receptors to neuromodulatory inputs . Such a mechanism enables spatiotemporal control of a network state at multiple resolutions . Fig 4 illustrates two simple forms of spatiotemporal control in our model . Fig 4A shows how the network state of a 160-cell population is affected by the number of active hyperpolarizing neuromodulatory pathways . For the sake of the illustration , the neuromodulatory input is equally divided into 8 pathways . A global network rhythm is induced through the hyperpolarization of a sufficient subpopulation of I-cells ( 4 pathways out of 8 in the chosen illustration ) . Both the rhythmic spectral power and the frequency distribution are modulated by the number of active pathways . This example shows that the temporal properties of a network rhythm can be modulated by spatially varying neuromodulatory inputs , even in the absence of changes in either intrinsic or synaptic parameters of the network . Fig 4B illustrates the spatial modulation of a network state in a fully connected population . In the proposed excitatory-inhibitory topology , a spatial clustering in GABABR only is sufficient to create clusters in the network switch of the excitatory population ( Fig 4B , left ) . When one of the neuromodulatory pathways is activated , it only affects the E-cells that have sufficiently strong GABAB connections with the modulated inhibitory subpopulation . As a result , the LFPs of the excitatory subpopulations are orchestrated individually and the temporal rhythm is modulated along the spatial axis: each spatially localized neuromodulatory pathway can switch ON and OFF a corresponding rhythm ( Fig 4B , right ) . Each rhythm has a specific spectral power signature , and the spatial organization of the network state is controlled at the spatial resolution of the neuromodulation . In contrast to the excitatory population , the rhythmic activity of the inhibitory population is spatially uniform in our illustration . This is because the AMPA connections from the excitatory to the inhibitory population are not clustered . As a result , the spatiotemporal organization of the network states is different in each subpopulation . The spatiotemporal control of the network entirely rests on decoupling the switching control from the tuning control . The switching is always at a cellular resolution . The tuning is at any temporal scale slower than the cellular rhythm and at any spatial scale between the cellular and network level . Because it controls the switch , a slow negative conductance at the cellular level is again critical to the spatiotemporal controllability of the network . Our analysis so far has focused on the tuning of the slow rhythms in the oscillatory network state . Fig 5 illustrates how the model also accounts for robust tuning of the active state , which is critical to modulate the transmission properties of a population . The figure shows how E-cells process impulses and sinusoidal inputs in two different configurations . The only difference between the two configurations is the ratio between GABAA and GABAB synaptic connection strength . This is consistent with a physiological regulation through synaptic plasticity [56 , 57] . The switch control of the network is insensitive to the synaptic ratio but the transmission properties of the E-cells are markedly affected ( Fig 5A ) . A high ratio enhances a linear-like response of E-cells: a short pulse of excitatory current triggers a spike , and a sinusoidal input entrains a phase-locked train of bursts ( Fig 5A , top ) . In contrast , a low ratio enhances a detector-like response: a short pulse of excitatory current triggers a burst of spikes , and a sinusoidal input triggers burst of spikes whose frequency is maximum at the onset of the rising phase of the input signal ( Fig 5A , bottom ) . This modulation of transmission properties is further quantified in Fig 5B , which illustrates the average response of a neuronal population to a sinusoidal input for three different synaptic ratios . For high ratio , the peak response is reached at the peak amplitude of the input signal . For low ratio , the peak response is reached when the input signal crosses a threshold from below . Such a modulation of the transmission mode is reminiscent of physiological observations in thalamocortical loops [7 , 8] . We emphasize that the modulation of the active state illustrated in Fig 5 is once again critically dependent on a slow negative conductance at the cellular level . The ration between GABAAR and GABABR primarily regulates the membrane polarization . When GABAAR connections dominate , the inhibitory drive from the I-cells maintains E-cells close to GABAAR reversal potential , i . e . chloride reversal potential ( set to -70mV in our model ) . At this potential , T-type calcium channels are inactivated , and the slow negative conductance is turned off: the E-cells exhibit the physiological signatures of the fast excitability mode: spike excitability and tonic firing [45 , 47 , 48] . In contrast , when GABABR connections dominate , the inhibitory drive maintains E-cells close to GABABR reversal potential , i . e . potassium reversal potential ( set to -85mV in this model ) . At this potential , T-type calcium channels are deinactivated , and the slow negative conductance is turned on: the E-cells exhibit the physiological signatures of slow excitability: burst excitability and endogenous bursting [45 , 47 , 48] .
In an effort to model the spatiotemporal organization of brain states , we proposed a simple neuronal network architecture that exploits modulation and heterogeneity of rhythms at a cellular resolution to tune the spatiotemporal signature of large rhythmic populations . At the core of our model lies a separation between switching mechanisms and tuning mechanisms . The mechanism of the switch is simple . It involves a single cellular property , occurs at a single temporal scale , and it is uniformly shared in the population . In contrast , the tuning mechanisms are multiple . They involve both cellular and synaptic properties , occur at many temporal and spatial scales , and can be highly heterogeneous in the population . A transition between two discrete states exists at every scale , from cells to networks , because of the uniform cellular switch . In contrast , the spatiotemporal signatures of each discrete state can be continuously tuned across the entire population , shaping robust and tunable network states . The central contribution of our model is to show that the cellular switch is essential to shape the network properties . It is necessary to the robustness of the network switch and enables the tunability of the network states by the remaining intrinsic and extrinsic conductances . The network control described in this paper only rests on two specific features: a cellular property to control the intrinsic slow negative conductance , provided by T-type calcium channels in our model , and a network topology that reciprocally interconnects an excitatory subpopulation and an inhibitory subpopulation . Those two properties are widely shared among a variety of circuits that exhibit fast control of network rhythms [7 , 8 , 13–19 , 30 , 53–55 , 58–67] . The canonical example is the thalamus , where both the role of rebound rhythms between ( excitatory ) thalamocortical neurons and ( inhibitory ) reticular neurons and the importance of T-type calcium channels have long been recognized in controlling network oscillatory states associated to sleep and attention [7 , 8 , 30 , 53–55 , 58–62 , 65] . Basal ganglia provide another example where the control of beta oscillations has been linked to rebound rhythms between ( excitatory ) subthalamic neurons and ( inhibitory ) external globus pallidus ( GPe ) neurons . A large amount of recent experimental evidence also demonstrates the importance of T-type calcium channels in the modulation of those rhythms . At the cellular level , experimental evidence shows that both STN and GPe neurons possess the ionic currents to undergo an excitability switch [53 , 54 , 64] . At the network level , oscillations have been recorded in the STN-GPe network in vitro and in vivo [13 , 14 , 39] . Fluctuations of the network state , and more specifically in the coherence and strength of beta oscillations have been linked to voluntary movement initiation , both in animals [68 , 69] and in humans [14 , 68 , 70–73] . Experimental studies show a prospective increase in beta synchrony prior to voluntary movements [74] and an event-related desynchronization in the beta band during movement [70 , 72] . Initiation of voluntary movements is also linked to an increase in dopamine and , in particular , to a transient increase in the activity of nigrostriatal circuits ( phasic dopamine release ) [75 , 76] . This dopamine transient increase triggers the decrease of the beta-band activity coherence and power [77] . Those observations are consistent with the predictions of our model under a transient modulatory input . Fluctuating brain states have also been described in the cortex . Models of the layer V of the cortex involved in vision include an excitatory-inhibitory network and T-type calcium currents [78] . In this brain region , oscillatory activity in the alpha band ( 8–12 hertz ) gates incoming signals by inhibiting task-irrelevant regions , thus routing signals to task-relevant regions [79 , 80]: alpha oscillations provide a functional inhibition and reduce the processing capabilities . At a cellular level , the excitability switch modeled in this paper is responsible for rebound bursting and hyperpolarization-induced bursting , two mechanisms that have been widely observed in experiments [7 , 53–55 , 58–60] . We stress that the excitability switch is distinct from the extensively studied post-inhibitory rebound ( PIR ) [7 , 49 , 51 , 52] . For instance , T-type calcium channels contribute to the switch mechanism through their slow activation , which is an intrinsic source of slow negative conductance , while they contribute to the post-inhibitory rebound through their inactivation , which is an intrinsic source of positive conductance . Other channels , such as HCN channels , only contribute to the rebound but do not contribute to the switch . The importance of T-type calcium channels has long been emphasized for their contribution to rebound mechanisms , both in central pattern generators and in mammalian brain rhythms [7 , 49 , 51 , 52] . The novelty of our model in that regard is to stress the importance of T-type calcium channels for their contribution to the switch in network state control . In the absence of the switch , rebound mechanisms alone do not suffice for network control of robust and tunable network rhythms . In the absence of slow regenerativity , a rebound rhythm in an excitatory-inhibitory network requires a specific resonance between the PIR and the kinetics of synaptic connections [41] . In this case , the circuit rhythm is fragile to changes in neuron intrinsic properties and synaptic connectivity . This fragility severely restricts the heterogeneity of rhythms in the population . A particular manifestation of the distinction between excitability switch and rebound property is provided in Fig 5 . In this figure , excitatory cells exhibit bursts both in the active and oscillatory states of the network . However , they participate in a rebound mechanism only in the oscillatory state . This change of rebound properties involves no change in the connectivity . It only results from a switch between two types of excitability . The novelty and significance of the switch mechanism at a network level is that it is largely independent of the network connectivity . Our paper differs in that regard from earlier computational studies that have studied modulations of network rhythms through modulations of the connectivity [39–41] . A common mechanism in those models is that stronger synchrony in the population relies on stronger connectivity [81] . In such models , active network states are associated to weak connectivity and asynchronous rhythmic activity whereas oscillatory states are associated to strong connectivity and synchronous rhythmic activity . Instead , in the present paper , the switch between the active and oscillatory state occurs without changes in connectivity . The connectivity is always weak , allowing for heterogeneous rhythms both in the active and oscillatory state . Stronger connectivity reduces the heterogeneity of rhythms in the population . Instead , a shared cellular switch allows for synchronous events even in the presence of heterogeneity . Most network computational models in the literature lack the cellular switch studied in the present paper . This is evident for all models that focus on the synaptic connectivity and only model rate or spiking properties at a cellular level . But it is also the case for many models that account for bursting properties at a cellular level but lack a switch of excitability . Even models that include T-type calcium channels often model the activation as instantaneous [82–88] . Those models can simulate bursts or rebound properties for specific parameter values but the absence of a slow negative conductance makes them fragile and not tunable [42] . The fact that most models neglect the slow kinetics of the calcium channel activation provides further evidence of computational models that account for rebound mechanisms but do not account for a cellular switch of excitability . It illustrates that the distinction between the two mechanisms has not received much attention . This is not to say that modeling the switch requires more biophysical details than modeling the rebound . The recent paper [89] shows that a simple integrate-and-fire model is sufficient to model the switch provided that it contains distinct fast and slow thresholds to account for the two distinct types of excitability . It also explains why existing integrate-and-fire models , which have only one threshold , cannot model the excitability switch even in the presence of adaptation variables .
All simulations were performed using the Julia programming language . Analysis were performed either in Julia or in Matlab . Julia and Matlab code files are freely available at http://www . montefiore . ulg . ac . be/~guilldrion/Files/Drionetal2018-Code . zip and https://osf . io/k86en . Single-compartment Hodgkin-Huxley models were used for all neuron models , following the equation CmV˙=−∑Iion+Iapp , where Iion corresponds to the ionic currents and Iapp is an externally applied current . The model is composed of a leak current Ileak=g¯leak ( V−Vleak ) , a transient sodium current INa=g¯NamNa3hNa ( V−VNa ) , a delayed-rectifier potassium current IK , D=g¯K , DmK , D4 ( V−VK ) , a T-type calcium current ICa , T=g¯Ca , TmCa , T3hCa , T ( V−VCa ) , a calcium-activated potassium current IK , Ca=g¯K , CamK , Ca ( [Ca] ) ( V−VCa ) and a hyperpolarization-activated cation current IH=g¯HmH ( V−VH ) , where m represents activation variables and h represents inactivation variables . The dynamics of voltage-gated activation and inactivation variables are modeled using the equation τx ( V ) m˙x=mx , ∞ ( V ) −mx , where mx , ∞ ( V ) =1/ ( 1+exp ( V−VhalfVslope ) ) and τx ( V ) =A−B/ ( 1+exp ( V−DE ) ) , except for τh , Na ( V ) , which is given by τhNa ( V ) = 0 . 67/ ( 1 + exp ( ( V + 62 . 9 ) /−10 . 0 ) ) ) * ( 1 . 5 + 1/ ( 1 + exp ( V + 34 . 9 ) /3 . 6 ) ) ) . The values of Vhalf , Vslope , A , B , D et E for each variable are given in Table 1 , and the corresponding curves are plotted in Fig 6 . The calcium-dependent activation of the calcium-activated potassium current is modeled as follows: mK , Ca ( [Ca] ) = ( [Ca]/ ( [Ca] + KD ) ) 2 . Parameters used in simulations were as follows: Cm = 1 μF/cm2 , VNa = 50 mV , VK = −85 mV , VCa = 120 mV , Vleak = −59 mV , VH = −20 mV . All maximal conductance values were picked randomly with respect to a uniform distribution in the following ranges ( in mS/cm2 ) : g¯leakϵ[0 . 0475 , 0 . 5575] , g¯Naϵ[135 , 205] , g¯K , Dϵ[20 , 60] , g¯Ca , Tϵ[0 . 375 , 0 . 725] , g¯k , Caϵ[3 , 5] , g¯Hϵ[0 . 0095 , 0 . 0105] . Calcium dynamics followed the equation [Ca]˙=−k1ICa , T−k2[Ca] where k1 and k2 were also picked randomly with respect to a uniform distribution ( k1ϵ [0 . 075 , 0 . 125] and k2ϵ [0 . 0075 , 0 . 0125] ) . In the case of instantaneous T-type calcium channel activation , the T-type calcium current was modeled as ICa , T=g¯Ca , TmCa , T , ∞3 ( V ) hCa , T ( V−VCa ) . Neuron models were connected via AMPA , GABAA and GABAB connections using the following equations: IAMPA=g¯AMPAAMPA ( V−0 ) , IGABA , A=g¯GABA , AGABAA ( V−VCl ) , and IGABA , B=g¯GABA , BGABAB ( V−VK ) , where AMPA , GABAA and GABAB are variables whose variation depends on the presynaptic potential Vpre following the equations AMPA˙=1 . 1Tm ( Vpre ) [1−AMPA]−0 . 19AMPA , GABAA˙=0 . 53Tm ( Vpre ) [1−GABAA]−0 . 19GABAA , GABAB˙=0 . 016Tm ( Vpre ) [1−GABAB]−0 . 0047GABAB , Tm ( Vpre ) = 1/ ( 1 + exp ( − ( Vpre − 2 ) /5 ) ) . Synaptic weights were taken randomly with respect to a uniform distribution around a central value ( g¯syn=g¯syn , central±g¯syn , central/8 ) . AMPA receptor reversal potential was set to 0 mV , GABAA receptor reversal potential was set to chloride reversal potential ( VCl = −70 mV ) and GABAB receptor reversal potential was set to potassium reversal potential ( VK = −85 mV ) . GABAB receptor activation was considered 50 time slower than GABAA and AMPA receptor activation . The local field potential ( LFP ) dynamics results from the collective synaptic activity of the neuronal population and is modeled by the normalized sum of the postsynaptic currents . The LFPs are low-pass filtered at 100 Hz via a fourth order Butterworth filter to reflect the use of macro-electrodes in LFP acquisition . The spectrogram analyses , or time-frequency plots , result from a logarithmic representation of the spectrogram of the short-time Fourier transform of the LFP . For the short-time Fourier transform , we consider a sampling frequency Fs = 1 kHz . Spatial clustering of GABAB connections was introduced by adding a Gaussian decay in the synaptic strength from neuron i to neuron j: gSD=gsyne− ( j−i ) 22cij2 where gsyn is the maximal synaptic strength , cij is the space constant controlling the spread of connectivity ( set to 0 . 8 ) , and i , j are the positions of the neuron in the populations E and I . gSD is normalized over the presynaptic population to get the same overall connection strength for each neuron in the postsynaptic population .
|
Brain information processing involves electrophysiological signals at multiple temporal and spatial timescales , from the single neuron level to whole brain areas . A fast and local control of these signals by neurochemicals called neuromodulators is essential in complex tasks such as movement initiation and attentional focus . The neuromodulators act at the cellular scale to control signals that propagate at potentially much larger scales . The present paper highlights the critical role of a cellular switch of excitability for the fast and localized control of cellular and network states . By turning ON and OFF the cellular switch , neuromodulators can robustly switch large populations between distinct network states . We stress the importance of controlling the switch at a cellular level and independently of the connectivity to allow for tunable spatiotemporal signatures of the network states .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"neurochemistry",
"action",
"potentials",
"neural",
"networks",
"engineering",
"and",
"technology",
"nervous",
"system",
"electrical",
"circuits",
"membrane",
"potential",
"electrophysiology",
"neuroscience",
"ion",
"channels",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"proteins",
"calcium",
"channels",
"biophysics",
"hyperpolarization",
"physics",
"biochemistry",
"cellular",
"neuroscience",
"neuromodulation",
"cell",
"biology",
"anatomy",
"synapses",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"electrical",
"engineering",
"neurophysiology"
] |
2018
|
Switchable slow cellular conductances determine robustness and tunability of network states
|
The ability to screen compounds in a high-throughput manner is essential in the process of small molecule drug discovery . Critical to the success of screening strategies is the proper design of the assay , often implying a compromise between ease/speed and a biologically relevant setting . Leishmaniasis is a major neglected disease with limited therapeutic options . In order to streamline efforts for the design of productive drug screens against Leishmania , we compared the efficiency of two screening methods , one targeting the free living and easily cultured promastigote ( insect–infective ) stage , the other targeting the clinically relevant but more difficult to culture intra-macrophage amastigote ( mammal-infective ) stage . Screening of a 909-member library of bioactive compounds against Leishmania donovani revealed 59 hits in the promastigote primary screen and 27 in the intracellular amastigote screen , with 26 hits shared by both screens . This suggested that screening against the promastigote stage , although more suitable for automation , fails to identify all active compounds and leads to numerous false positive hits . Of particular interest was the identification of one compound specific to the infective amastigote stage of the parasite . This compound affects intracellular but not axenic parasites , suggesting a host cell-dependent mechanism of action , opening new avenues for anti-leishmanial chemotherapy .
Leishmaniasis is caused by protozoan parasites of the genus Leishmania . The disease is endemic in the tropics , subtropics and the Mediterranean basin . There are three main clinical syndromes caused by different species of Leishmania . Cutaneous and muccocutaneous leishmaniasis result in large , painful sores that can take many months to heal [1] . Visceral leishmaniasis results in fever , weight loss , and damage to internal organs such as the spleen and the liver and may be fatal if left untreated [2] . Leishmania parasites are transmitted to mammalian hosts through the bite of phlebotomine sandflies . The parasites that develop in the mid-gut of the flies , called promastigotes , are flagellated and extracellular . Upon injection in the bloodstream of a mammalian host , promastigotes are rapidly phagocytosed by macrophages where they differentiate into the amastigote form . Amastigotes multiply in the macrophage parasitophorous vacuole , leading to destruction of the host cell and release of free amastigotes into the bloodstream , where they are capable of infecting new phagocytic cells [3] . Current treatment for leishmaniasis relies on chemotherapy , as no efficient vaccine is available . Sodium stibogluconate and amphotericin B have been the first line treatment; however , they have significant side effects and unresponsiveness to sodium stibogluconate has been reported for many years [4]–[6] . A few new anti-leishmanial drugs have been recently released ( miltefosine , paromomycin ) , but they also have drawbacks including cost and toxicity [7] . In addition , it has been shown in vitro that in some cases resistance can be easily induced [8] . New therapeutics are therefore urgently needed . Recognition of this need in recent years has led to partnerships between a number of foundations , agencies and universities to support the discovery of anti-parasitic agents , including anti-leishmanials . Lead discovery , one of the bottlenecks in the pipeline for novel anti-leishmanial drugs , would be facilitated by improved high-throughput technology allowing for the ability to screen large number of candidates [9] , [10] . Several anti-leishmanial high-throughput screens have been reported [11]–[13] . Primary screens often target the parasite promastigote stage because of ease of culture and manipulation . Indeed , promastigotes from several Leishmania species are easily maintained as cell suspension in vitro . However , as the promastigote is the form of the parasite in the insect vector , it is not the appropriate target for an anti-leishmanial drug [14] . Culture conditions for axenic amastigotes have been developed in order to facilitate the study of this stage of the parasite [15] , [16] . This has allowed amastigotes to be screened in a high-throughput manner [17] . However , expression arrays comparing L . infantum axenic amastigotes and amastigotes isolated from macrophages have shown differences in several cellular processes , including metabolism , intracellular transport and response to oxidative stress [18] . These observations highlight the importance of the host macrophage in driving the parasite to specific adaptations . The axenic amastigote model therefore has limitations as it does not encompass many aspects of intracellular parasite development [19] . Compounds active against axenic forms might be unable to reach the intracellular amastigote because of their inability to cross host cell membranes or maintain stability under low pH . Other compounds may need to be metabolized by the macrophage to gain activity . Finally , the macrophage itself might be directly targeted , thereby leading to parasite growth inhibition [20] . We have developed a host cell-based screening assay using a human macrophage cell line infected with L . donovani , one of the agents of visceral leishmaniasis . This assay format enables screening of compounds directly against the intracellular stage of the parasite . This assay was used to screen a library of 909 bioactive compounds consisting largely of FDA approved small molecules . In order to compare the efficiency of this screening method with traditional high-throughput screening assays , the same compound library was screened against free living promastigotes . A compound leading to sixty percent parasite growth inhibition at 10 µM was considered a hit in both assays . 59 hits were identified in the promastigote assay of which only 26 were also considered hits in the intracellular amastigote assay . Only one compound was specifically active against the intracellular amastigote stage . We conclude that the promastigote assay fails to identify all active compounds and leads to a rate of 56% false positives .
THP-1 cells ( human acute monocytic leukemia cell line – ATCC TIB202 ) were grown in RPMI supplemented with 10% Foetal Bovine Serum ( FBS ) and 50 µM 2-Mercaptoethanol at 37°C in 5% CO2 . L . donovani promastigotes [strain 1S , clone 2D ( MHOM/SD/62/1S-cl2D ) ] were grown at 27°C in RPMI supplemented with 10% FBS and 10% Brain Heart Tryptose medium ( BHT ) [21] . Differentiation of promastigotes into axenic amastigotes was achieved by dilution of 5×105 promastigotes in 3 ml of low-pH axenic amastigote media ( 15 mM KCl; 136 mM KH2PO4; 10 mM K2HPO4·3H2O; 0 . 5 mM MgSO4·7H2O; 24 mM NaHCO3; 22 mM glucose; 1 mM glutamine , 1× RPMI 1640 vitamin mix , 10 µM folic acid , 100 µM adenosine , 1× RPMI amino acid mix , 5 µg/ml hemin , 50 U/ml of penicillin , 50 µg/ml of streptomycin , 25 mM MES and 20% FBS . The pH was adjusted to pH 5 . 66 at 22°C , yielding a final pH of 5 . 5 at 37°C ) [22] . Axenic amastigotes were grown in ventilated flasks at 37°C in 5% CO2 . A library of 909 bioactive compounds was donated by Iconix Biosciences . These compounds were dissolved in DMSO at a stock concentration of 1 mM . Amphotericin B ( Sigma ) was used as a positive control . L . donovani promastigotes from an exponentially growing culture were diluted to 106/ml in RPMI containing 10% FBS and 10% BHT . The diluted culture ( 99 µl/well ) was dispensed in sterile 96-well flat white opaque assay plates ( Greiner Bio-One ) using a WellMate multichannel dispenser ( Matrix ) . 1 µl of 1 mM test compound dissolved in DMSO was added to the plates for a final concentration of 10 µM compound and 1% DMSO . Amphotericin B was added as a positive control ( final concentration 2 µM , 1% DMSO ) and as a negative control , 1 µl DMSO was added ( 1% final concentration ) . Compounds and controls were added to the assay plate with the robotic dispenser Biomek FXp liquid handler ( Beckman Coulter ) . Promastigotes were incubated with the compounds for 72 h at 27°C . The parasites were then lysed by adding 50 µl of CellTiter-Glo ( Promega ) and placed on an orbital shaker for 5 min at room temperature . After lysis , the resulting ATP-bioluminescence was measured using the Analyst HT plate reader ( Molecular Devices ) . Percentage inhibition of parasite growth was calculated for each well as [1- ( RLUx-RLU+ ) / ( RLU--RLU+ ) ]*100 where RLUx , RLU+ and RLU- are respectively the Relative Light Units for each well , positive ( amphotericin B ) and negative ( DMSO ) controls . A screening window coefficient , denoted Z' factor , was used to evaluate the performance of the assay . The Z' factor , calculated as 1- ( 3σc++3σc− ) / ( µc+−µc− ) where σc+ , σc− , µc+ and µc− are respectively the standard deviation and mean values of positive and negative controls , is reflective of the assay signal dynamic range and the data variation associated with signal measurement [23] . For GI50 determinations ( half maximal inhibitory concentration ) , compounds were serially diluted 3-fold in DMSO , with final assay concentrations ranging from 50 µM to 0 . 02 µM ( 1% final concentration of DMSO ) . GI50 curve fitting was carried out using GraphPad Prism 4 Software ( GraphPad Software Inc . , San Diego , CA ) . Sterile , black , 96-well , clear bottom plates ( Greiner Bio-One ) were seeded with exponentially growing THP-1 ( 5×105cells/ml ) . THP-1 were treated with 0 . 1 µM phorbol myristate acetate ( PMA , Sigma ) at 37°C for 48 h to achieve differentiation into adherent , non-dividing macrophages . Maturation of THP-1 cells towards monocyte-macrophage like cells is essential to avoid parasitized cells being overgrown by replicating cells . After activation by PMA , cells were washed and incubated with complete RPMI medium containing stationary phase L . donovani promastigotes at a macrophage/promastigote ratio of 1/15 . After 4 h incubation at 37°C , non-internalized promastigotes were removed by 2–3 successive washes with RPMI containing 5% FBS and 5% horse serum . Test compounds ( 10 µM ) , positive control ( 2 µM amphotericin B ) or negative control ( 1% DMSO ) were then added to the cultures using a Biomek FXp liquid handler ( Beckman Coulter ) . Cultures were incubated at 37°C for 72 h . Cells were then washed with phosphate-buffered saline ( PBS ) , fixed for 30 minutes with 4% formaldehyde , rinsed again with PBS , stained for 2 h with 4′ , 6′-diamidino-2-phenylindole ( DAPI 300 nM ) and finally washed with PBS . For GI50 determination , compounds were serially diluted 3-fold in DMSO , with final assay concentrations ranging from 50 µM to 0 . 02 µM ( 1% final concentration of DMSO ) . Images were acquired with an INCell Analyzer 1000 automated epi-fluorescent microscope ( G . E . Healthcare ) . The excitation and emission filters used to detect DAPI were 350/50 nm and 460/40 nm respectively . Eight image fields were acquired per well with a 20X objective . The proprietary INCell Developer Toolbox 1 . 7 software was used for image analysis . Segmentation parameters were set to identify host nuclei with a minimum area of 250 µm2 and parasite kinetoplast with an average area of 1 µm2 . The intensity of parasite nucleus was too low to be detected with a 20X objective . A border , representing the boundary of the cell , was drawn around the nucleus ( total area between 700 and 2000 µm2 ) . Only parasites found within this area were included in the calculation to eliminate extracellular parasites . False positive parasite detection in the nucleus was also excluded from the calculation . Host cell nuclei and parasite kinetoplasts were counted and the ratio of parasites DNA to host nuclei was selected as the measurement output . Percentage inhibition of parasite growth was calculated as [1- ( P/hcx-P/hc+ ) / ( P/hc−-P/hc+ ) ]*100 where P/hcx , P/hc+ and P/hc− are parasite per host cell ratio for every well , positive control ( amphotericin B ) and negative control ( DMSO ) respectively . Calculation of Z' factor and GI50 curve fitting were carried out as described above . L . donovani axenic amastigotes ( 5×105 cells/ml in axenic amastigote media ) were dispensed in sterile 96-well flat white opaque assay plates ( Greiner Bio-One ) using a WellMate multichannel dispenser ( Matrix ) . Compounds were serially diluted 3-fold in DMSO , with final assay concentrations ranging from 50 µM to 0 . 02 µM ( 1% final concentration of DMSO ) . 1% DMSO and amphotericin B ( 2 µM , 1% DMSO final concentration ) were added as negative and positive controls respectively . Axenic amastigotes were incubated with the compounds for 72 h at 37°C with 5% CO2 . Parasite viability was then measured using CellTiter-Glo as described above . Calculation of Z' factor , percentage of parasite growth inhibition and GI50 curve fitting were carried out as described above .
We developed a 96-well plate , cell-based assay simple to manipulate and reproducible , enabling screening of a high number of compounds against intra-macrophage L . donovani . The human leukemia monocyte cell line THP-1 has been commonly used as a model for Leishmania infection and has been described as a suitable model for drug screening [24] , [25] . In vitro infection of macrophages by Leishmania and analysis of intracellular parasite growth requires a method allowing for robust detection , discrimination and counting of parasites and host cells . In our setting , THP-1 cells infected with L . donovani were stained with the DNA marker DAPI ( 4′ , 6′-diamidino-2-phenylindole ) allowing the visualization of host cell nuclei and parasite kinetoplasts . Images collected with an INCell Analyzer 1000 fluorescent microscope showed a significant size difference between host cell nuclei and parasite kinetoplasts . This feature was exploited for image segmentation and determination of the number of host cells and parasites ( Figure 1A–D ) . The ratio between total number of parasites and total number of host cells was calculated for each well . In addition , counts of host cell nuclei were used as a quantitative measure of cell toxicity induced by the compounds . Incubation of L . donovani with THP-1 for 4 hours at a ratio of 15 parasites per host cell led to an average infection of 4 . 1 +/− 0 . 32 parasites per host cell after 72 h incubation , with an average of 30 +/− 9 percent of the cells infected and no change in the number of host cells ( Figure 1E ) . Growth of parasite and host cells was not affected by 1% DMSO ( Figure 2A–B ) . Amphotericin B , the first line drug used against leishmaniasis , was used as a positive control . At 2 µM amphotericin B did not affect THP-1 host macrophages ( Figure 2A ) but significantly inhibited growth of intracellular L . donovani ( Figure 2B ) with an estimated GI50 of 0 . 08 µM ( Figure 2C ) . This is comparable to GI50 values from previous reports [11] , [20] . The intracellular amastigote imaging assay described above was used to screen a library of 909 bioactive small molecules ( Iconix library ) . In the primary screen , compounds were assayed in duplicates at 10 µM . The average Z' value calculated per plate based on the positive and negative controls was 0 . 63 , indicating a satisfactory robustness of the assay . Sixty percent parasite growth inhibition in at least one of the replicates was the cut-off arbitrarily determined for hit selection . This low threshold was purposely selected to evaluate sensitivity of the assay and guarantee identification of all active compounds . In addition , compounds toxic to the host cell , determined as inducing more than 20% reduction in THP-1 numbers , were excluded . A total of 27 compounds met these criteria and were selected for further analysis . This list of active chemicals included previously identified anti-leishmanials such as amphotericin B , pentamidine isothionate and tamoxifen citrate , thus validating the ability of the screen to identify molecules active against Leishmania . The Iconix library was screened in parallel against L . donovani free-living promastigotes . Promastigote viability was determined after 72 hours incubation with the compounds , using an ATP-bioluminescence assay previously described for high-throughput screening against Trypanosoma brucei [26] . This assay measures luminescence produced by luciferase in presence of cellular ATP; the intensity of light is proportional to the amount of ATP released and correlates with the number of viable parasites ( data not shown ) [27] . Amphotericin B at 2 µM was used as a positive control and 1% DMSO as a negative control . In the primary screen , compounds were assayed at 10 µM . The assay was robust with an average Z' value of 0 . 72 . Consistent with the image-based assay targeting intracellular L . donovani , 60% parasite growth inhibition was the cut-off used for active compound selection . Fifty-nine compounds were selected as hits for further validation . The comparison of the results obtained for these screens showed that out of the 27 hits identified in the amastigote screen , 26 were also present in the promastigote screen . Only one compound , naloxonazine , showed complete specificity for the intracellular amastigote stage . Out of the 59 compounds identified in the promastigote screen , 19 were considered toxic to the THP-1 macrophage ( Figure 3 and Table 1 ) . GI50 values for these 60 hits ( 59 identified in the promastigote screen and one intracellular amastigote-specific hit ) were then established for both stages of the parasite . 15 compounds ( 25% of the hits ) were equipotent against both stages of the parasite . 14 compounds ( 23% ) were more potent against the intracellular amastigotes while 13 compounds ( 22% ) were more active against the promastigotes . The remaining compounds were toxic to the host cell ( Table 1 ) . As axenic amastigotes have been considered to mimic the intracellular stage of the parasite , we analyzed their sensitivity to the 60 hits described above . This study indicated that compound activity against axenic amastigotes mostly correlated with promastigotes . The specific activity of naloxonazine against intracellular amastigotes was confirmed as this compound showed no activity against promastigotes or axenic amastigotes ( Table 1 ) . The Iconix collection contained two opioid receptor antagonists , naloxone and naloxonazine . The first was not selected as a hit in any of the screens described above while the latter showed specific activity against the intracellular amastigote stage . To confirm these primary observations , the activity of both compounds was tested against promastigotes , intracellular and axenic amastigotes . Naloxonazine exhibited specific activity against intracellular amastigotes ( GI50 intracellular amastigote: 3 . 45 µM; GI50 THP-1: 33 . 8 µM; GI50 promastigote: >50 µM; GI50 axenic amastigote: >50 µM ) , while naloxone was inactive against all parasite forms and not toxic to the host macrophage ( Figure 4 ) . At a curative concentration , the selectivity window of naloxonazine was reduced ( GI90 intracellular amastigote: 12 . 5 µM; GI90 THP-1: 50 µM ) , limiting the possibility of using naloxonazine for treatment .
Current chemotherapy for Leishmaniasis has several drawbacks , including cost , toxicities , route of administration , and the emergence of drug resistance . The pipeline for anti-leishmanial drugs therefore needs to be filled with new compounds . As the discovery of new and original leads suitable for optimization and drug development is dependent on the ability to screen many compounds , assays should be rapid , inexpensive and reproducible [14] . In addition , for pathogens displaying several life stages like Leishmania , there is a need to determine the best parasite stage to target . In the case of Leishmania there are three major options: first , targeting the extracellular living promastigote stage , second the axenic amastigote and third the intracellular amastigote stage . The first and second options meet the reproducibility , rapidity and low cost requirements for high-throughput screens , due to the ease in manipulating promastigotes or axenic amastigotes in vitro . This has been demonstrated by Siqueira-Neto et al . and Sharlow et al . who described screening 26 , 500 and 200 , 000 compounds against the Leishmania promastigote stage , respectively [11] , [12] . These assays monitored parasite viability by measuring products from metabolically active cells ( The Alamar Blue assay involves reduction of resazurin into fluorescent resorufin by live cells , while CellTiter-Glo luciferase catalyzes the production of luminescence in the presence of cellular ATP ) . The axenic amastigotes offer the ability to screen easily the relevant-like stage of the parasite and allow testing the potency of compounds under low pH conditions . However , the major disadvantage of these two approaches is the absence of the host cell in the assays: the natural niche of the parasite is not taken into account and aspects of parasite biology such as host-parasite interactions or accessibility of the target are ignored . The intracellular amastigote stage has been logically designated as the more relevant target for primary screening against Leishmania , but previous methods were labor intensive and would not support automation [19] , [28] . Methods traditionally used to detect and estimate the number of intracellular Leishmania include Giemsa staining or the use of reporter gene-expressing parasites ( green fluorescent protein , luciferase or beta-galactosidase ) [29] , [30] . Giemsa staining is cumbersome as it needs manual counting . Reporter gene-expressing parasites are a powerful alternative , but stable recombinant parasite populations are required and they do not allow concomitant analysis of the host cell . Here we describe a high content assay that allows the simultaneous visualization of both host cell nuclei and parasite kinetoplasts by using the DNA marker DAPI . The significant difference in the size of these two organelles facilitates discrimination between host cells and parasites , and thus accurate counting of both entities . Reduction in the number of kinetoplasts gives a measure of inhibition of parasite growth , while a reduction in the number of host cell nuclei is indicative of compound cytotoxicity . Thus , this image-based assay allows the identification of leishmaniocidal as well as leishmaniostatic compounds . All steps of this assay are amenable to automation and could be reduced to 384-well format resulting in a robust high-throughput screening methodology . To evaluate what differences might be obtained from screening against the extracellular versus the intracellular parasite stages , we screened the same set of 909 compounds against both L . donovani promastigotes and intracellular amastigotes . We observed that the majority of the hits identified with the intracellular amastigote screen , defined as inducing 60% parasite growth inhibition at 10 µM , were also found in the promastigote screen . One compound showed specific inhibition of intracellular amastigote and was completely inactive against promastigotes . Fifty-six percent of the hits from the promastigote screen were not found in the intracellular amastigote screen . These results indicated that a promastigote screen failed to identify all active compounds and led to 56% of compounds being likely false positives . Thus , while the promastigote stage appears suitable for high-throughput screening , a fraction of the hits would be missed; furthermore , a high rate of false positives is characteristic of primary screens against promastigotes , underlying the importance of evaluating compound activity against intracellular amastigotes at least in a secondary screen . This is in accordance with the findings of Siqueira-Neto et al . who found that only 4% of their hits identified in a promastigote primary screen were active in an intracellular context [11] . Advantages of the intracellular amastigote assay include cell-health information , very low cost of consumables and a reduced necessity for secondary assays . The importance of the host cell in the assay was also demonstrated by the dose response study against axenic amastigotes ( i . e . amastigote-like stage obtained from differentiation of promastigotes in vitro in the absence of a host cell ) ; although this parasite form should mimic the intracellular stage , the activity of compounds against axenic amastigotes mostly correlated with promastigotes rather than intracellular amastigotes . A similar assay was also successfully developed for screening drugs against the intracellular stage of the related parasite Trypanosoma cruzi [31] . Screening the Iconix library against intracellular T . cruzi identified 56 hits , among which 8 were also hits in the Leishmania screen presented here . Six of these were found to be more active against the intracellular amastigote stage of L . donovani compared to promastigotes , indicating inter-species activity of compounds only for the intracellular stages of these two different parasites . Fifty percent of the compounds that were preferentially active against intracellular amastigotes are known to bind mammalian/eukaryotic G protein coupled receptors ( opioid receptors , serotonin or dopamine receptors and adrenergic receptors ) . Heterotrimeric G proteins are absent in trypanosomatids [32] and we could not find convincing homologs of opioid receptors in the Leishmania genome . Compounds described as ligands of G protein coupled receptors may have different targets among parasitic proteins , leading to mechanisms of inhibition independent of the host cell . This is the case of a serotonin receptor agonist that interferes with P . falciparum growth by blocking a surface membrane channel [33] , or a κ-opioid agonist active against T . brucei whose target remains to be identified as no homolog of κ-opioid receptor is found in the T . brucei genome [34] . However , the fact that the ligands of G protein coupled receptors identified in this study , showed a preferential activity towards the intracellular amastigote stage , also highlights the potential value of these host cell signaling pathways as targets . Previous reports demonstrated the involvement of such receptors in inhibition of infection by several intracellular pathogens including Leishmania [35]-[37] . Targeting host factors essential for parasite development is an emerging drug discovery paradigm . It is assumed to be less likely to induce drug resistant pathogens and offers the possibility to repurpose drugs by exploiting compounds currently used for diseases unrelated to microbial infection [38] , [39] . Interestingly , one compound out of 909 was active against the intracellular amastigote stage but was completely inactive against promastigotes . This compound , naloxonazine , was also inactive against axenic amastigotes , indicating that its activity is dependent on a macrophage function . Naloxonazine is described as an irreversible µ1-opioid receptor antagonist [40] . There is evidence for the presence of opioid receptors on cells of the immune system [41] , [42] and it is known that opioids are involved in modulation of host resistance to infectious diseases [43] , [44] . The immune response of mice infected with L . donovani has been shown to be influenced by the opioid receptor agonist morphine , but the receptors involved and the mechanism leading to this immunomodulation remain unknown [45] . Loperamide , a µ-opioid receptor agonist , was also identified in this study as inhibiting parasite growth , and this compound was also more potent against the intracellular stage of the parasite . Another µ-opioid receptor antagonist , naloxone , also present in the Iconix library , did not show any activity against L . donovani . The differential selectivity of naloxone and naloxonazine for opioid receptor binding sites might explain their differential activity against intracellular L . donovani [46] . Naloxone is monomer-like while naloxonazine appears as an inverted dimer ( Figure 4 ) . The presence of the macrophage appears to be essential for the activity of naloxonazine against L . donovani , but the underlying cellular and molecular mechanisms remain to be elucidated . Although naloxonazine itself would not meet the requirement for a therapeutic drug due to the reduced selectivity window at a curative concentration , it would be interesting to analyze other compounds that target the same host cell pathway . In summary , we report an automated screen against intracellular amastigotes of L . donovani . It has the advantage of screening against the relevant stage of the parasite , taking into consideration crucial aspects of its biology , and giving the opportunity to identify host factors critical for the establishment of infection . This is essential for the identification of new , original and diverse lead compounds for anti-leishmanial therapy .
|
Leishmaniasis , a disease caused by protozoan parasites of the genus Leishmania , is a poverty-related disease threatening 350 million people throughout the world . Drugs currently available to treat this disease are toxic to the patient and drug-resistant parasites are emerging . New therapeutics are therefore needed . Fortunately , interest in confronting the treatment challenges has grown and new technology has led to an increase in high-throughput screens conducted against Leishmania . In order to gain insight into the most efficient screening strategy , we compared two methods , one targeting the easily cultured insect-infective promastigote stage of the parasite , and the other , targeting the clinically relevant but more difficult to culture intracellular amastigote stage . We show that while a screen against promastigotes is amenable to automation , it fails to recognize all active compounds . These compounds revealed only by an intracellular assay might act on host cell pathways important for parasite development . Targeting such pathways is an emerging strategy in drug discovery against infectious diseases .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"small",
"molecules",
"microbiology",
"host-pathogen",
"interaction",
"parasitology",
"parastic",
"protozoans",
"leishmania",
"neglected",
"tropical",
"diseases",
"infectious",
"diseases",
"microbial",
"pathogens",
"biology",
"drug",
"discovery",
"biochemistry",
"leishmaniasis",
"protozoology"
] |
2011
|
A Screen against Leishmania Intracellular Amastigotes: Comparison to a Promastigote Screen and Identification of a Host Cell-Specific Hit
|
The Neurospora crassa mitochondrial tyrosyl-tRNA synthetase ( mtTyrRS; CYT-18 protein ) evolved a new function as a group I intron splicing factor by acquiring the ability to bind group I intron RNAs and stabilize their catalytically active RNA structure . Previous studies showed: ( i ) CYT-18 binds group I introns by using both its N-terminal catalytic domain and flexibly attached C-terminal anticodon-binding domain ( CTD ) ; and ( ii ) the catalytic domain binds group I introns specifically via multiple structural adaptations that occurred during or after the divergence of Peziomycotina and Saccharomycotina . However , the function of the CTD and how it contributed to the evolution of splicing activity have been unclear . Here , small angle X-ray scattering analysis of CYT-18 shows that both CTDs of the homodimeric protein extend outward from the catalytic domain , but move inward to bind opposite ends of a group I intron RNA . Biochemical assays show that the isolated CTD of CYT-18 binds RNAs non-specifically , possibly contributing to its interaction with the structurally different ends of the intron RNA . Finally , we find that the yeast mtTyrRS , which diverged from Pezizomycotina fungal mtTyrRSs prior to the evolution of splicing activity , binds group I intron and other RNAs non-specifically via its CTD , but lacks further adaptations needed for group I intron splicing . Our results suggest a scenario of constructive neutral ( i . e . , pre-adaptive ) evolution in which an initial non-specific interaction between the CTD of an ancestral fungal mtTyrRS and a self-splicing group I intron was “fixed” by an intron RNA mutation that resulted in protein-dependent splicing . Once fixed , this interaction could be elaborated by further adaptive mutations in both the catalytic domain and CTD that enabled specific binding of group I introns . Our results highlight a role for non-specific RNA binding in the evolution of RNA-binding proteins .
RNA-binding proteins play critical roles in post-transcriptional regulation of gene expression in all domains of life [1] . However , the complexity of this regulation is far greater in eukaryotes than in prokaryotes , reflecting both the larger number of RNAs requiring regulation and the evolution of new RNA processing and regulatory mechanisms . The latter include extensive RNA splicing and alternative splicing to produce different protein isoforms; an increased importance of RNA localization in larger and more complex eukaryotic cells; nonsense-mediated decay to prevent translation of intron-containing RNAs; and combinatorial regulation of mRNA translation and stability by RNA-binding proteins and miRNAs acting in ribonucleoprotein complexes [2]–[5] . These new modes of post-transcriptional regulation necessitated and were enabled by corresponding increases in the number and diversity of RNA-binding proteins and the evolution of new RNA-binding functions [3] , [6] . Thus far , however , the molecular mechanisms underlying the evolution of new RNA-binding functions have remained unclear . Cellular proteins that adapted to splice autocatalytic group I and group II introns provide powerful model systems for investigating how proteins evolve new RNA-binding functions . Group I and group II introns are found in prokaryotes and in the mitochondrial ( mt ) and chloroplast DNAs of some eukaryotes , with group I introns also found in the nuclear rRNA genes of certain fungi and protozoa [7] , [8] . Both types of introns are ribozymes that catalyze their own splicing as well as mobile genetic elements that can be horizontally transferred to different hosts where they propagate by inserting into new genomic sites [9] , [10] . Although some group I and II introns self-splice in vitro , most have acquired mutations that impair formation of a catalytically active RNA structure , necessitating the recruitment of cellular proteins to promote RNA folding for efficient splicing in vivo [11] , [12] . These group I and group II intron splicing factors include both host-encoded proteins , such as aminoacyl-tRNA synthetases ( aaRSs ) and translation factors , and intron-encoded proteins , such as DNA endonuclease and reverse transcriptases , that evolved secondarily to function in RNA splicing [11] . Such co-option of pre-existing proteins to function in splicing is pertinent to the evolution of splicing mechanisms in higher organisms , as emphasized by recent findings that a key spliceosomal protein , Prp8 , was derived from a group II intron-like reverse transcriptase [13] , [14] . One of the most extensively studied examples of a cellular protein that evolved to function in RNA splicing is the Neurospora crassa mtTyrRS ( CYT-18 protein ) , which acts as a splicing factor for mt group I introns [15]–[18] . Biochemical and structural studies showed that CYT-18 functions in splicing by recognizing and stabilizing the conserved phosphodiester backbone structure of group I intron RNAs [19]–[22] . This splicing function has been found only for those mtTyrRS of fungi belonging to the subphylum Pezizomycotina and can be traced to a series of structural adaptations of the protein that were acquired during or after the divergence of Pezizomycotina from Saccharomycotina [23] . CYT-18 and other mtTyrRSs are class 1 aaRSs that are closely related to bacterial TyrRSs [24] . They consist of an N-terminal catalytic domain , which binds the acceptor stem of tRNATyr , followed by an intermediate α-helical domain and a C-terminal anticodon-binding domain ( CTD ) , which bind the anticodon and variable arms ( Figure 1A and 1B; the catalytic and intermediate α-helical domains together are denoted the N-terminal domains or NTDs ) . Like its bacterial counterparts , CYT-18 functions as a homodimer , with each dimer binding either one molecule of tRNATyr or group I intron RNA [25]–[28] . CYT-18 binds group I introns by using both its N-terminal catalytic domain and CTD , but only some introns require the CTD for RNA splicing [29]–[31] . Both the N-terminal catalytic domain and CTD of Pezizomycotinia mtTyrRSs have distinctive structural adaptations that are absent in non-splicing mtTyrRSs , including the closely related Saccharomycotina mtTyrRSs [23] . These structural adaptations include a small N-terminal α-helical extension ( denoted H0 ) and a series of small insertions ( Ins 1–5 ) , whose presence correlates with RNA splicing activity ( Figure 1B ) [32] , [33] . The Pezizomycotina mtTyrRS also have a non-essential C-terminal tail of variable length ( C-tail; 13–152 amino acids ) appended to the CTD ( [29] and this work ) . Structural studies , including a co-crystal structure of a splicing-active CYT-18 protein lacking the CTD ( here denoted CYT-18 NTDs ) bound to a group I intron RNA ( the bacteriophage Twort orf142-I2 ribozyme ) , provided insight into group I intron binding by the N-terminal catalytic domain [22] , [33] . These studies showed that CYT-18 binds group I introns asymmetrically across the two subunits of the homodimer by using a newly evolved group I intron-binding surface on the side of the catalytic domain opposite that which binds tRNATyr . This new RNA-binding surface includes the N-terminal extension H0 , Ins 1 , and Ins 2 and provides an extended scaffold for the conserved phosphodiester backbone structure of the group I intron catalytic core . The CYT-18 constructs used for crystallography lacked the flexibly attached CTD , which has been problematic for X-ray crystallography of TyrRSs [33]–[36] . Recently , we determined an NMR structure of the isolated CTD of the splicing-active Aspergillus nidulans mtTyrRS , which is closely related to CYT-18 [37] . The structure showed that the mtTyrRS CTD resembles those of bacterial TyrRSs in having a fold similar to that of bacterial ribosomal protein S4 , but with novel structural features . The latter include three Pezizomycotina-specific insertions ( Ins 3–5 ) , with Ins 3 corresponding to an expansion of the flexible linker between the NTDs and CTD . Modeling of the NMR structure onto the CYT-18 NTDs+Twort co-crystal structure using distance constraints from directed hydroxyl-radical cleavage assays suggested that the two CTDs of the homodimeric protein bind opposite ends of a group I intron RNA . This model requires that the CTD of one subunit of the CYT-18 homodimer undergo a large shift on its flexible linker to interact with either tRNATyr or the group I intron RNA bound on opposite sides of the catalytic domain [37] . Thus far , however , there has been no structural data for a CYT-18 protein that contains both the NTDs and CTD , and the role of the CTD in promoting group I intron splicing has remained unclear . CYT-18 has been used as a model for the theory of constructive neutral evolution ( referred to here as “pre-adaptive evolution” ) . This theory holds that complex multi-protein or RNP complexes arise by a “ratchet-like process” in which a pre-existing neutral or mildly deleterious interaction is “fixed” by a mutation in one partner that makes it dependent upon the other to perform a biological function . Once fixed , this dependence can be further elaborated by adaptive changes in both partners , which increase reaction efficiency and co-dependence [38]–[40] . In the case of CYT-18 , this hypothesis suggests that an ancestral non-splicing fungal mtTyrRS had a pre-existing ability to bind group I introns , which became fixed when the intron RNA acquired mutations that impaired self-splicing , resulting in dependence upon the bound protein for structural stabilization [11] . After the interaction was fixed , further adaptive mutations in both the RNA and protein increased both the efficiency of RNA splicing and its protein-dependence . Early studies suggesting that CYT-18 recognized tRNA-like structural features of group I intron RNAs were cited as a prime example of a pre-adaptive interaction leading to the evolution of a new RNA-splicing function [39] , [41] . However , subsequent findings that CYT-18's N-terminal catalytic domain binds group I introns specifically by using a separate non-tRNA-binding surface [22] , [33] made the nature of the initial non-adaptive interaction unclear . Here , we used small angle X-ray scattering ( SAXS ) and biofchemical assays to investigate the solution structures of full-length CYT-18 protein and its CTDs and their mode of interaction with group I intron RNAs . The SAXS analysis shows that the CTDs of both subunits of the CYT-18 homodimer extend outward from the NTDs , but move inward to bind opposite ends of the group I intron RNA . Surprisingly , we find that the CTD of CYT-18 has a high non-specific RNA binding affinity , which may contribute to its interaction with group I intron RNAs , and that the Saccharomyces cerevisiae ( yeast ) mtTyrRS , which diverged prior to the evolution of splicing activity , can also bind intron RNAs non-specifically via its CTD . Finally , experiments with chimeric proteins show that the yeast CTD can replace CYT-18's to promote aminoacylation but not group I intron splicing . Our results suggest a scenario of pre-adaptive evolution in which the initial non-adaptive interaction between an ancestral mtTyrRS and group I intron RNA was non-specific binding by the CTD and highlight a role for non-specific binding in the evolution of new RNA-binding functions .
First , we used SAXS to investigate the conformational changes of CYT-18 and the position of its CTDs in the absence and presence of a group I intron RNA . Scattering data were collected for three CYT-18 constructs: CYT-18* , a wild-type protein truncated to delete most of the non-essential C-tail in order to simplify modeling and analysis; CYT-18 NTDs , which contains the N-terminal catalytic and α-helical domains , but lacks both the CTD and C-tail; and CTD , the isolated C-terminal anticodon-binding domain ( Figure 1C ) . CYT-18* is fully active in tyrosyl-adenylation , which measures the number of TyrRS active sites , and it functions similarly to full-length CYT-18 both in aminoacylation of Escherichia coli tRNATyr , a standard substrate for this protein , and in splicing the N . crassa mt large subunit rRNA ( Nc mt LSU ) intron , which requires a functional CTD ( Figure S1 ) . The CYT-18 NTDs construct is also fully active in tyrosyl-adenylation , but cannot aminoacylate tRNATyr as expected because of the lack of the CTD ( Figure S1A and S1B ) [30] . Figure 2 shows SAXS curves for all three proteins , and Table 1 summarizes size parameters calculated from the SAXS curves , including the protein molecular weight; the maximum dimension of the particle ( Dmax ) ; and the radius of gyration ( Rg ) , which is the root mean square distance to the center of mass of a particle and provides an estimate of the overall particle size [42] . For all three proteins , Kratky plots of the SAXS data show a bell shape curve with a distinct peak , indicative of a folded globular protein ( Figure S2 ) . Focusing first on the CYT-18 NTDs protein , the scattering curve overlays well ( χ = 1 . 9 ) with a theoretical scattering curve calculated from the previous CYT-18 NTDs crystal structure [33] by using the program CRYSOL ( Figure 2A , top curve ) [43] . The SAXS curve gave an estimated molecular weight of 84 . 4 kDa and Rg and Dmax values of 35 . 6 and 123 Å , respectively , in good agreement with the molecular weight calculated from protein sequence ( 89 . 6 kDa ) and with Rg and Dmax values calculated from the crystal structure using CRYSOL ( 35 . 2 and 125 Å , respectively ) ( Table 1 ) . The distance distribution function P ( r ) for the CYT-18 NTDs displays a single peak with a tail ( Figure 2B ) , a pattern indicative of a protein having an elongated structure [44] . Ab initio models of the CYT-18 NTDs protein were built from the SAXS data by simulated annealing of either dummy atoms by DAMMIN or a chain-like ensemble of dummy residues by GASBOR ( Figures 2C and S3 , respectively ) [45] , [46] . The DAMMIN and GASBOR models show good fits to the experimental SAXS curve ( χ = 1 . 8 for both models ) and are similar in shape to each other and to the high-resolution structure as shown by the superposition of the crystal structure within the SAXS model envelopes . The final DAMMIN and GASBOR models are the result of analyzing multiple solutions and either averaging the models ( DAMMIN ) or picking the most representative one ( GASBOR ) . The normalized spatial discrepancy ( NSD ) value , which describes the similarity between the different models produced by the programs , is low for both the DAMMIN and GASBOR models ( 0 . 63±0 . 03 and 1 . 10±0 . 02 , respectively ) , indicating that the multiple solutions built by the programs are similar to each other ( Table 2 ) . Taken together , these results indicate that the conformation adopted by CYT-18 NTDs in solution is similar to that in the crystal structure [33] . The SAXS data for the isolated CTD overlays well with a theoretical scattering curve calculated from a homology model of CYT-18's CTD constructed from the NMR structure of the A . nidulan CTD using I-TASSER ( χ = 2 . 8 ) ( Figure 2A , middle curve ) [47] . The molecular weight of 13 . 3 kDa estimated from the scattering data ( Table 1 ) indicates that the CTD is monomeric in solution . The Rg and Dmax values for the CTD calculated from the SAXS data ( 17 . 7 and 62 Å , respectively ) are in good agreement with those for the I-TASSER model ( 17 . 0 and 55 . 9 Å , respectively ) ( Table 1 ) . The DAMMIN and GASBOR models of the CYT-18 CTD ( χ = 1 . 7 and 2 . 2 , respectively ) also superpose well with the homology model ( Figures 2D and Figure S3 , respectively ) . Thus , the SAXS analysis indicates that the CYT-18 CTD folds independently of the remainder of the protein and that the I-TASSER model provides a good representation of the structure of the CYT-18 CTD in solution . The latter finding validates the use of the I-TASSER model in building high-resolution structures of CYT-18* from the SAXS data ( see below ) . CYT-18* is the first CYT-18 protein to be investigated structurally that contains both the NTDs and CTD . The molecular weight for this protein estimated from the SAXS data is 119 kDa , confirming that CYT-18* is a dimer in solution ( Table 1 ) . The Rg and Dmax values from the CYT-18* scattering data are 46 . 9 and 170 Å , respectively , both larger than that for the CYT-18 NTDs , as expected . Ab initio models of the CYT-18* homodimer indicate an open conformation with both CTDs extending outward from the NTDs ( χ = 1 . 5 and 2 . 1 for the DAMMIN and GASBOR models , respectively ) ( Figures 2E and S3 ) . A rigid-body model of CYT-18* was also built by CORAL , which constructs models that fit the SAXS data by combining high-resolution models of individual components , in this case the crystal structure of the CYT-18 NTDs and the I-TASSER model of the CTD ( see above ) , with different conformations of flexible dummy residue linkers [48] . The CORAL model overlays well with the scattering curve ( χ = 1 . 8 ) ( Figure 2A , bottom curve ) and superposes well into the SAXS envelopes of the ab initio models ( Figures 2E and S3 ) . These findings indicate that in the absence of intron RNA , CYT-18* adopts an S-shaped configuration with the two CTDs of the homodimer extending outward in opposite orientations . The co-crystal structure of the CYT-18 NTDs bound to Twort intron RNA indicated that the RNA binds asymmetrically across the dimer interface and that the structure of the NTDs does not change substantially upon binding the intron RNA [22] . To elucidate interacting regions and conformational changes of the CTDs upon binding to the intron RNA , we obtained SAXS data for complexes of both the CYT-18 NTDs and CYT-18* bound to the same Twort group I intron RNA . The experimental scattering curve of the CYT-18 NTDs+Twort RNA complex overlays reasonably well with the scattering curve calculated from the co-crystal structure ( χ = 4 . 4 ) ( Figure 3A , top curve ) , and gave Rg and Dmax values ( 39 . 2 and 137 Å , respectively ) in agreement with those calculated from the co-crystal structure ( 39 . 1 and 134 Å , respectively ) ( Table 1 ) . Likewise , a rigid-body model of the CYT-18 NTDs+Twort RNA complex built using CORAL shows a good fit to the experimental SAXS curve ( χ = 2 . 2 ) ( Figure 3B; top curve ) and is similar in shape to the co-crystal structure ( Figures 3A and 3B , compare insets above the top curves ) . These findings indicate that the CYT-18 NTDs+Twort RNA co-crystal structure is similar to the structure of the complex in solution and can be used as a component for structural modeling of the CYT-18*+Twort RNA complex from the SAXS data . Finally , the scattering data for the Twort RNA complex with CYT-18* , which contains both the NTDs and CTD , gave Rg and Dmax values of 41 . 9 Å and 146 Å , respectively ( Figure 3A , bottom curve; Table 1 ) . The relatively small difference between these values and those for the CYT-18 NTDs+Twort complex ( Rg and Dmax values of 39 . 2 Å and 137 Å , respectively; Table 1 ) suggests that CYT-18* forms a compact complex with the RNA in which the CTDs make a smaller than expected contribution to the overall particle dimensions . This conclusion was supported by ensemble optimization analysis using the program EOM , which generates a large random pool of conformations and picks an optimized ensemble that best fits the scattering data ( see Materials and Methods ) . This optimized ensemble pool displayed a smaller , tighter range of Rg and Dmax values than a random pool of protein-RNA conformations consistent with a compact rigid complex ( Figure S4 ) . A rigid-body model of the CYT-18*+Twort RNA complex built using CORAL indicates that both CTDs are positioned near the Twort intron ( Figure 3C ) . This CORAL model can be compared to a previous model of CYT-18*+Twort built using biochemical data ( Figure 3D ) [37] . While both the CORAL and biochemical models show that both CTDs are located near the intron RNA , the CORAL model better fits to the scattering data than does the biochemical model ( χ = 3 . 2 and 8 . 8 , respectively ) . In both models , the CTD of one subunit of the CYT-18 homodimer is close to and may interact with P2 , P6–P6a , and P8 of the intron RNA , while the CTD of the other subunit is close to and may interact with P4–P5 , and P9 of the intron RNA ( Figure 3C and 3D ) . Considered together , the SAXS analyses indicate that upon binding a group I intron RNA , CYT-18* forms a compact complex in which both CTDs of the CYT-18 homodimer clamp down to interact with opposite ends of the group I intron RNA . To investigate how the RNA-binding properties of CYT-18's CTD enable it to interact with the two structurally distinct ends of a group I intron RNA , we analyzed the interaction of the CYT-18 NTDs and the isolated CTD with various RNAs by equilibrium-binding assays at 25°C and 37°C ( Figures 4 and S5 , respectively ) . The RNAs compared were three group I introns ( the N . crassa mt large ribosomal subunit-ΔORF intron ( Nc mt LSU ) ; the N . crassa NADH dehydrogenase subunit 1-ΔORF intron ( Nc ND1m ) ; and the Twort intron RNA ) ; a group II intron RNA ( Lactococcus lactis Ll . LtrB-ΔORF; Ll . LtrB ) ; and poly ( U ) 30 , which presumably lacks higher-order structure . The binding curves for the CYT-18 NTDs to the Nc mt LSU , Nc ND1m , and Twort group I intron RNAs were best fit by hyperbolic functions with Kds ranging from 200 to 590 nM at 25°C ( Figure 4A–4C ) and 440 to 740 nM at 37°C ( Figure S5A–S5C ) . The Kd values for the Nc ND1m intron are substantially higher than that calculated from previous koff measurements , which assumed that the kon of the construct lacking the CTD is the same as that for the wild-type protein ( 71±24 pM ) [30] . This difference suggests that the CTD might make a major contribution to kon by mediating the initial interaction with intron RNA substrates . At both temperatures , the strongest binding group I intron RNA was the Nc ND1m intron and the weakest was the Nc mt LSU intron , consistent with previous findings that the CTD is required for tight binding and splicing of the Nc mt LSU , but not the Nc ND1m intron [30] . At 25°C , the CYT-18 NTDs bound the Ll . LtrB group II intron RNA with a K1/2 = 240 nM , within the range of Kds for group I intron RNAs , but the binding curve was sigmoidal , with n = 1 . 6 , suggesting cooperative and possibly non-specific binding ( Figure 4D ) , whereas at 37°C , binding of the group II intron RNA was weaker ( Kd = 440 nM ) and the binding curve was hyperbolic ( Figure S5D ) . The CYT-18 NTDs did not bind appreciably to poly ( U ) 30 at either 25°C or 37°C ( Figures 4E and S5E ) . Surprisingly , the isolated CYT-18 CTD bound group I and II intron RNAs and poly ( U ) 30 more strongly than did the CYT-18 NTDs and with similar affinities for all five RNAs tested ( Figure 4 ) . Indeed , the binding curves for the isolated CTD to these radically different RNAs were remarkably similar to each other , each being sigmoidal with K1/2s = 57–64 nM at 25°C and 51–77 nM at 37°C . These sigmoidal binding curves ( i . e . , Hill coefficients ( n ) >1 ) suggest cooperative binding of two or more CTDs to each RNA . The ability of the isolated CTD to bind similarly to group I and II intron RNAs , as well as unstructured poly ( U ) 30 indicates that it is a non-specific RNA binding domain . The finding that the CYT-18 CTD is a non-specific RNA-binding domain led us to wonder whether an ancestral Pezizomycotina mtTyrRS might have initially bound group I intron RNAs non-specifically . To address this question , we turned to the S . cerevisiae ( Sc ) mtTyrRS , which is closely related to CYT-18 but branched from the Pezizomycotina mtTyrRSs prior to the evolution of splicing activity [23] . We compared two constructs that were expressed in E . coli: recombinant wild-type Sc mtTyrRS and a derivative lacking the CTD ( Sc NTDs ) . We confirmed that both proteins are fully active in tyrosyl-adenylation , indicating correct folding of the catalytic domain ( Figure S1A ) . Aminoacylation assays showed that the Sc mtTyrRS has higher activity with E . coli tRNATyr than does CYT-18 ( Figure S1B ) , possibly reflecting that the E . coli tRNATyr and the Sc mt tRNATyr are more similar to each other than to the Nc mt tRNATyr . All three tyrosyl-tRNAs share the same N73 identity element ( A73 ) and anticodon , but differ in the N1-N72 identity element at the end of the acceptor stem ( G-C in E . coli tRNATyr and Sc mt tRNATyr , but A-U in Nc mt tRNATyr ) and the length of the variable arm ( 13–14 nt in E . coli tRNATyr and Sc mt tRNATyr , but unusually long at 16 nt in Nc mt tRNATyr ) ( Figure S6 ) [49]–[51] . Equilibrium binding assays showed that the Sc mtTyrRS , although incapable of splicing group I intron RNAs [23] , can bind both the Nc mt LSU group I intron RNA and Ll . LtrB group II intron RNA with Kds = 430 and 440 nM , respectively ( Figure 5A and 5B ) , within the range of Kds for specific binding of CYT-18 NTDs to group I intron RNAs ( see above ) . However , the similar affinity of the Sc mtTyrRS for the group I and group II intron RNAs suggests that this binding is non-specific . Strikingly , the ability of the Sc mtTyrRS to bind group I and II intron RNAs was entirely dependent upon its CTD , with the Sc NTDs protein showing no detectable binding of either intron RNA over the concentration range tested ( Figure 5A and 5B ) . To investigate if the Sc mtTyrRS CTD is a non-specific RNA-binding domain like CYT-18's CTD , we expressed and purified this domain separately including a small segment of the upstream linker region ( denoted Sc CTD ) . We then assayed equilibrium binding of the Sc CTD to three group I introns ( Nc mt LSU , Nc ND1m and Twort ) , a group II intron ( Ll . LtrB ) , and poly ( U ) 30 at 25°C . These assays showed that the Sc CTD is capable of binding all the RNAs tested with Kd or K1/2 values ranging from 110 nM to 1 µM ( Figure 5C ) . The Sc CTD had the highest affinity for the Ll . LtrB group II intron RNA ( K1/2 = 110 nM ) , which it bound in a cooperative manner ( n = 1 . 8 ) , and the lowest affinity for poly ( U ) 30 ( K1/2 = 1 µM; n = 1 . 6 ) . Notably , for all RNAs tested , the non-specific binding by the isolated Sc mtTyrRS CTD is 1 . 5- to 15-fold weaker than the binding of CYT-18's CTD to the same RNA ( K1/2s at 25°C = 57–64 nM ) ( Figure 4 ) . Together , these findings indicate that the CTD of the Sc mtTyrRS is also a non-specific RNA-binding domain and that the Sc mtTyrRS binds both group I and group II intron RNAs non-specifically via its CTD . Finally , we investigated whether the Sc CTD could replace the CYT-18 CTD to promote group I intron splicing by making CYT-18/Sc mtTyrRS chimeric proteins . Two chimeric constructs were made differing in whether they contain the flexible linker region from CYT-18 or the Sc mtTyrRS ( Figure 6 ) . Chimera 1 consists of the CYT-18 catalytic and α-helical domains fused to the Sc CTD via the Sc mtTyrRS linker region , whereas chimera 2 contains the same CYT-18 NTDs fused to the Sc CTD via the CYT-18 linker , which includes Ins 3 . Both chimeric proteins showed tyrosyl-adenylation activity similar to wild-type CYT-18 , but displayed differences in aminoacylation activity ( Figure 7A and 7B ) . Chimera 1 , which contains both the Sc mtTyrRS linker and CTD , is similar to the yeast Sc mtTyrRS in having higher aminoacylation activity with E . coli tRNATyr than does CYT-18 , likely due to its better recognition of the bacterial tRNA ( see above ) . By contrast , chimera 2 , which contains the CYT-18 linker followed by the yeast mtTyrRS CTD , has substantially lower aminoacylation activity than CYT-18 , indicating that the CYT-18 linker impairs charging of E . coli tRNATyr . The findings for chimera 1 indicate that higher TyrRS activity with E . coli tRNATyr correlates with presence of the yeast CTD and linker , regions of the TyrRS that recognize the tRNA variable arm and anticodon stem , and not with the catalytic domain , which recognizes the acceptor stem [26] , [28] , [52] . The Sc mtTyrRS linker may contribute to the recognition of E . coli tRNATyr , either by contacting the tRNA directly or by facilitating binding of the CTD to the variable arm and/or anticodon stem . To determine whether the Sc mtTyrRS CTD could function in splicing , we compared the ability of the chimeric proteins to splice the Nc mt LSU and ND1m group I introns , which do and do not require the CTD for splicing , respectively [30] . Group I introns splice via two sequential transesterification reactions initiated by the addition of guanosine nucleotide to the 5′ end of the intron , resulting in ligated exons and excised linear intron RNA with a non-coded G residue at its 5′ end [53] . The ability of the chimeric proteins to splice the Nc mt LSU and ND1m group I introns was assayed by using 200 nM 32P-labeled precursor RNA containing the introns , 100 nM protein , and unlabeled GTP at three different temperatures ( 25°C , 30°C , and 37°C ) at either 100 mM KCl ( the standard condition for CYT-18 ) or a lower salt concentration , 25 mM KCl ( Figures 7C–7F , S7A , and S7B ) . The assays showed that the chimeric proteins could splice the Nc ND1m intron , which does not require the CTD , but could not splice the Nc mt LSU intron , which requires the CTD , under all conditions examined . The inability of the chimeric proteins to splice the Nc mt LSU intron was additionally confirmed by splicing assays done with higher protein concentration ( 500 nM; protein excess conditions ) to compensate for potentially weaker binding of the intron RNA by the Sc CTD ( Figure S7C ) , and by using a more sensitive assay in which [α-32P]GTP is incubated with unlabeled precursor RNA to label the 5′ end of the intron RNA during the first step of splicing ( Figures 7G and S7D ) . Notably , although the chimeric proteins were capable of splicing the ND1m intron ( Figures 7D and S7B ) , they did so at a slower rate than full-length CYT-18 , with the rate decreasing further at higher salt conditions ( Figure 7E and 7F ) . The slower rate of ND1m intron splicing by the chimeric proteins is similar to that found previously for the CYT-18 NTDs alone [30] , suggesting that it reflects a lack of contributing but nonessential interactions with the CTD . Interestingly , chimera 2 has higher splicing activity with the ND1m intron than does chimera 1 , the reverse of what was found for TyrRS activity ( Figure 7B ) . This finding suggests that the longer CYT-18-linker region , which is present in chimera 2 but not chimera 1 , contributes to splicing activity . This contribution could involve either specific or non-specific interaction of Ins 3 with the group I intron RNA or increased conformational flexibility of the CTD due to expansion of the linker . Considered together , the findings for chimera 1 and chimera 2 indicate that although both the Sc mtTyrRS and CYT-18 CTDs bind group I intron RNAs non-specifically , the Sc CTD lacks further adaptations required for group I intron splicing activity .
Our results provide insight into the function of CYT-18's CTD and its contribution to the evolution of group I intron splicing activity , highlighting a role for non-specific binding interactions in the evolution of new RNA-binding functions . First , the SAXS analysis indicates that the CTDs of both subunits of the CYT-18 homodimer have a preferred orientation in solution extending outward in opposite directions from the NTDs , but move inward to bind opposite ends of a group I intron RNA . The CORAL model of CYT-18* bound to Twort intron RNA based on the SAXS data suggests that the CTD of one subunit binds the intron near P2 , P6–P6a , and P8 , while the CTD of the other subunit likely interacts with P4–P5 and P9 ( Figure 3C ) . These interaction sites agree with a previous biochemical model based on directed hydroxyl-radical cleavage assays in which Fe-EPD with a cleavage radius of 25 Å was conjugated at two sites in the CTD ( G493C and C494 ) [37] . These assays found cleavage sites in P6–P6a , P3–P8 , and P5 in the Nc ND1m intron and P2 , P4 , and P6–P6a in the Nc mt LSU intron [37] , [54] . The additional cleavages in the Nc mt LSU intron P2 helix , which is considerably longer than P2 of the Twort or Nc ND1 introns , are consistent with its proximity to P8 . The putative interaction sites between the CTDs and intron RNA in our CORAL model are also consistent with genetic experiments showing that CTD binding can suppress intron RNA mutations that impair long-range tertiary interactions P5-L9 and P2-L8 on opposite ends of the intron RNA [31] . The relatively fixed orientation of the CTDs in the free CYT-18 protein agrees with previous 15N-1H- two-dimensional NMR analysis showing that the CTDs of the full-length A . nidulans mt and Geobacillus stearothermophilus TyrRSs do not tumble independently in solution [37] . Nevertheless , the linker must be sufficiently flexible to allow the CTDs to bind to group I introns or tRNATyr on opposite sides of the catalytic domain , and the SAXS analysis provides the first direct evidence for this conformational flexibility by showing the two CTDs of the homodimer swing downward from their starting position in the free protein to interact with different regions of a group I intron RNA . We were surprised to find that the CTDs of both CYT-18 and the non-splicing yeast mtTyrRS are non-specific RNA-binding domains . The isolated CTDs of both proteins bind structured group I and group II intron RNAs , or the simple homopolymer , poly ( U ) 30 with similar affinities , with this non-specific binding 1 . 5- to 15-fold stronger for the CYT-18 CTD than the yeast mtTyrRS CTD ( see Results ) . The non-specific RNA-binding activity of the TyrRS CTDs may contribute to its function in aminoacylation by augmenting its specific-binding interactions with tRNATyr , which include recognition of the variable arm and anticodon bases [28] , [55] . Likewise , the high non-specific binding activity of the CYT-18 CTD does not preclude and may bolster specific-binding interactions of this domain with group I intron RNAs . The latter could result either from further adaptive evolution of the CTD or simply from positioning of the CTD on the intron RNA via specific binding of the NTDs . Although non-specific RNA binding was unexpected for an aaRS domain involved in tRNA recognition , yeast and higher eukaryotic aaRSs have been shown previously to have appended non-specific RNA-binding domains that are not present in their bacterial counterparts and contribute to aminoacylation efficiency . Thus , the yeast glutaminyl-tRNA synthetase ( GlnRS ) has an N-terminal non-specific RNA binding domain , which when fused to a bacterial GlnRS enabled it to functionally replace the yeast enzyme in vivo , as did fusion of the yeast Arc1 protein , a non-specific RNA-binding protein that ordinarily helps mediate tRNA/aaRS interactions in trans [56] , [57] . Similarly , some higher eukaryotic aaRSs have tandem repeats of a small non-specific RNA-binding motif that enhances tRNA binding [58] . These non-specific RNA-binding domains are thought to act by adding sufficient binding energy to compensate for relatively weak specific binding interactions of aaRSs with tRNA substrates , similar to the augmentation of specific binding of tRNA and intron RNA substrates suggested above for the TyrRS CTD . Notably , the ribosomal protein S4-like fold , which forms the core of bacterial and mitochondrial TyrRS CTDs , has been identified previously as an ancient RNA-binding domain . This domain is found in all three kingdoms of life in a variety of proteins that bind structurally different RNAs , including two families of pseudouridine synthetases , a family of predicted RNA methylases , an RNA-modification enzyme with both pseudouridine synthetase and cytidine deaminase activity , threonyl-tRNA synthetases , and a heat-shock protein [59]–[61] . The S4-like fold consists of two α-helices arranged as a helical hairpin packed against three or four β-sheets . Connecting two of the β-sheets is a characteristic L-shaped loop , which together with the two α-helices is termed the αL motif . This motif generally contains clusters of basic and polar residues that are capable of interacting with various nucleic acid substrates in the different S4-like fold containing proteins . In TyrRSs , the αL motif interacts in a region between the variable and anticodon arms [61] , [62] . We suggest that the inherently high non-specific RNA-binding affinity of the S4-like fold was the key factor enabling it to evolve interactions with different RNA substrates in the course of evolution . Indeed , the fungal mtTyrRSs provide a dramatic example of a case in which the S4-like fold of a single enzyme may bolster specific-binding interactions with three different regions of two different RNA substrates , a mt tRNATyr and a group I intron RNA . Although we suggest that the non-specific binding of the CTD played a key role in initial interaction with group I intron RNAs , the CTDs of present-day fungal mtTyrRS appear to have evolved specific interactions with group I intron RNAs . Thus chimeric proteins containing the CYT-18 NTDs linked to the yeast CTD can efficiently aminoacylate E . coli tRNATyr , as well as splice the Nc ND1 intron , which requires only the NTDs [30] . However , the chimeric proteins splice the Nc ND1m intron less efficiently than full-length CYT-18 at a rate expected for loss of contributing CTD interactions , and they are unable to splice the Nc mt LSU intron , which requires the CTD [30] . Additional adaptations of the CYT-18 CTD required to promote splicing may include RNA-binding contacts by Ins 3–5 , which are found in the CTDs of splicing-competent Pezizomycotina mtTyrRS , but not in the Sc mtTyrRS [37] . Both the previous biochemical model of CYT-18*+Twort [37] and the new CORAL model based on the SAXS data ( Figure 3C ) place Ins 4 and 5 in position to bind group I intron RNAs . Since the discovery of the splicing function of CYT-18 [17] , there have been numerous additional examples of aaRSs that have acquired new functions unrelated to translation , in most cases via addition of non-catalytic domains [63] , [64] . The acquisition of these new domains and functions is thought to reflect that aaRS are ancient essential enzymes whose presence early in evolution of the cell provided a robust scaffold for the addition of new structural elements [65] . In archael and eukaryotic TyrRSs , the N-terminal catalytic domain is followed by a different anticodon-binding domain , known as the C-W/Y domain , which is homologous to the anticodon-binding domain of TrpRSs [66] . Two additional structural elements were acquired during the evolution of higher eukaryotes and function in receptor-mediated signaling pathways associated with angiogenesis: the ELR motif in the catalytic domain and a C-terminal EMAP II-like domain , which has non-specific RNA-binding properties [63] , [67]–[69] . The ELR motif is on the intron-binding side of the catalytic domain [70] and incorporated in the same α-helix as Ins1 in the fungal mtTyrRS , suggesting that this region may be a particularly robust location for insertion of new functional elements . Finally , our results provide evidence that non-specific binding can play a key and perhaps widespread role in pre-adaptive interactions that lead to the evolution of new RNA-binding functions of proteins . For the group I intron splicing activity of fungal mtTyrRSs , our findings suggest a scenario outlined in Figure 8 in which an initial non-specific interaction between the CTD of an ancestral mtTyrRS and a group I intron RNA was fixed by an intron RNA mutation that made formation of active ribozyme structure dependent upon interaction with the protein . After the interaction was fixed , the mtTyrRS and group I intron were forced to co-evolve , with further adaptive mutations in the protein leading to specific binding of both the catalytic domain and CTD to the intron RNA . These specific-binding interactions extended the intron RNA-binding surface , both increasing the efficiency of splicing and permitting additional mutations in the intron RNA that made it more dependent upon the protein for structural stabilization . RNA-editing enzymes such as APOBEC1 , which evolved from enzymes that acted on mononucleotide substrates , may be additional examples of constructive neutral evolution in which a relatively non-specific pre-adaptive interaction with an RNA substrate was fixed by a deleterious mutation , in this case one that could be corrected by RNA editing , and then elaborated by further adaptive mutations [71] , [72] . Indeed , a similar evolutionary pathway may have been used more generally for other RNA-modification enzymes , including the ones mentioned above that contain an S4-like non-specific RNA-binding domain . Beyond the initial pre-adaptive phase , the extensive structural data for the interaction of fungal mtTyrRSs with group I intron RNAs provide strong evidence for a ratchet-like process in which multiple adaptive mutations , including six different Peziomycotina-specific insertions , led to the evolution of an efficient splicing apparatus for group I introns . It is highly unlikely that the multiple adaptive mutations in the protein leading to an extensive group I intron-binding surface occurred in one step . The surprising finding that the structural adaptations of the mtTyrRS catalytic domain utilized a non-tRNA-binding surface could reflect that the tRNA-binding site in the catalytic domain could not be easily modified to function in group I intron splicing without inhibiting mtTyrRS activity , which is essential in an obligate aerobe . Additionally , the non-tRNA-binding side of the catalytic domain may have had a pre-existing auxiliary RNA-binding function , as found for some aaRSs [73] , [74] . By contrast to the catalytic domain , the regions of the CTD needed for splicing activity overlap tRNA-binding regions requiring co-evolution with both the intron RNA and mt tRNATyr . Indeed , the unusually long variable arm of Pezizomycotina mt tRNATyrs ( see Results ) ( Figure S6 ) may be an example of a feature that co-evolved with the CTD to allow it to better accommodate group I intron RNAs [37] . We also note that although the initial interaction of an ancestral fungal mtTyrRS likely involved a single group I intron RNA , perhaps the mt LSU intron , which is dependent upon the mtTyrRS for splicing in all Pezizomycotina fungi examined [23] , the fungal mtTyrRSs ultimately evolved to function in splicing multiple group I introns by recognizing the conserved phosphodiester backbone structure of the catalytic core . This binding mode has the evolutionary advantages of enabling the fungal mtTyrRSs to coordinate the splicing of multiple group I introns as well as the ability to accommodate new group I introns that invade genomes as mobile genetic elements .
Recombinant plasmids used for protein expression in E . coli are derivatives of the phage T7 promoter-driven expression vectors pET3a , pET11a , or pET11d ( EMD Millipore ) . pEX560 , which expresses a wild-type CYT-18 protein ( amino acids 33–669 ) , contains the cyt-18 ORF ( nucleotides 97–2 , 010 ) cloned downstream of the T7 promoter in pET3a [29] . pCYT18/ΔC-tail , which expresses CYT-18* ( C-terminal truncation of the non-essential C-tail; amino acids 584–669 ) , was derived from pEX560 by introducing three stop codons ( TAATAGTAG ) after Leu583 by site-directed mutagenesis ( QuikChange; Agilent Technologies ) . pHISTEV602 expresses the CYT-18 NTDs ( C-terminal truncation of both the CTD and C-tail; amino acids 424–669 ) , with an N-terminal tobacco etch virus ( TEV ) protease-cleavable 6× His-tag . It was constructed by PCR of pEX560 using primers that amplify nucleotides 97–1 , 251 of the CYT-18 ORF and append NcoI and BamHI sites , and then cloning the resulting PCR product between the NcoI and BamHI sites of pET11d . pCYT18-CTD , which expresses the CYT-18 CTD ( amino acids 448–583 ) with an N-terminal TEV-cleavable 6× HIS-tag , was constructed by PCR of pEX560 using primers that amplify nucleotides 1 , 342–1 , 749 of the CYT-18 ORF and append NdeI and BamHI sites , and then cloning the resulting PCR product between the NdeI and BamHI sites of pET11a . All CYT-18 expression constructs lack the mt targeting sequence ( amino acids 1–32 ) . Wild-type CYT-18 and CYT-18* have an extra N-terminal methionine , while CYT-18 NTDs and CTD have an extra N-terminal glycine resulting from TEV-protease cleavage of the N-terminal 6× His-tag . pHISTEVScTyrRS , which expresses the full-length mature S . cerevisiae mtTyrRS with an N-terminal TEV-cleavable 6× HIS-tag , contains Sc mtTyrRS codons 38–492 ( lacking the mt target sequence; amino acids 1–37 ) cloned between the Nco1 and BamHI sites of pET11d [23] . pHISTEVSc/ΔCTD expresses Sc mtTyrRS lacking the CTD ( denoted Sc NTDs ) and was derived from pHISTEVScTyrRS by using site-directed mutagenesis to add three stop codons ( TAATAATAA ) after Asp400 . pMAL-ScCTD , which expresses the Sc mtTyrRS CTD ( denoted Sc CTD ) , contains Sc CTD codons 414–492 cloned between the BamHI and HindIII sites of pMAL-c2t [75] , a derivative of plasmid pMAL-c2x ( New England Biolabs ) that expresses the protein with an N-terminal maltose-binding protein tag followed by a TEV-protease site . Chimeric proteins containing the N-terminal catalytic domain of CYT-18 and the CTD of the Sc mtTyrRS were made by overlap PCR . Chimera 1 contains the CYT-18 NTDs ( amino acids 33–417 ) fused to the Sc mtTyrRS flexible linker and CTD ( amino acids 397–492 ) . Chimera 2 contains the CYT-18 NTDs and linker including Ins 3 ( amino acids 33–451 ) fused to the Sc mtTyrRS CTD ( amino acids 416–492 ) . The chimeric protein ORFs were cloned between the BamHI and HindIII sites of pMAL-c2t ( see above ) , enabling the expression of fusion proteins with an N-terminal TEV-protease cleavable maltose-binding protein tag . Recombinant plasmids used for in vitro transcription contain group I or II introns cloned downstream of a phage T3 or T7 promoter . pBD5a contains the N . crassa mt large subunit rRNA-ΔORF ( Nc mt LSU ) intron cloned downstream of a T3 promoter in pBS ( + ) [19] . Transcription of pBD5a linearized with BanI yields a 503-nt RNA containing a 65-nt 5′ exon , the 388-nt mt LSU intron , and a 50-nt 3′-exon . pND1m contains the N . crassa NADH dehydrogenase subunit 1-ΔORF ( Nc ND1m ) intron cloned downstream of a T7 promoter in pUC18 [18] . Transcription of pND1m linearized with NdeI yields a 209-nt RNA containing a 6-nt 5′ exon , the 196-nt ND1 intron , and a 7-nt 3′ exon . pTWORT-P2 contains a ribozyme derivative of a group I intron of the Staphylococcus aureus bacteriophage Twort orf142 gene ( intron nucleotides 9-250 ) cloned downstream of a T7 promoter in pUC19 [76] . Transcription of pTWORT-P2 linearized with EarI yields a 242-nt transcript of the Twort ribozyme . pSSltrBΔA contains a derivative of the L . lactis Ll . LtrB-ΔORF intron with a deletion of the branch-point nucleotide to prevent splicing during binding assays cloned downstream of a T7 promoter in pUC19 [77] . Transcription of a DNA template made by PCR of the pSSltrBΔA plasmid ( forward primer 5′-ATGAATTCTAATACGACTCACTATAGGGTTATAATTATCCTTACACATCCATAAC and reverse primer 5′-CGCTGCAGAATTGATATCAAAAATGATATG ) yields an 807-nt RNA containing a 28-nt 5′ exon , the 749-nt intron , and a 30-nt 3′ exon . Proteins were expressed from the recombinant plasmids indicated above in E . coli HMS174 ( DE3 ) ( CYT-18 , CYT-18* , and CYT-18 NTDs ) ; BL21 ( DE3 ) ( CYT-18 CTD , chimera 1 , and chimera 2 ) ; or Rosetta 2 ( DE3 ) ( EMD Millipore ) ( Sc mtTyrRS , Sc NTDs , and Sc CTD ) . Overnight cultures of fresh transformants were inoculated into LB media , and the proteins expressed via auto-induction [78] . Cells expressing CYT-18 and CYT-18* were grown at 35°C overnight with shaking at 260 rpm . Cells expressing all other proteins were grown at 37°C for 4 h then shifted to 25°C overnight with shaking at 260 rpm . Wild-type CYT-18 and CYT-18* were purified as described [22] , [27] . Briefly , cells were lysed by incubation with lysozyme at 1 mg/ml for 30 min followed by polyethyleneimine precipitation to remove nucleic acids , and ammonium sulfate precipitation [27] . The ammonium sulfate pellet was dissolved in 500 mM KCl , 25 mM Tris-HCl ( pH 7 . 5 ) and then dialyzed overnight in 25 mM KCl , 25 mM Tris-HCl ( pH 7 . 5 ) . The protein was purified from the dialysate by using a HiTrap SP XL cation exchange column ( GE Healthcare Life Sciences ) , followed by a size-exclusion column ( HiLoad 16/60 Superdex 200; GE Healthcare Life Sciences ) [22] . The 6× HIS-tagged proteins CYT-18 NTDs , CYT-18 CTD , Sc mtTyrRS , and Sc NTDs were purified similarly , except that the ammonium sulfate pellet was dissolved in 500 mM KCl , 25 mM Tris-HCl ( pH 7 . 5 ) , and 30 mM imidazole , and the proteins were purified by nickel-affinity chromatography using a HisTrap HP column ( GE Healthcare Life Sciences ) [23] , followed by TEV protease-cleavage of the 6× HIS-tag in dialysis buffer ( 500 mM KCl , 25 mM Tris-HCl [pH 7 . 5] , 5 mM DTT ) to remove imidazole . The proteins were then further purified by an additional round of nickel-affinity chromatography , followed by size-exclusion chromatography ( HiLoad 16/60 Superdex 200; GE Healthcare Life Sciences ) . The maltose-binding protein ( MalE ) fusions MalE-ScCTD , MalE-chimera 1 , and MalE-chimera 2 were purified by polyethyleneimine precipitation of nucleic acids , as described above for CYT-18 , and then loaded onto an amylose affinity column ( New England Biolabs ) in buffer containing 25 mM Tris-HCl ( pH 7 . 5 ) , 500 mM KCl , 1 mM DTT , 1 mM EDTA , and 10% glycerol followed by elution with 10 mM maltose in the same buffer . The proteins were further purified using a heparin-sepharose column ( HiTrap heparin HP column; GE Healthcare Life Sciences ) in 300 mM KCl , 25 mM Tris-HCl ( pH 7 . 5 ) , 1 mM DTT , and 1 mM EDTA and eluted with a salt gradient of 300 mM to 1 . 5 M KCl in the same buffer . The final purification step was size-exclusion chromatography ( HiLoad 16/60 Superdex 200; GE Healthcare Life Sciences ) in 25 mM Tris-HCl ( pH 7 . 5 ) , 200 mM KCl , and 10% glycerol . Proteins used for SAXS were stored in buffer containing 100 mM KCl , 5 mM MgCl2 , 10 mM Tris-HCl ( pH 7 . 5 ) , 5% glycerol at −80°C . Proteins used for biochemical assays were dialyzed into 100 mM KCl , 25 mM Tris-HCl ( pH 7 . 5 ) , and 50% glycerol and stored at −80°C . Protein yields ranged from 14 to 44 mg/l ( monomer concentrations ) , and all proteins were >99% pure as judged by SDS-polyacrylamide gels stained with Coomassie blue . Protein concentrations were determined by measuring A280 under denaturing conditions ( 6 M guanidine hydrochloride ) . Concentrations of wild-type CYT-18 , the Sc mtTyrRS , and C-terminal truncations of these proteins refer to the homodimer , while CTD concentrations refer to the monomer . Intron-containing RNA substrates for SAXS and biochemical assays were transcribed from the linearized recombinant plasmids indicated above . The Twort intron for SAXS was synthesized by large-scale in vitro transcription reactions ( 10–30 ml ) with T7 polymerase at 37°C in reaction buffer containing 40 mM Tris-HCl ( pH 8 . 1 ) , 1 mM spermidine , 10 mM DTT , 8 mM NTPs , and 15 mM MgCl2 . Transcription reactions were incubated at 37°C for 8 h and terminated by adding 50 mM EDTA followed by extraction with phenol-chloroform-isoamyl alcohol ( 25∶24∶1; phenol-CIA ) . The RNA was then purified through a 5-ml HiTrap desalting column ( GE Healthcare Life Sciences ) and a size exclusion column ( HiLoad 16/60 Superdex 200; GE Healthcare Life Sciences ) . T7 RNA polymerase for the large-scale transcriptions was expressed with an N-terminal 6× HIS-tag from pRC9 and purified as described [79] . 32P-labeled Nc mt LSU , Nc ND1m , and Twort RNAs for equilibrium-binding assays were synthesized by using a MAXIscript transcription kit ( Life Technologies ) , with the concentration of unlabeled UTP changed from that recommended in the manufacturer's protocol ( 0 . 5 mM ) to 10 µM UTP to obtain higher specific activity transcripts ( 300 Ci/mmol ) . The Ll . LtrB group II intron was synthesized by using a mutant T7 polymerase that can read through a T7 polymerase transcription termination site within the intron [80] in reaction medium containing 40 mM Tris-HCl ( pH 7 . 9 ) , 6 mM MgCl2 , 10 mM DTT , 2 mM spermidine , 1 mM GTP , 1 mM CTP , 1 mM ATP , 250 nM UTP , and 1 µM [α-32P]UTP ( 3 , 000 Ci mmol−1; Perkin Elmer ) . After transcription and DNase treatments ( MAXIscript transcription kit; Life Technologies ) , transcripts were purified by extraction with phenol-CIA , followed by gel filtration through two consecutive 1-ml Sephadex G-50 columns ( Sigma-Aldrich ) . Intron-containing RNA substrates for splicing reactions were transcribed from linearized DNA template using a MEGAscript transcription kit ( Life Technologies ) with 1 µCi [α-32P]UTP ( 3 , 000 Ci mmol−1; PerkinElmer ) added for standard 32P-labeled substrates and 3 µCi [α-32P]UTP ( 3 , 000 Ci mmol−1 ) added for higher specific activity subtrates ( Figures 7C , S7A , and S7C ) . The Nc mt LSU intron substrate was synthesized by in vitro transcription of pBD5a ( BanI digested ) using a MEGAscript T3 kit , while the Nc ND1m intron substrate was synthesized by in vitro transcription of pND1m ( NdeI digested ) using a MEGAscript T7 kit . The intron RNAs were purified as described above . The poly ( U ) 30 oligonucleotide used for binding assays was synthesized and HPLC-purified by Integrated DNA Technologies . The oligonucleotide was dissolved in 10 mM HEPES ( pH 7 . 5 ) , 1 mM EDTA and stored at a concentration of 25 µM . For equilibrium-binding assays , 25 pmoles of the oligonucleotide was 5′-end labeled with [γ-32P]ATP ( 3 , 000 Ci mmol−1; PerkinElmer ) using T4 kinase ( New England Biolabs ) and then purified by phenol-CIA extraction followed by desalting through a Sephadex G-25 column . Proteins and RNAs used for SAXS analysis were prepared as described above . RNA-protein complexes were formed by mixing protein dimer and RNA at a 1∶1 molar ratio in 1 . 2 ml of 100 mM KCl , 5 mM MgCl2 , 10 mM Tris-HCl ( pH 7 . 5 ) , and 5% glycerol . After incubation at room temperature for 15 min , RNP complexes were purified by size-exclusion chromatography ( Hi-Load 16/60 Superdex 200 column; GE Healthcare Life Sciences ) in the same buffer . RNP complexes and proteins for SAXS were concentrated by using Amicon Ultra-4 centrifugal filter units ( EMD Millipore ) and frozen for storage at −80°C . The size-exclusion chromatography column buffer was used as a solvent blank for SAXS . SAXS data were collected on beamline 12-ID-C at the Advanced Photon Source ( Argonne , Illinois ) . Each sample had 20 1-s exposures taken at a sample-to-detector distance of 2 . 0 m , covering a momentum transfer range of 0 . 007<q<0 . 35 Å−1 . Samples were continuously passed through the beam using a flow-cell to minimize radiation damage . The 20 consecutive exposures were compared and showed no change in scattering intensity , indicating no radiation damage . Radially averaged scattering data were buffer subtracted and analyzed by using ATSAS [42] and IGOR-Pro ( WaveMetrics ) . Scattering curves were displayed as the scattering intensity ( I ( q ) ) as a function of momentum transfer q = ( 4πsinθ ) /λ , where λ is the wavelength of the incident X-ray beam and θ is half the angle between the incident and scattering radiation . SAXS data were obtained for least three different concentrations of each protein and checked for aggregation and interparticle interference by examination of the Guinier region [81] . Guinier plots ( log ( I ( q ) ) versus log ( q ) ) were checked for linearity in the Guinier region , a diagnostic of sample quality . For globular proteins , the Guinier approximation is valid for qRg<1 . 3 . The q range used for SAXS analysis was 0 . 015<q<0 . 3 Å−1 for CYT-18 protein constructs and 0 . 02<q<0 . 3 Å−1 for CYT-18+Twort complexes . The I ( 0 ) ( extrapolated forward scattering at zero angle ) and Rg ( radius of gyration ) were evaluated using the Guinier approximation for scattering intensity ( I ( q ) ) according to the equation:I ( 0 ) and Rg were also computed from the scattering curve by using the indirect Fourier transform program AUTOGNOM , which additionally provides an estimate of the maximum particle dimension ( Dmax ) from the distance distribution function P ( r ) [42] . The Rg values determined by using the Guinier approximation were consistent with those determined by AUTOGNOM . Molecular weights were calculated by comparing the extrapolated forward scattering at zero angle , I ( 0 ) , with that of a protein standard , bovine serum albumin ( BSA ) , by using the equation:where MMp and MMst are the molecular weights of the protein sample and protein standard , respectively , cp and cst are their concentrations in g/l , and I ( 0 ) p and I ( 0 ) st are the forward scattering intensities of the protein and standard , respectively . Agreement with the calculated molecular weights of the samples indicates sample quality and monodispersity [81] . Experimental scattering curves were compared with theoretical scattering curves calculated by the program CRYSOL ( for qmax<0 . 3 ) from the crystal structures of those macromolecules with known atomic structures ( CYT-18 NTDs , CYT-18 NTDs+Twort , CYT-18 CTD homology model ) [43] . Ab initio shape reconstructions were done by using DAMMIN ( for qmax<8/Rg ) and GASBOR , which use simulated annealing methods to build low resolution protein models from dummy atoms or residues , respectively [45] , [46] . The program DAMMIN uses dummy atoms packed into a sphere with the beads determined to be either protein or solvent . The final DAMMIN model was obtained by using the DAMAVER program suite to align ten models from independent DAMMIN runs and produce an average model . The latter was further refined by using DAMMIN to produce the final model [82] . GASBOR represents the protein as a chain-like ensemble of dummy residues equal to the number of residues in the protein . The final GASBOR model was chosen as the one with the lowest NSD value after running DAMSEL to compare ten models from independent GASBOR runs [82] . No symmetry was specified for the building of CYT-18* or CYT-18 CTD ab initio models , while P2 symmetry was specified for the CYT-18 NTDs models based on prior knowledge from the CYT-18 NTDs crystal structure . DAMMIN and GASBOR produced similar models of CYT-18* with or without P2 symmetry enforced . Rigid-body models of CYT-18* by itself and of CYT-18* and the CYT-18 NTDs bound to Twort RNA were built by using the program CORAL [48] . This program employs a simulated annealing method to place high resolution models of individual components in orientations that minimize the discrepancy between the calculated SAXS profile and the experimental SAXS data , with distances between the structured components constrained by randomized dummy residue linkers chosen from a generated library of non-clashing loop structures . To build models of CYT-18* , a homology model of the CYT-18 CTD ( amino acids 448–583 ) was generated by I-TASSER [47] , using the A . nidulans CTD NMR structure ( PDB:2KTL ) as a template for modeling [37] . The confidence ( C-score ) and TM-scores of the CYT-18 CTD homology model , which are indicators of model quality , are high at 0 . 98 and 0 . 85 , respectively . This CYT-18 CTD model and available high-resolution crystal structures for CYT-18 NTDs+Twort RNA ( PDB:2RKJ ) were used for rigid-body modeling . The final CORAL models were chosen from among ten independently derived models based on the best fit to the experimental scattering data as indicated by a low χ value [48] . Ensemble optimization analysis to characterize the flexibility of CYT-18*+Twort system was conducted by using the program , EOM [83] , [84] . This program generates a random pool of 10 , 000 structures and creates an optimized ensemble from this pool , such that the average scattering pattern of the ensemble fits the experimental SAXS data . Comparison of the shape of the Rg and Dmax distributions of the optimized ensemble with those of the random pool provides information about the size and flexibility of the structure , with a broad peak resembling that of the random pool suggesting a flexible , extended structure and a peak narrower than the random pool suggesting a more rigid structure . Tyrosyl-adenylation assays were done by incubating 100 nM protein in a 50-µl reaction containing 5 mM ATP , 100 mM KCl , 10 mM MgCl2 , 144 mM Tris-HCl ( pH 7 . 5 ) , 2 mM DTT , 0 . 1 mg/ml BSA ( New England Biolabs ) , 0 . 1 unit of yeast inorganic phosphatase ( New England Biolabs ) , and 5 µCi of L-[3 , 5-3H]-tyrosine ( 53 Ci mmol−1; Amersham Biosciences Corp . ) [30] . Reactions were initiated by adding protein and incubated at 30°C for 10 min . Reactions were terminated by adding 1 ml of reaction medium and immediately filtering through a nitrocellulose membrane to trap protein bound tyrosyl-adenylate . Radioactivity was measured by Beckman Coulter LS 6500 scintillation counter using Ready Protein scintillation cocktail ( Beckman ) . Aminoacylation assays were done as described previously with protein concentrations normalized to tyrosyl-adenylation activity [23] . Reactions of 120 µl contained 100 nM protein and 6 µM E . coli tRNATyr ( Sigma-Aldrich ) in 100 mM KCl , 15 mM MgCl2 , 50 mM Tris-HCl ( pH 7 . 5 ) , 5 mM ATP , and 10 mM L-tyrosine ( a 1∶10 mixture of L-[3 , 5-3H]-tyrosine and unlabeled L-tyrosine ) . Reactions were initiated by adding protein and incubated at 30°C . For time courses , 20-µl portions were removed after times ranging from 2 to 60 min , and the reaction was terminated by precipitation with 0 . 8 ml of a solution containing 10% trichloroacetic acid and 20 mM sodium pyrophosphate . Reactions were filtered through Whatman 3 MM filter paper to collect the precipitates , and the filters were washed three times with 1 ml of a solution containing 5% trichloroacetic acid and 20 mM sodium pyrophosphate followed by 2 ml of 95% ethanol . The filters were then dried and quantified by using a Beckman Coulter LS 6500 scintillation counter as above . 32P-labeled RNAs ( 5 pM; 300 Ci/mmol ) were incubated with increasing concentrations of protein in a 50-µl reaction containing 100 mM KCl , 5 mM MgCl2 , 20 mM Tris-HCl ( pH 7 . 5 ) , 5 mM DTT , 0 . 1 mg/ml BSA , and 10% glycerol at either 25°C ( Figures 4 and 5 ) or 37°C ( Figure S4 ) . Binding reactions were initiated by adding 10 µl protein and terminated after 30 min by filtering 10 µl of the reaction through a nitrocellulose membrane ( Amersham Hybond ECL nitrocellulose; GE Healthcare Life Sciences ) backed by a nylon membrane ( Amersham Hybond-N+; GE Healthcare Life Sciences ) . The nitrocellulose membrane retains protein-bound RNA and the nylon membrane retains free RNA . The end point ( 30 min ) was chosen after determining that incubations times of 20 , 30 , or 60 min gave indistinguishable results for all proteins assayed . After application of samples , the membranes were washed three times with 20-µl wash buffer containing 100 mM KCl , 5 mM MgCl2 , and 20 mM Tris-HCl ( pH 7 . 5 ) , then dried and quantified using a PhosphorImager and the program ImageQuant ( GE Healthcare Life Sciences ) . Splicing time courses for the Nc mt LSU and Nc ND1m introns were done by pre-incubating 32P-labeled precursor intron RNA ( 50 or 200 nM; 0 . 13–0 . 4 Ci mmol−1 ) with protein ( 25 or 100 nM ) in a 100-µl reaction containing 100 mM KCl , 5 mM MgCl2 , 20 mM Tris-HCl ( pH 7 . 5 ) , 1 mM DTT , 0 . 1 mg/ml BSA , and 10% glycerol for 10 min on ice , followed by 5 min at reaction temperature . Reactions were initiated by adding 1 mM GTP-Mg2+ . Portions ( 8 µl ) were removed at different times , and the reaction terminated by adding 50 mM EDTA , followed by phenol-CIA extraction and mixing 10 µl of sample with 10 µl of 2× gel loading dye ( 95% formamide , 0 . 02% SDS , 0 . 02% bromophenol blue , 0 . 01% xylene cyanol , and 1 mM EDTA ) . End-point splicing assays were done similarly in reaction medium containing 25 or 100 mM KCl for 60 min . Splicing assays comparing wild-type CYT-18 and chimera CYT-18/Sc mtTyrRS proteins were also done with higher concentrations of 32P-labeled precursor RNA ( 200 nM , 0 . 13–0 . 4 Ci mmol−1 ) and protein ( 100 nM or 500 nM dimer ) and with 200 nM unlabeled precursor , 500 nM protein , and 500 nM [α-32P]GTP ( 3 , 000 Ci mmol−1; PerkinElmer ) . In all cases , samples were analyzed by electrophoresis in a denaturing 4% polyacrylamide gel , which was dried and quantified with a PhosphorImager , and data were analyzed by using ImageQuant TL .
|
The acquisition of new modes of post-transcriptional gene regulation played an important role in the evolution of eukaryotes and was achieved by an increase in the number of RNA-binding proteins with new functions . RNA-binding proteins bind directly to double- or single-stranded RNA and regulate many cellular processes . Here , we address how proteins evolve new RNA-binding functions by using as a model system a fungal mitochondrial tyrosyl-tRNA synthetase that evolved to acquire a novel function in splicing group I introns . Group I introns are RNA enzymes ( or “ribozymes” ) that catalyze their own removal from transcripts , but can become dependent upon proteins to stabilize their active structure . We show that the C-terminal domain of the synthetase is flexibly attached and has high non-specific RNA-binding activity that likely pre-dated the evolution of splicing activity . Our findings suggest an evolutionary scenario in which an initial non-specific interaction between an ancestral synthetase and a self-splicing group I intron was fixed by an intron RNA mutation , thereby making it dependent upon the protein for structural stabilization . The interaction then evolved by the acquisition of adaptive mutations throughout the protein and RNA that increased both the splicing efficiency and its protein-dependence . Our results suggest a general mechanism by which non-specific binding interactions can lead to the evolution of new RNA-binding functions and provide novel insights into splicing and synthetase mechanisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"rna-binding",
"proteins",
"rna",
"rna",
"structure",
"rna",
"folding",
"protein",
"structure",
"determination",
"rna",
"processing",
"molecular",
"complexes",
"proteins",
"enzymes",
"protein",
"structure",
"biology",
"and",
"life",
"sciences",
"enzymology",
"rna",
"splicing",
"molecular",
"biology",
"ribozymes",
"macromolecular",
"structure",
"analysis"
] |
2014
|
Evolution of RNA-Protein Interactions: Non-Specific Binding Led to RNA Splicing Activity of Fungal Mitochondrial Tyrosyl-tRNA Synthetases
|
Neutrophils are known to play a major role in the egg granulomatous lesions caused by Schistosoma japonicum , but the precise mechanism by which eggs recruit or active neutrophil is unknown . Here we report S . japonicum egg specific EF-hand protein-SjE16 . 7 is a potent neutrophil recruiter and initiates the egg associated inflammatory granuloma in schistosomiasis . We show that the expression of SjE16 . 7 at level of both mRNA and protein is restricted to the egg stage . It locates in the miracidium and subshell area of the egg and can be secreted by the egg . The antigenic properties of SjE16 . 7 strongly suggest a role for SjE16 . 7 as an egg-derived molecule involved in host-parasite interactions . To study SjE16 . 7 functions in vivo , we challenged murine air pouch with recombinant SjE16 . 7 . The results showed SjE16 . 7 trigged more inflammatory cell infiltration than vehicle or control protein . Using peritoneal exudate neutrophils from mice , we found that SjE16 . 7 significantly induced neutrophil chemotaxis in vitro , and the observed phenotypes were associated with enhanced Rac GTPase activation in SjE16 . 7 treated cells . Finally , in vivo hepatic granuloma formation model showed SjE16 . 7 coupled beads recruited more inflammatory cell infiltration than control beads . Our findings suggest SjE16 . 7 is an important pathogenic factor derived from egg . By recruiting neutrophils and inducing local inflammation , SjE16 . 7 facilitates eggs to be excreted through gut tissues and also initiates pathology in the liver; therefore SjE16 . 7 is a possible target for the prevention and treatment of schistosomiasis .
Schistosomiasis is a neglected tropical disease caused by trematode parasites of the genus Schistosoma . It is estimated to affect 200 million people globally and cause nearly 280 , 000 deaths reported annually [1] . S . japonicum is the major causative agent of schistosomiasis in South East Asia and China , which mainly cause “intestinal” and “hepatic schistosomiasis” . Deposited in the host liver or intestinal tissue , schistosome eggs are the major cause of pathology in schistosomiasis . They are viable metabolically active organisms , and highly antigenic . Eggs evoke inflammation leading to a granulomatous response that is necessary for its translocation into the intestinal lumen and excretion in the feces . Meanwhile this process initiates the pathology in host liver and intestine [2] . Neutrophils are believed to play a significant role in S . japonicum granulomatous pathology [3]–[5] . In the initiation of granuloma formation , neutrophils are recruited to the core of granuloma leads a neutrophil-mediated inflammatory response , which ultimately cause tissue damage [4] . At the later stage , neutrophils are recruited again and accumulated at the periphery of the granuloma , where they release a number of granule proteins involved in collagen degradation and reabsorption . It is well known , granuloma formation is a T cell-mediated immune response . The T cell-mediated response , especially CD4+ T cell response has been reported to participate in neutrophil induction in schistosomiasis [5] , [6] . However , in CD4+ T cell depleted mice , neutrophils still can be observed and become the dominated population in the cellular infiltrate around the egg [7] . These results suggest that neutrophils can be attracted by T cell independent way . Using crude extracts or antigen fractions from S . japonicum egg , Owhashi and Horri showed schistosome egg components have high neutrophil chemotactic activity [8] , [9] , but up to now the detail molecule and mechanism involved in chemoattractants for neutrophils are still not identified yet . Previously transcriptomic analyses of S . japonicum showed egg 16 kD calcium-binding protein ( SjE16 . 7 ) is specifically expressed in eggs , but the function of this protein and whether it plays roles in egg associated pathogenesis are unknown [10] . In this study , we cloned SjE16 . 7 from S . japonicum eggs , and then studied its function in host innate immune response . We showed as an egg-derived molecule , SjE16 . 7 promotes neutrophil chemotaxis through Rac GTPase pathway . It plays impotent roles in schistosome egg induced inflammatory granuloma; therefore SjE16 . 7 is a potential target for prevention and treatment of schistosomiasis .
The conducts and procedures involving animal experiments were approved by the Animal Ethics Committee of Shanghai Jiao Tong University School of Medicine ( project number 2012008 ) according to Regulations for the Administration of Affairs Concerning Experimental Animals ( approved by the State Council of the People's Republic of China ) and Guide for the Care and Use of Laboratory Animals ( Department of Laboratory Science , Shanghai Jiao Tong University School of Medicine , laboratory animal usage license number SYXK 2008-0050 , certificated by Shanghai Committee of Science and Technology ) . Chemicals were purchased from Sigma-Aldrich Co . unless otherwise noted . Ca2+ , Mg2+ , and phenol red-free Hanks balanced salt solution ( HBSS ) , phosphate-buffered saline ( PBS; PH7 . 2 ) and Dulbecco's Modified Eagle Medium ( DMEM ) were obtained from Life Technologies . NSC23766 was purchased from Tocris Bioscience . Polymyxin B was ordered for Sigma-Aldrich . Male C57BL6/J ( 6–7 week ) mice were purchased from Shanghai Laboratory Animal Center , Chinese Academy of Sciences . Mice were housed under specific pathogen-free conditions and fed autoclaved food and water as needed . Male New Zealand White rabbits ( 2 . 0–2 . 5 kg ) were provided and housed in the Department of Laboratory Animal Science , Shanghai Jiao Tong University School of Medicine . C57BL6/J mice were percutaneously infected with 30 S . japonicum cercariae ( Chinese mainland strain , Anhui population , National Institute of Parasitic Diseases , Chinese Center for Disease Control and Prevention ) . Adult worms were collected from infected mice 6 weeks post infection . Mice were perfused with PBS to remove worms from mesenteric veins . The male and female adult parasites were separated and stored in liquid nitrogen until use . Eggs were harvested from mouse liver . Tissues were minced with scissors in ice-cold 1 . 2% NaCl and passed through a crude sieve . The filtrate was passed through a series of sieves with decreasing pore size and finally eggs were collected from top sieve ( 45 µm ) . To collect the mature eggs , eggs were purified using a Percoll gradient , and then washed and concentrated by centrifugation . Eggs were used for circumoval precipitin test or stored in liquid nitrogen immediately . Total RNA was extracted from male , female or coupled adult worms using an RNAeasy Mini Kit ( Qiagen ) following the instructions of the manufacturer . The kit was also used to extract total RNA from eggs . First strand cDNA was synthesized using a Sensiscript RT Kit ( Qiagen ) . SjE16 . 7 gene sequence ( NCBI/GenBank AY816133 ) was obtained from NCBI/GenBank . A pair of primers 5′- ATGTCGGATGAAAACCGATGGATTGC-3′ and 5′-TTATTCATACGTTTGACGTACATAAGC-3′ was designed to amplify the ORF of cDNA using the first strand cDNA from females , males , coupled adults and eggs respectively , as templates . The PCR products were separated by running an agarose gel and a DNA band matching the designated size was cut and extracted using a Qiaquick Gel Extraction kit ( Qiagen ) . The DNA was then ligated into a cloning pGEM-T vector ( Promega , Madison , WI ) . Positive clones were selected and sequenced . To express the eukaryotic SjE16 . 7 in vitro , we designed a pair of primers: 5′-CGGGAATTCATGTCGGATGAAAACC-3′ and 5′- ATTGCGGCCGCTTAGTGGTGGTGGTGGTGGTGTTCATACGTTTG-3′ to subclone the SjE16 . 7 into the pPIC9K vector ( Invitrogen ) and transformed in Pichia pastoris strain GS115 ( Invitrogen ) according to the manufacturer's instructions . The generated protein was fused with His tag at the C terminal for affinity purification . The recombinant His-SjE16 . 7 protein was collected from the yeast culture supernatant and purified by Ni-NTA Superflow Cartridges according to the manufacturer's instructions ( Qiagen ) . The molecular weight and purity of recombinant proteins were identified by SDS-PAGE . Prokaryotic SjE16 . 7 protein was expressed in E . coli . A pair of primes: 5′-CGGGAATTCATGTCGGATGAAAACC-3′ and 5′-ATTGCGGCCGCTTATTCATACGTTTG-3′ was used to amplify the target gene and subcloned into the pGEX-4T-1 plasmid ( GE Healthcare life Sciences ) . The target gene was fused in frame with the N terminal GST tag . The plasmid was transformed into E . coli BL21 cells for protein expression . Protein expression was initiated by IPTG and cells were harvested after 4 h culture . Bacteria were lysed and sonicated . The recombinant fusion protein GST-SjE16 . 7 or GST control protein from E . coli lysates was purified using Glutathione Sepharose 4B ( GE Healthcare Life Sciences ) . LPS from prokaryotic expressed proteins was removed by Triton X-114 phase separation as literature described [11] . Briefly , Triton X-114 was added to the proteins to a final concentration of 1% and incubated for 30 min at 4°C with constant stirring , followed by 10 min incubation at 37°C and a centrifugation step at 16 , 000×g at 25°C , for 10 min . Six cycles of Triton X-114 phase separation was performed for a sufficient LPS depletion . Trace amounts of Triton X-114 were removed by dialysis against PBS . Finally LPS content was detected by Tachypleus Amebocyte Lysate Kits ( Gulangyu , Xiamen China ) according to the manufacturer's recommendation . In some experiments , purified GST-SjE16 . 7 was cleaved with thrombin ( GE Healthcare life Sciences ) to remove the GST tag . GST free SjE16 . 7 protein was used in ELISA . GST-SjE16 . 7 fusion protein was used in other experiments ( air pouch model , transwell migration assay , PBD pull-down assay and inflammatory hepatic granuloma model ) , while GST protein was used as control . SjE16 . 7 antiserum was prepared in C57BL6/J mice and New Zealand White rabbits . Recombinant SjE16 . 7 was formulated with either Freunds complete ( primary ) or Freunds incomplete ( two boosts at two weekly intervals ) adjuvants and the preparations were subcutaneously injected into the animals . Animals were sacrificed 2 weeks after the final antigen immunization and sera were collected from the blood . Antibodies were purified from rabbit anti-SjE16 . 7 sera . Immunoglobulins were precipitated with ammonium sulfate first , and then purified using Protein A/G agarose beads according to the product's instruction ( Pierce Biotechnology ) Soluble egg antigen ( SEA ) and adult worm antigen ( AwAj ) of S . japonicum were prepared as described previously [12] . Briefly , eggs or adult worms were suspended in PBS and homogenized in an ice-chilled water bath . The mixture was centrifuged at 100 , 000 g for 1 h . The supernatant was used as SEA or AwAj . Protein concentration was measure by standard Bradford protein assay ( Biorad ) using bovine serum albumin as a standard . SDS-PAGE was performed by the procedure of Laemmli using 10–12% polyacrylamide gels in presence of 5% 2-mercaptoethanol in sample buffer . Protein molecular-weight markers ( Fermentas or Biorad ) were used as MW standards . Proteins were visualized by Coomassie brilliant blue staining . Western blotting was performed on nitrocellulose filter ( Biorad ) . Blots were immunostained with 100 times diluted mouse anti-SjE16 . 7 serum , and 1000 times diluted horseradish peroxidase ( HRP ) conjugated goat anti-mouse IgG ( Cell Signaling ) . Enhanced chemiluminescence ( ECL , Pierce ) was used as substrates and signals were analyzed by Luminescent imager ( ImageQuant Las 4000 , GE Healthcare ) . S . japonicum eggs were isolated from the livers of infected mice . Live eggs were cultured at 37°C with normal rabbit sera , rabbit anti-SjE16 . 7 sera or sera from S . japonicum infected rabbits . Eggs were examined for the presence of precipitates around eggs 48 h later . Liver specimens were fixed in 10% formalin , embedded in paraffin and sectioned at 3 µm . Following antigen retrieval by boiling in 0 . 01 M sodium citrate , pH 6 , for 20 minutes in water bath , endogenous peroxidase activity was blocked by incubation with 3% ( v/v ) H2O2 for 20 min at room temperature . Slides were washed three times with PBS , and blocked with 5% bovine serum albumin for 20 min at room temperature . Tissue sections were then probed with 1∶50 diluted rabbit anti-SjE16 . 7 sera for 1 hour at 37°C . Slides were washed three times with PBS , and HRP conjugated secondary antibody was added ( anti-rabbit IgG in MaxVisionTM HRP-Polymer anti-Rabbit IHC Kit , Maxim , Fujian , China ) and incubated for 12 min at room temperature . Slides were washed again in PBS 3 times and developed using the AEC ( 3-amino-9-ethylcarbazole ) substrate system ( Maxim , Fujian , China ) . Counterstaining was done with hematoxylin . Identical concentration of normal rabbit sera were used as negative control . Sections were examined with an Olympus BX51 microscope and acquired with an Olympus DP12 digital camera controlled by CellSens Standards software . Mouse sera were collected from mice before and 2 , 4 , 6 , 8 , 10 weeks after S . japonicum infection . Sera were kept at −80°C until use . Rabbit sera before and after S . japonicum infection were kindly provided by Dr . Wei Hu . Antibody reactivity of animal sera against recombinant prokaryotic SjE16 . 7 ( GST-tag free ) , eukaryotic His-SjE16 . 7 and SEA were determined by enzyme-linked immunosorbent assay ( ELISA ) using adaptations of previously described methods [13] . In short , microtitration plates ( Nunc , Denmark , 96 wells , flat bottom ) were coated with 100 µl 5 µg/ml SEA , SjE16 . 7 or His-SjE16 . 7 respectively . Mouse antisera were diluted 1∶100 and HRP conjugated goat anti-mouse/rabbit IgG or IgM Abs ( Sigma ) was used as the secondary antibody at a dilution of 1∶1000 . Reactions were developed using 3 , 3′ , 5 , 5′-Tetramethylbenzidine ( TMB ) substrates and stopped with 2 N H2SO4 . The optical densities were read at 450 nm in a microwell reader system ( μQuant , Bio-Rad , USA ) . Dorsal air pouches were induced in mice using previously described methods [14] . In brief , 4 ml of sterile-filtered air was injected subcutaneously into the back of C57BL6/J mice , and the pouch was reinflated with 3 ml of sterile air 3 d later . The dorsal air pouches of groups of 5–6 mice were either injected with 1 ml 0 . 5% carboxymethylcellulose ( CMC ) or 0 . 5% CMC with 50 µg recombinant antigens ( His-SjE16 . 7 , GST , or GST-SjE16 . 7 ) 3 d later . Four hours or 24 h later , the mice were sacrificed and air pouches were lavaged with 3 ml sterile PBS . The aspirate was centrifuged at 500 g 10 min at 4°C . Supernatants were separated and stored at −80°C until testing . Cell pellets were re-suspended in PBS and counted in a standard hemocytometer chamber . Cells were incubated with FITC-conjugated anti-CD11b monoclonal antibody ( MAb ) , Percy- Cyanine5 . 5 conjugated anti-Ly6G ( Gr-1 ) MAb and APC conjugated anti-F4/80 MAb . The cells were analyzed on a FACSCalibur flow cytometer ( BD Bioscience ) equipped with Cell Quest software . Neutrophils were defined as cells that were CD11b+Ly6GhighF4/80− and macrophages as cells that were CD11b+Ly6G−F4/80+ . Leukocyte myeloperoxidase ( MPO ) activity was assessed by measuring the H2O2-dependent oxidation of TMB as previously reported [15] . Supernatants collected from air pouch aspirates were used for the assay . Aliquots of 30 µl were incubated with 120 µl of TMB substrates in 96 well plates . Plates were incubated for 5 min at room temperature and stopped with 2 N H2SO4 . The optical densities were read at 450 nm in a microwell reader system ( μQuant , Bio-Rad , USA ) . All the samples were performed in duplicates , and samples from the one experiment were tested on the same plate . To prepare peritoneal neutrophils , 6–7 week old mice were injected intraperitoneally with 2% casein/PBS . After 4 hour , peritoneal exudates cells were collected , spun down , and suspended in HBSS supplemented with 0 . 1% BSA ( HBSS/0 . 1%BSA ) . Neutrophil viability was >95% according to the results from trypan blue staining . Purity was typically 90% as assessed by flow cytometry based on the forward and side scatter and high Ly6G staining . Assay were performed in modified Boyden Chambers ( Transwell from Costar , Corning Life Sciences ) ( 6 . 5 nm in diameter; 3 µm pore size ) , according to the manufacturer's protocol . In brief , the bottom sides of the inserts were coated with 2 µg/ml fibronectin . Murine neutrophils were plated ( 1×106/well in DMEM , 0 . 1% BSA ) in the top chamber of Transwell inserts , and 0 . 1%BSA/DMEM with antigens ( SjE16 . 7 or SEA ) was added to the bottom chamber . After incubation for 3 h at 37°C and 5% CO2 humidified environment , 50 µl 70 mM EDTA was then added to the top and bottom chambers to release the cells that have adhered to the well and bottom of the membrane . After incubation for 5 min , the number of transmigrated cells in the lower compartment was determined with a hemocytometer . Murine neutrophils ( 5×106 ) were treated with 1 µM SjE16 . 7 or control protein for 10 min or the time as indicated . After treatment , cells were washed in PBS , and lysed in PBD lysis buffer ( 50 mM Tris pH 7 . 5 , 10 mM MgCl2 , 0 . 2 M NaCl , 0 . 5% NP-40 , and 1× protease inhibitors cocktail ( Roche ) ) . The lysate was incubated with 20 µg of PAK-GST protein beads ( Cytoskeleton ) for 30 min at 4°C . After washing , protein on beads and in total cell lysates was subjected to Western blot ( Rabbit anti-Rac1/2/3 Ab , Cell Signaling ) to determine the level of active Rac . Glutathione Sepharose 4B beads ( GE Healthcare Life Sciences ) were incubated with GST or GST-SjE16 . 7 protein ( 2 mg antigen/ml beads ) for 2 h at room temperature . The mixture was centrifuged for 2 min at 5 , 000 g , and the resulting pellet was washed twice with 10 bed volumes of sterile PBS . The beads were resuspended in sterile PBS at the concentration of 1×105 beads/ml before injection . Inflammatory hepatic granulomas were induced in mice by injecting antigen-coupled beads in the cecal vein of the mouse as described [16] , [17] . The mouse was anesthetized ( 60 mg/kg pentobarbital intraperitoneally ) , and a midabdominal incision was made . 20 , 000 beads coupled with GST or GST-SjE16 . 7 dissolved in 0 . 2 ml of sterile PBS were injected in the cecal vein using an insulin syringe ( Becton Dickinson ) . Unloaded beads were injected as negative control . 4–6 mice were used for each group . To check the role of SjE16 . 7 in initiation of inflammatory hepatic granulomas , the animals were sacrificed 6 h after injection . To evaluate the acute granulomatous response , consecutive 3 µm-thick formalin-fixed paraffin-embedded liver sections were stained with H&E . H&E-stained sections were examined with an Olympus BX51 microscope and acquired with an Olympus DP12 digital camera controlled by CellSens Standards software . A quantitative analysis of inflammatory response around the beads ( diameter between 50 and 100 µM ) was performed by counting the leukocytes in the granuloma . 30 granulomas per mouse were calculated . Groups were compared using the two-tailed student t test and analysis of variance ( ANOVA ) with GraphPad Prism 5 software . A nonparametric Mann-Whitney U-test was used for analysis of the western blot data because of the relatively small sample size in each experiment . Results were considered significant at a P value of <0 . 05 . The GenBank accession numbers for SjE16 . 7 and SmE16 are AAW27865 . 1 and AAA29859 . 1 respectively . The GenBank accession numbers for Clonorchis sinensis calcium-binding protein , S . mansoni calmodulin , C . sinensis 16 kDa calcium-binding protein and Fasciola hepatica CaM3 are GAA27855 . 1 , XP_002580524 . 1 , GAA40982 . 1 and AFM84631 . 1 respectively . The GenBank accession number for S . japonicum protein SJCHGC05185 is AAW26060 . 1 .
The cDNA clone encoding the full-length sequence of SjE16 . 7 of S . japonicum was obtained by RT-PCR amplification with total RNA extracted from egg . In agree with the results reported by other groups [10] , no SjE16 . 7 cDNA could be amplified from male , female or coupled-adult RNA ( Fig . 1A ) . The amplified full-length cDNA sequence of SjE16 . 7 was verified by sequencing ( Majorbio , China ) . The sequence was comprised a 435 bp ORF encoding 145 amino acid residues with the predicted molecular mass of ∼16 . 725 kDa and theoretical isoelectric point of pH 4 . 88 . ( The molecular mass and isoeletric point of SjE16 . 7 were calculated using the Compute pI/Mw tool ( <http://web . expasy . org/compute_pi/> ) . Comparison of the predicted amino acid sequence showed that SjE16 . 7 was 70% identical with the homolog from S . mansoni ( SmE16 ) according to multiple sequence alignment analysis ( GenBank AAW27865 . 1 vs . AAA29859 . 1 ) ( Fig . 1B ) . To prepare the recombinant SjE16 . 7 protein , the gene was cloned into eukaryotic expression vector pPIC9K and prokaryotic vector pGEX-4T-1 respectively . The eukaryotic recombinant protein was expressed in the yeast Pichia pastoris as His fusion protein ( His-SjE16 . 7 ) with an expected molecular mass of ∼17 kDa ( Fig . 1C ) . Prokaryotic protein was expressed in Escherichia coli as a GST-tagged protein ( GST-SjE16 . 7 ) of ∼42 kDa in size . The molecular weight of purified SjE16 . 7 after thrombin treatment is around 16 kDa ( Fig . 1D ) . After 6 cycles of Triton X-114 phase separation , LPS in GST and GST-SjE16 . 7 was between 0 . 10 and 0 . 15 EU/ml , which is the FDA endotoxin limit for drugs ( data not shown ) . SjE16 . 7 transcript is highly expressed in egg stage but not in adults . To investigate whether also SjE16 . 7 protein occurs specifically in the egg stage , a polyclonal mouse antiserum was raised against His-tagged SjE16 . 7 . We analyzed protein extracts obtained from adult and egg stages of SjE16 . 7 by SDS-PAGE followed by Western blotting . To ensure that comparable amount of proteins of each stage were present , the samples were duplicated . One SDS-PAGE gel was stained with Coomassie brilliant Blue . The other gel was transferred to the membrane and then blotted with mouse SjE16 . 7 antiserum . As on the mRNA level , SjE16 . 7 could be clearly detected only in the egg stage ( Fig . 2A ) . Concerning the site of expression of SjE16 . 7 within the eggs , sections of S . japonicum-infected mouse liver were stained with rabbit anti-SjE16 . 7 serum . Immunohistochemistry results showed that inside the egg , SjE16 . 7 locates in the miracidium and presents between the eggshell and miracidium , the subshell area of the egg ( Fig . 2B ) . SjE16 . 7 also shows within the granuloma surrounding the egg which indicates it is secreted by the egg . To further clarify that SjE16 . 7 is a secretion protein , we performed circumoval precipitin tests ( COPT ) . Live schistosome eggs were cultured at 37°C with normal rabbit sera or anti-SjE16 . 7 sera in vitro for 48 hours . The formation of precipitates showed around the eggs cultured with anti-SjE16 . 7 sera but not eggs cultured with normal sera ( Fig . 2C ) As an egg specific protein , SjE16 . 7 can be secreted by the egg . We then asked the antigenicity of SjE16 . 7 , whether this protein provokes immune responses during infection with schistosomes . To address this question , mice were infected with S . japonicum cercariae . The antibody production against complete SEA , or SjE16 . 7 ( prokaryotic SjE16 . 7 or eukaryotic His-SjE16 . 7 ) was detected by ELISA respectively . The ELISA results showed that the infected mice developed specific antibodies against SjE16 . 7 as SEA , while sera from control mice without infection revealed no specific antibodies ( Fig . 2D ) . The levels of specific IgM and IgG antibodies were increased significantly 6 weeks after infection and remained at high level until the mice were sacrificed at 10 weeks after infection ( Fig . 2E ) . Similar phenotypes were observed in S . japonicum infected rabbits . As shown in Fig . 2F , 6 weeks post infection , the infected rabbits develop significantly enhanced antibodies against SjE16 . 7 . The antibody levels reach a peak 12 weeks post infection . Although the antibody levels declined in the late infection , compare to the antibody levels before infection , the differences are still significantly . Schistosome eggs are highly immunogenic . Within infected hosts , eggs induce vigorous immune responses and are rapidly surrounded by inflammatory cells , creating a granuloma . Above results suggest that SjE16 . 7 is an egg specific antigen and recognized early during infection with S . japonicum in mice . To investigate the role of SjE16 . 7 in induction of inflammation , we used a murine air pouch model of inflammation . As shown in Fig . 3 , injection of vehicle or control protein produced a modest infiltration of the cells into the pouch . In contrast , administration of either eukaryotic or prokaryotic recombinant SjE16 . 7 produced an intense accumulation of inflammatory cells at 4 h and 24 h . Myeloperoxidase ( MPO ) is one of the principal enzymes released from neutrophils during inflammatory responses . Consistent with enhanced inflammatory cells , the level of MPO was found increased significantly in SjE16 . 7 induced exudates ( Fig . 3A ) . The analysis of cell population in the exudates induced by SjE16 . 7 showed CD11b+ myeloid cells , especially CD11b+Ly6Ghigh neutrophils were the predominant cell type ( Fig . 3B ) . SjE16 . 7 recruited neutrophil infiltration in vivo . To further address the effect of SjE16 . 7 on chemotaxis of neutrophil , we examined migration of the cells in a transwell chamber assay . As shown in Fig . 4A and 4B , the presence of SEA or SjE16 . 7 in the bottom chamber dramatically promoted migration and the function of SjE16 . 7 was dose dependent . Furthermore , the attraction of neutrophils by SjE16 . 7 can be neutralized by anti-SjE16 . 7 antibody ( Fig . 4C ) . Prokaryotic SjE16 . 7 protein was prepared from E . coli . We removed majority LPS ( from 10 EU/ml to <0 . 25 EU/ml ) in antigen purification , but the residual LPS contamination still might be a chemoattractant for neutrophils . To test this possibility , LPS inhibitor , Polymyxin B was added in our chemotaxis assay . As shown in Fig . 4D , Polymyxin B didn't inhibit prokaryotic SjE16 . 7 induced neutrophil migration . Since cell movement requires dynamic reorganization of the action cytoskeleton and membrane polarization , and both of them regulated by Rac GTPases , we measured Rac GTPase activation in SjE16 . 7 treated neutrophils . Using GST pull-down assay that specifically recognized active GTP-bound Rac , we tested the SjE16 . 7 effect on neutrophils . We found that SjE16 . 7 treated neutrophils exhibited an increase in active Rac levels over controls ( Fig . 5A ) . SjE16 . 7 treatment induces an intense accumulation of GTP-Rac that peaked at 10–20 min , declining over 25 min ( Fig . 5B ) . To determine whether SjE16 . 7 regulates chemotaxis through Rac , in transwell assay neutrophils were pretreated with Rac antagonist NSC23766 . As shown in Fig . 5C and Fig . 5D , NSC23766 depressed the SjE16 . 7 induced cell chemotaxis in a dose-dependent manner . In schistosome infection , neutrophils are known as a major cellular component in the early phase of egg-associated granulomatous lesion and they are important in initiation of egg granuloma . We next measured the ability of SjE16 . 7 in initiation of mouse hepatic granuloma . Agarose beads were coated with GST-SjE16 . 7 or GST control protein . C57BL6/J mice were injected with 20 , 000 beads and sacrificed at 6 h after injection . Beads were found lodged in the peripheral perisinusoidal ramifications of the portal vein . Uncoated beads and beads coated with control protein showed no cellular reaction , except for a monolayer of leucocytes ( Fig . 6 ) . Beads coated with GST-SjE16 . 7 elicited significant granulomatous inflammatory reaction . Quantification of the leucocytes around the beads revealed that GST-SjE16 . 7 beads recruited more leucocytes ( 78 . 88±15 . 83/bead ) than control beads ( 12 . 78±13 . 31/bead ) or GST beads ( 39 . 64±10 . 73/bead ) . The difference between GST-SjE16 . 7 group and GST group or control group was significant ( GST-SjE16 . 7 vs . GST , p<0 . 01; GST-SjE16 . 7 vs . Control , p<0 . 001 ) .
Neutrophils are known to play a major role in the disease caused by S . japonicum [4] . For example , in the initiation of egg-induced granulomatous pathology , neutrophils are recruited to the egg and become the dominant cell population . The accumulating neutrophils cause necrotic lesions in the liver or intestinal tissue of the host . In this study , we identified SjE16 . 7 as a potent neutrophil recruiter derived from S . japonicum egg . It significantly induces neutrophil chemotaxis via Rac GTPase pathway and stimulates neutrophil infiltration in vivo . We propose SjE16 . 7 is an important pathogenic factor which facilitates eggs to migrate through gut tissues and also initiate pathology in the liver . It is a potential target for the disease prevention and treatment . SjE16 . 7 transcript was firstly identified by Hu and her colleagues . They compared the expressed sequence tags ( ESTs ) derived from cercariae , schistosomulae , adults and eggs of S . japonicum , then found transcript encoding SjE16 . 7 is expressed specifically in egg stage [10] . Consistent with this result , our PCR results revealed that SjE16 . 7 mRNA is restricted to the egg stage . Furthermore , Western blotting performed on the adult and egg stages of S . japonicum confirmed that at the protein level , SjE16 . 7 is present in eggs , and not detectable in adult worms . Searching with BLASTP at the NCBI database , SjE16 . 7 displays besides 70% identity with S . mansoni SmE16 , 43% identity with S . mansoni calmodulin , 41% and 39% identity with Clonorchis sinensis calcium-binding protein and 16 kDa calcium-binding protein respectively , and 38% identity with Fasciola hepatica CaM3 . SmE16 is most related with SjE16 . 7 , while the S . mansoni calmodulin , 16 kDa calcium-binding protein of C . sinensis and F . hepatica CaM3 are more related to another S . japonicum protein SJCHGC05185 . SmE16 is an S . mansoni egg specific calcium-binding protein , firstly reported by Moser in 1992 [18] . Mathieson and Wilson compared the proteome of the undeveloped and developed S . mansoni egg and showed SmE16 was expressed at a low level in undeveloped , but increased in abundance in developed eggs [19] . In mature eggs , SmE16 could be assigned to miracidium and the water-soluble protein washed from the egg as the miracidium emerged ( hatch fluid ) . These results are consistent with SjE16 . 7 distribution in the egg . SjE16 . 7 presents in the miracidium and subshell area of the S . japonicum egg . To identify egg secretin proteins of S . mansoni , Mathieson and Wilson collected egg secretion protein by incubating mature eggs in serum-free RPMI . In their study egg secretion protein was characterized by multiple isoforms/variants of just 6 proteins and SmE16 was not in it . However in the similar work done by another group in the Unite State , SmE16 was identified as one of the abundant S . mansoni egg secreted proteins [20] . As an egg specific protein , SjE16 . 7 is a secretion protein and can be recognized by the host immune system . This conclusion is based on the following observations: ( a ) In COPT , the formation of precipitates showed around the live eggs cultured with anti-SjE16 . 7 sera in vitro . ( b ) In immunocytochemistry , SjE16 . 7 was detected within the egg and within the granuloma surrounding intact eggs in the liver of mice with S . japonicum infection . ( c ) 6 weeks post S . japonicum infection , animals develop significantly specific antibodies against SjE16 . 7 . Immunohistochemstry staining of S . japonicum eggs in liver sections of infected mice revealed that SjE16 . 7 locates in the miracidium and subshell area of the egg . This subshell area is supposed to be responsible for egg secretion protein synthesis and storage [21] . Several secretary proteins , such as IPSE/alpha-1 , Omega-1 and Kappa-5 have been found in this area [22]–[24] . SjE16 . 7 shows in subshell area , but whether it is synthesized in that area , or it is a secretary/excretory protein produced by miracidium and stored in subshell area is presently under study . The localization and its antigenic properties strongly suggest a role for SjE16 as an egg-derived molecule involved in host-parasite interactions . Analysis of SjE16 . 7 protein sequence revealed SjE16 . 7 has 2 EF-hand domains . The classical EF-hand is composed of 2 perpendicular alpha helices separated by a loop , which form a helix-loop-helix motif . The loop integrated in this motif can accommodate Ca2+ or Mg2+ with distinct geometries and the affinity for these ions is determining factor for the function of the protein [25] . Previous studies of schistosome showed proteins containing EF-hand domain may play roles in modulating host immune responses during the process of host-parasite interaction [26] . Sm22 . 6 and Sm20 . 8 are schistosome tegumental antigens exclusively expressed in adult worms . Both of them have 2 EF-hand motifs and are dominant targets for human IgE responses in the process of schistosome infection and therefore , are important for parasites survival [27] , [28] . SmE16 is a stage specific calcium-bind protein expressed in eggs of S . mansoni . It has two EF-hands and share 70% identical sequence with SjE16 . 7 . The detection of specific antibodies to the SmE16 in sera of schistosomiasis patients suggests a role for SmE16 as an egg-derived molecule involved in host-parasite interactions , but the function of this protein in the egg is still unclear [18] . EF-hand structural motif has been found in a large number of protein families , possessing diverse functions . In inflammation , EF-hand proteins participate in many steps of inflammatory processes , including leukocyte migration , inflammatory cell activation , cytokine production and et al . [25] . For example , mammalian S100A8 and S100A9 are members of S100 family which has 2 EF-hand motifs . They are regarded as marker proteins for a number of inflammatory diseases [29] . Studies in human and mouse showed S100A8 and S100A9 have potent chemotactic activity for neutrophils and monocytes [29] , [30] . They regulate leukocytes trafficking by acting as a chemoattractant and influencing expression of adhesion molecules . Similar to S100A8 and S100A9 , SjE16 . 7 has 2 EF-hand motifs and potent chemotactic activity for neutrophils . It might be an EF-hand protein used by parasite to mediate the inflammatory responses which facilitate egg passage through the tissues . Neutrophils are highly motile leukocytes . For neutrophil migration to occur there is a constant need for the cell to coordinate a variety of intracellular activities both spatially and temporally [31] . The small Rho guanosine triphosphatases ( GTPases ) Rac1 and Rac2 are key players in this process and their ability to cycle between active ( GTP-bound ) and inactive ( GDP-bound ) status allows the cell to respond rapidly to extracellular signals . Regulating membrane polarization or cytoskeletal dynamics , Rac GTPases promote cell migration . The up-regulation of GTP-Rac in SjE16 . 7 treated neutrophils suggested SjE16 . 7 may induce cell migration trough Rac . This mechanism was confirmed by Rac antagonist NSC23766 which depressed the SjE16 . 7 induced cell migration . The major activator of Rho family GTPase is G-Protein coupled Receptor ( GPCR ) [32] . It has been reported GPCR was used by other EF-hand protein to stimulate myeloid cell chemotaxis [33] . Whether SjE16 . 7 uses the same receptor to activate Rac and simulate neutrophil migration remains to be determined . Besides Rac GTPases , a huge variety of intracellular signaling molecules have been implicated in neutrophil migration , including MAPK cascades , PI3Ks , phospholipases , scaffold proteins and other GTPases [31] , [34] . In the cells , those signaling molecules interplay and cooperate with each other , and then cause the cell migration . For example , after receptor-chemoattractant interaction , PI3Ks activation leads to PIP3 accumulation at the leading edge of the cells . PIP3 coordinates Rac-GEFs recruitment and in turn Rac activation that allows actin polymerization . In this study , we checked Rac GTPase in SjE16 . 7 induced neutrophil chemotaxis , however other signaling molecules in neutrophil migration are still worth to be considered . The granulomatous response to the egg is a dynamic process . It is orchestrated by neutrophils , macrophages , CD4+ T cell , and multiple other cell types . In this study we focus on the relationship between SjE16 . 7 and neutrophil , and found it recruits neutrophil infiltration . It would be interesting to know the effects of SjE16 . 7 on other cell types and whether it influences on the development and resolve of granuloma . Macrophages are one of the major cells populations in egg granuloma . They are involved in the initiation of granuloma formation [35] and play essential roles in granuloma developing and resolving . In our hands , SjE16 . 7 dramatically chemoattracts macrophages as neutrophils ( our unpublished data ) . Moreover , SjE16 . 7 enhanced the cytokine expression and secretion in macrophages through intracellular MAPK signal pathway ( our unpublished data ) . Macrophage recruitment and activation are essential for the development and resolve of the granuloma . For example , the presence of alternative active macrophages ( AAMs ) in the granuloma downmodulates the initiate inflammatory responses and provides a readily available supply of proline to the fibroblasts resulting in collagen synthesis [2] . SjE16 . 7 promotes macrophage migration may contributes to the initiation of the granuloma . Whether SjE16 . 7 induced macrophage activation is associated with AAM polarization , therefore affects granuloma development is worth to be studied . An effective T-cell response is known to be critical for the development of the granulomatous response . To date , we haven't worked on it yet . Further work is required to know the relationship between SjE16 . 7 and other cell types , and fully elucidate roles of SjE16 in the developing and resolving granuloma .
|
As a neglected disease , schistosomiasis continues to be a significant cause of parasitic morbidity and mortality worldwide . Schistosoma japonicum is one of the major causative agents of human schistosomiasis . Trapped in the liver or intestinal tissue , S . japonicum eggs are the main cause of pathology following infection . They induce vigorous immune responses from the host , which facilitate the passage of the eggs from the tissue to the gut lumen and cause the pathology in liver . In this paper , we described , for the first time , S . japonicum egg specific EF-hand protein-SjE16 . 7 is a potent neutrophil recruiter and initiates the egg associated inflammatory granuloma in schistosomiasis . This study presents a precise mechanism by which eggs recruit neutrophil and induce inflammatory response . It furthers our understanding of the immunopathogenesis of human schistosomiasis . In addition , it provides a potential target for the prevention and treatment of this globally important parasite .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"parasitology",
"biology",
"microbiology"
] |
2014
|
Schistosoma japonicum Egg Specific Protein SjE16.7 Recruits Neutrophils and Induces Inflammatory Hepatic Granuloma Initiation
|
Bolivia has the highest prevalence of Chagas disease ( CD ) in the world ( 6 . 1% ) , with more than 607 , 186 people with Trypanosoma cruzi infection , most of them adults . In Bolivia CD has been declared a national priority . In 2009 , the Chagas National Program ( ChNP ) had neither a protocol nor a clear directive for diagnosis and treatment of adults . Although programs had been implemented for congenital transmission and for acute cases , adults remained uncovered . Moreover , health professionals were not aware of treatment recommendations aimed at this population , and research on CD was limited; it was difficult to increase awareness of the disease , understand the challenges it presented , and adapt strategies to cope with it . Simultaneously , migratory flows that led Bolivian patients with CD to Spain and other European countries forced medical staff to look for solutions to an emerging problem . In this context , thanks to a Spanish international cooperation collaboration , the Bolivian platform for the comprehensive care of adults with CD was created in 2009 . Based on the establishment of a vertical care system under the umbrella of ChNP general guidelines , six centres specialized in CD management were established in different epidemiological contexts . A common database , standardized clinical forms , a and a protocolized attention to adults patients , together with training activities for health professionals were essential for the model success . With the collaboration and knowledge transfer activities between endemic and non-endemic countries , the platform aims to provide care , train health professionals , and create the basis for a future expansion to the National Health System of a proven model of care for adults with CD . From 2010 to 2015 , a total of 26 , 227 patients were attended by the Platform , 69% ( 18 , 316 ) were diagnosed with T . cruzi , 8 , 567 initiated anti-parasitic treatment , more than 1 , 616 health professionals were trained , and more than ten research projects developed . The project helped to increase the number of adults with CD diagnosed and treated , produce evidence-based clinical practice guidelines , and bring about changes in policy that will increase access to comprehensive care among adults with CD . The ChNP is now studying the Platform’s health care model to adapt and implement it nationwide . This strategy provides a solution to unmet demands in the care of patients with CD , improving access to diagnosis and treatment . Further scaling up of diagnosis and treatment will be based on the expansion of the model of care to the NHS structures . Its sustainability will be ensured as it will build on existing local resources in Bolivia . Still human trained resources are scarce and the high staff turnover in Bolivia is a limitation of the model . Nevertheless , in a preliminary two-years-experience of scaling up this model , this limitations have been locally solved together with the health local authorities .
The prevention and control of CD in Bolivia was declared a national priority in Law 3374 dated March 23 , 2006 [12] . However , no additional regulations were developed . The Chagas National Programme ( ChNP ) , the responsible body for the prevention , diagnosis and treatment of CD in the country , elaborated drafted its Strategic Plan 2010–2015 [13] . Fig 1 provides a holistic and inter-sectorial overview of the Strategic Plan . Given the epidemiological situation of Bolivia in the 1990s ( infestation rates >50% ) , the priority area was vector control , which was financed by the Inter-American Development Bank ( Banco Interamericano de Desarrollo , BID ) from 1999 to 2006 . This support included treatment of children but not adults with chronic CD , and the control of congenital disease ( supported with funds from the Belgian Government until 2009 ) was prioritized . [14] Diagnosis and treatment of children under 18 years old should in theory be provided by primary healthcare centers in Bolivia , however many centers do not systematically provide this service . Even though CD affected mainly adults in Bolivia , no model for care of adults with chronic infection was defined , and therefore primary health center do not contemplate etiological treatment for this population . In 2009 , the ChNP reported that 178 , 012 persons had been screened . Most were pregnant women who were monitored for congenital transmission , although only 3 , 103 were treated ( 10% of those confirmed as having T . cruzi infection in the same year , representing only around 0 . 5% of all estimated patients with the infection ) ( Table 1 ) . Additionally , health professionals’ knowledge of CD was limited . Since their training included information on the very frequent adverse effects of aetiological treatment and the autoimmune origin of cardiac involvement [16] , medical staffs were reluctant to recommend aetiological treatment of chronic CD in adults . Consequently , treatment of adults was neglected , as occurred in other endemic countries . Today , there is sufficient evidence on the role of T . cruzi in triggering and sustaining the inflammatory response [8 , 16] and , therefore , on the importance of early anti-parasitic treatment . Furthermore , the lack of information on the benefits of treatment , together with fear and an alternative understanding of CD by at-risk persons , limited patients’ active demand for treatment [17 , 18] . Additionally , access to healthcare was hampered by the absence of symptoms , the non-specific nature of symptoms when present , and limited access to health centres , especially in rural settings . Following a change in international consensus , the Chagas Platform was developed as a joint initiative that arose from the need to offer diagnosis and anti-parasitic treatment to adult patients in the chronic phase of T . cruzi infection . [19 , 20] In 2009 , ISGlobal ( Barcelona , Spain ) and CEADES ( Cochabamba , Bolivia ) pooled their expertise in the comprehensive management of adult patients by developing a care model with the Bolivian Ministry of Health and Bolivian state universities to collaborate in research and training of health professionals . The primary objective of the Chagas Platform is to contribute to the control of CD , and the model designed to achieve this objective is based on 4 pillars: The Chagas Platform is therefore considered a translational model in which provision of care is the initial trigger of research needs , thus initiating a circular cycle where the results of research are applied in to healthcare and are used to train staff and effect changes in health policy . In this manuscript , we focus on comprehensive care and staff training as critical components for future scaling up of access to diagnosis and treatment . Comprehensive healthcare based on agreed protocols was initially provided in vertical , dedicated structures for adults . Even when these structures were located in existing health structures , they were conceived as specific units for CD , instead of as units for integrating care of CD patients in the normal outpatient care circuit . This strategy made it possible to create centres of expertise and ensured sufficient capacity to increase the number of people diagnosed and treated . It simultaneously generated a critical mass of patients that allowing to pilot the use of comprehensive care strategies that could subsequently be integrated in the national health system and to advance in key research areas . Additionally , the results of the Program revealed the magnitude of the problem and the need for a national strategy for patients in the chronic phase of CD . There are currently six centres in three highly endemic departments of Bolivia: Tarija ( one ) , Chuquisaca ( one ) , and Cochabamba ( four: two in rural areas and two in urban or semi-urban areas ) . The centres were established with different organizational set-ups that varied depending on the local partners that committed to the project in each area . Although each set-up has its particularities , they all share the same protocols and database . The network of Chagas Platform centres in both rural and urban areas enables patients to be transferred from one geographical area to another . These circuits ensure better coverage and improved access to healthcare . The Chagas Platform centres offer their services free of charge . The project was built in collaboration with the Spanish Agency for International Development Cooperation ( AECID ) and contributions from local partners . The ChNP covers drug costs . Of the 35 persons working in the six centres , two are covered by SEDES ( the departmental health authority in Chuquisaca ) and eight by the Juan Misael Saracho University ( Tarija ) ; the remaining 25 are covered by ISGlobal and CEADES with AECID funds . Local authorities are expected to progressively assume the cost of human resources in the future . The success of the model relies on the protocols and clinical guidelines used ( see Fig 2 ) . The main elements are as follows: Another key element for the success of the program is the capacity of the staff to provide quality healthcare for CD patients . As health professionals were not sufficiently well trained in CD during their formal education , training of staff on current protocols became critical . After the first centres acquired expertise , the Chagas Platform started offering primary care staff a 1-week training program in the centres to learn the protocols . Besides , the training and implementation of operational research included in the model has a relevant role giving to the National Health System personnel elements to analyze and reformulate priorities in Public Health interventions . Finally , as the pilot project proved effective and acceptable , the Chagas Platform healthcare model has been expanded to primary healthcare centres since 2015 , and a new strategy based on the network of centres managing CD was established , with the intention of expanding coverage in diagnosis and treatment in remote areas . The stages of development of the intervention have been reflected in Fig 3 . In addition to the adaptation of the protocols for managing CD in the health system care centres , this new horizontal approach was based on the training of health professionals ( physicians , nurses , and biochemists ) in these protocols and on referrals and counter-referral circuits between primary and specialized care centres .
Results from 2010 to 2015 . Since the implementation of the Chagas Platform in 2009 , a total of 26 , 227 adult patients have been attended in Bolivia , in the Platform centres . Around 69% had T . cruzi infection ( 18 , 316 ) . To date , 8 , 567 patients have started treatment and , on average , 80% have received the complete course . Data regarding coverage of patients with T . cruzi infection are summarized in Table 2 . The number of patients was initially low because only two centres were functioning at the outset . The number of centres increased gradually until 2013 , when the sixth and latest centre was opened . The increased demand has put strain on the system ( organization , logistics , regulation of stocks , appointments ) and the appropriate amount of medication has not always been available . Additionally , a benznidazole stock shortage in 2012 accounted for the low number of persons who started treatment during that year . Apart from the poor availability of drugs , which still limits the number of people who can be treated , the annual gap between persons with a positive diagnosis and persons treated can also be explained by the non-fulfilment of eligibility criteria and less importantly patient’s reluctance to be treated . Strict fulfilment of inclusion criteria has improved adherence to treatment . On average , 80% of persons who initiated treatment finished the 60-day course , and conscientious follow-up ensured that data on adherence were excellent . Around 10% of patients left treatment voluntarily and around 10% of patients were advised to stop treatment owing to adverse drug reactions ( ADRs ) that were partially controlled with symptomatic treatment . Benznidazol and nifurtimox ADRs have been pointed out as one of the main problems for adherence to treatment , and due to the relevance of the topic , the description of them in the context of the Platform model will be described in deep in a separate manuscript . The demand for the Platform was directly related to the implementation of community information activities . Since 2010 , more than 25 , 000 people have received direct information about CD while in the healthcare process at Platform centres . Additionally , more than 3 , 500 people attended community information sessions . More than the half of these sessions was in the original language ( mainly Quechua ) . The percentage of people attending the centres that had T . cruzi infection was higher than expected , and most of those who were treated with benznidazole had excellent adherence to treatment . The latest available data from the ChNP reveal the limitations of the ChNP for covering existing demand: in 2014 , 29 , 052 adults were diagnosed with T . cruzi infection in Bolivia , and only 4 , 444 were treated ( S1 ) . Most of these patients ( 1 , 868 , 42% ) were treated in the Chagas Platform . [23] In the areas where the Chagas Platform has its centres , yearly screening ranges from 0 . 7% to 2 . 1% of the estimated number of T . cruzi infected people . Almost two-thirds of current adult treatment is provided in the Platform centres , thus making them an important support structure for the ChNP in providing adult diagnosis and treatment . Despite the quantitative and qualitative improvement in CD healthcare , the annual number of treated patients is less than 0 . 5% of the estimated total requiring treatment . There is also a considerable gap between people with T . cruzi infection and the number of people treated ( only 10% among all the people with T . cruzi infection diagnosis ) . In this sense , the benefits of the project lie in the fact that it comprises a defined healthcare package , with concrete protocols to manage adults in the chronic phase of CD , that has proven effective and can be expanded to the National Health System . Unfortunately , the total number of patients treated in Bolivia increased by only 14% between 2009 and 2015 . While a favourable trend was observed for adults , the number of newborns and children treated decreased [15] Expanding the model to the primary health care system could reverse this trend . In collaboration with the Bolivian Ministry of Health , this is the next step proposed by CEADES and ISGlobal , with the support of AECID . Since 2010 , more than 1 , 600 health professionals have been trained in the specific management of patients with CD . A training fellowship program has been implemented between Universidad Mayor de San Simon ( Bolivia ) and Barcelona University , and five international professional exchanges have taken place since 2011 . Specialized conferences and specific training activities were held during a conference in 2014 , and more than 500 people participated .
Even if the Chagas Platform has proven to be highly effective , Platform centres have limited human resources to cover current demand in the management of CD in Bolivia . As recommended by the WHO , the main strategy for increasing access to healthcare requires the diagnosis and treatment of CD to be incorporated into the national health system , as part of their regular activities . In Bolivia , the recent establishment of a network including primary healthcare centres in rural areas following easy and realistic protocols is already showing positive results: it enables the management of CD in adults to be standardized and access to healthcare for people living in remote areas to be improved . To date , the integral CD healthcare model used in the Platform centres has been accepted for adaptation by the national health system before being implemented nationwide . Current Platform activities such as prevention , diagnosis , treatment , IEC , and training have been included in the implementation of the proposed model in national primary health centres . The main lessons learned from the implementation are as follows: The increase in the number of people diagnosed and treated must be made carefully in order to avoid excessive strain on the health system , which has to adapt gradually to growing demand . Finally , as CD is a neglected disease , several external factors can hamper patient management . The poor availability of current drugs ( benznidazole and nifurtimox ) and the lack of new and better-tolerated medicines are key limiting factors . Furthermore , the lack of biomarkers of response to treatment hinders research on new drugs [24] . Fortunately , in the last five years , international public and private initiatives have been launched to develop new drugs and to implement clinical trials , as have studies focused on biomarkers of cure and/or response to treatment , some of which have been performed in the Chagas Platform . Despite not being the only valid strategy , the Chagas Platform has proven to be a model of care for patients with T . cruzi infection that has been adopted by the Bolivian government . Moreover , the Chagas Platform has brought the benefits of reinforcing research capacity and training health professionals . Linking care , training , and research at the operational level is a very powerful tool that keeps health professionals updated and motivated . Additionally , the comprehensiveness of the program in the healthcare system should prove useful in other health-related issues . The expertise that has formed the basis of guidelines and protocols is now being expanded to the national health system , thus highlighting the success of the program and enabling diagnosis and treatment to be scaled up .
|
Bolivia has the highest prevalence of Chagas disease ( CD ) in the world ( 6 . 1% ) , with more than 607 , 186 people with Trypanosoma cruzi infection . In Bolivia , the management of CD has been declared a national priority . In 2009 , the Chagas National Program ( ChNP ) had neither a protocol nor a clear directive for diagnosis and treatment of adults . The Chagas Platform has been built as a model for comprehensive care of adults with CD . From 2010 to 2015 , a total of 26 , 227 patients were attended by the Platform , 69% ( 18 , 316 ) were diagnosed with T . cruzi , 8 , 567 initiated anti-parasitic treatment , more than 1 , 616 health professionals were trained . More than ten research projects were developed . The project has also produced evidence-based clinical practice guidelines , and brings about changes in policy that will increase access to comprehensive care among adults with CD . The ChNP is now studying the Platform’s health care model to adapt and implement it nationwide . It is an experience of collaboration and knowledge transfer between endemic and non-endemic countries .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"geographical",
"locations",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"health",
"care",
"organisms",
"age",
"groups",
"adults",
"protozoans",
"pharmaceutics",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"infectious",
"diseases",
"south",
"america",
"health",
"systems",
"strengthening",
"protozoan",
"infections",
"people",
"and",
"places",
"trypanosoma",
"cruzi",
"trypanosoma",
"chagas",
"disease",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"drug",
"therapy",
"health",
"care",
"policy",
"bolivia"
] |
2017
|
A strategy for scaling up access to comprehensive care in adults with Chagas disease in endemic countries: The Bolivian Chagas Platform
|
Depolarization of presynaptic terminals stimulates calcium influx , which evokes neurotransmitter release and activates phosphorylation-based signalling . Here , we present the first global temporal profile of presynaptic activity-dependent phospho-signalling , which includes two KCl stimulation levels and analysis of the poststimulus period . We profiled 1 , 917 regulated phosphopeptides and bioinformatically identified six temporal patterns of co-regulated proteins . The presynaptic proteins with large changes in phospho-status were again prominently regulated in the analysis of 7 , 070 activity-dependent phosphopeptides from KCl-stimulated cultured hippocampal neurons . Active zone scaffold proteins showed a high level of activity-dependent phospho-regulation that far exceeded the response from postsynaptic density scaffold proteins . Accordingly , bassoon was identified as the major target of neuronal phospho-signalling . We developed a probabilistic computational method , KinSwing , which matched protein kinase substrate motifs to regulated phosphorylation sites to reveal underlying protein kinase activity . This approach allowed us to link protein kinases to profiles of co-regulated presynaptic protein networks . Ca2+- and calmodulin-dependent protein kinase IIα ( CaMKIIα ) responded rapidly , scaled with stimulus strength , and had long-lasting activity . Mitogen-activated protein kinase ( MAPK ) /extracellular signal–regulated kinase ( ERK ) was the main protein kinase predicted to control a distinct and significant pattern of poststimulus up-regulation of phosphorylation . This work provides a unique resource of activity-dependent phosphorylation sites of synaptosomes and neurons , the vast majority of which have not been investigated with regard to their functional impact . This resource will enable detailed characterization of the phospho-regulated mechanisms impacting the plasticity of neurotransmitter release .
Depolarization of the presynaptic plasma membrane stimulates the opening of voltage-gated Ca2+ channels , and Ca2+ rapidly enters at the active zone . Increased Ca2+ triggers neurotransmitter release by binding to Ca2+ sensors such as synaptotagmin 1 , which facilitates fast synchronous fusion of docked synaptic vesicles [1–3] . Docking and priming of synaptic vesicles , as well as Ca2+ channel clustering [4 , 5] , are coordinated by scaffold proteins at the active zone . Ca2+ channels are tethered in proximity to synaptic vesicles directly by Rab3-interacting molecules ( RIMs ) [6 , 7] and/or by a complex including bassoon and RIM-binding protein [8] . As such , the protein composition of the active zone scaffold not only is important for vesicle release but also exerts influence on release probability and presynaptic homeostatic plasticity [9] . The influx of Ca2+ following depolarization also stimulates phosphorylation-based signalling . Phosphorylation and dephosphorylation of presynaptic proteins at the active zone are strongly coupled to Ca2+ influx in the vicinity of Ca2+ channels , rather than the cytosolic Ca2+ concentration [10] . Ca2+ binds calmodulin , and this complex activates downstream phospho-signalling pathways that directly affect release properties . For example , Ca2+- and calmodulin-dependent protein kinase II ( CaMKII ) phosphorylation of synapsin 1 has a role in changing the availability of synaptic vesicles for release [11] . Ca2+/calmodulin also activates the phosphatase calcineurin ( protein phosphatase 2B ) and downstream phosphatases . Three synapsin 1 sites—S62 , S67 , and S549—are dephosphorylated by calcineurin and may contribute to increased availability of synaptic vesicles [12] . Calcineurin has also been shown to directly regulate synaptic vesicle endocytosis [13 , 14] . The main protein kinases mediating presynaptic phospho-signalling and plasticity are protein kinase C ( PKC ) [15 , 16] , protein kinase A ( PKA ) [9] , cyclin-dependent kinase 5 ( CDK5 ) [17] , and extracellular signal–regulated kinase 1/2 ( ERK ) /mitogen-activated protein kinase 1/3 ( MAPK1/3 ) [18] . However , the identity of the presumably large number of substrates , the timing and interdependence of protein kinase activation or inactivation , and mechanistic consequences remain largely unresolved , as the complex activity-driven network of presynaptic kinases has previously not been explored systematically , to our knowledge . There are few global studies of activity-dependent neuronal phospho-signalling . The postsynaptic density has been enriched to reveal postsynaptic signalling [19 , 20] . Here , we use a subcellular compartment enrichment strategy to focus on presynaptic proteins . Sucrose homogenates of brain tissue produce a particulate fraction that contains both pre- and postsynaptic components but is primarily enriched in isolated presynaptic terminals , i . e . , synaptosomes [21 , 22] . Synaptosomes contain synaptic vesicles [22] , are metabolically active , generate adenosine triphosphate [23] , and maintain a membrane potential [24] and calcium homeostasis [25] . Thus , synaptosomes are a highly functional model for studying presynaptic phospho-signalling in isolation . Depolarization of synaptosomes with KCl stimulates concentration-dependent Ca2+ influx [25 , 26] . A subset of the targets of presynaptic phospho-signalling , which are dependent on the initial Ca2+ influx , has been identified [27] . However , the broader and functionally important question of how synaptic phospho-signalling develops following a period of neuronal activity remains to be addressed . The poststimulus period has only been examined with very limited scope for a few well-characterized phosphorylation sites in synapsin 1 and MAPK [28 , 29] . Such transient and persistent changes in the presynaptic phosphoproteome following activity are of crucial importance , as they are likely tightly linked to several forms of synaptic plasticity , which ensure that activity in neuronal networks remains within physiological limits and provide the basis for information storage , learning , and memory . Here , we used quantitative phosphoproteomics to extensively define the targets and regulators of presynaptic phospho-signalling by examining both the stimulus and poststimulus periods . Exploration of the data revealed distinct temporal patterns of regulation . The relevance of our data is further underscored by the finding that cultured hippocampal neurons exhibited a similar activity-dependent response and confirmed the identity of several predominantly presynaptic phospho-signalling integrator proteins . A new computational method , KinSwing , was used to determine the profile of relative protein kinase activity , based on protein kinase substrate prediction . The inferred protein kinase activity was matched to the patterns of presynaptic phospho-regulation , revealing that a poststimulus up-regulation of phosphorylation was likely to be substantially mediated by MAPK/ERK . Proteins regulating vesicle release were prominent substrates of poststimulus up-regulation of phosphorylation .
Synaptosomes were stimulated with 20 mM KCl or 76 mM KCl ( half of total monovalent salt ) —or mock treated by keeping 4 . 7 mM KCl—for 10 s and subsequently returned to 4 . 7 mM KCl to enable repolarization . We confirmed the acute nature of this stimulation protocol by showing that the depolarization only acted during the 10-s period ( S1 Fig , S1 Data ) . The synaptosomes were lysed at specific times to monitor changes to the phosphoproteome—i . e . , at the end of the depolarization ( 10 s ) and at three poststimulus time points ( 90 , 300 , and 900 s ) ( Fig 1A ) . Changes in phosphorylation levels relative to the mock stimulation over time were determined using a global quantitative phosphoproteomics workflow [30] ( Fig 1A and Materials and methods ) . The synaptosome data were processed and interrogated using a robust bioinformatics approach . Only high-confidence phosphopeptides with a probability score ≥ 0 . 75 for phosphorylation site assignment detected in at least three of the six biological replicates were used for further analyses . Our statistical workflow ( Fig 1A and Materials and methods ) involved normalization , missing value imputation , and correction for nonbiological sources of variation , as described [31 , 32] . Unsupervised principle component analysis ( PCA ) of all phosphopeptides separated the time points across the first principle component , indicating that phosphorylation of peptides was largely occurring in a time-dependent manner ( Fig 1B ) . A total of 5 , 715 unique phosphopeptides were quantified in the intersection of 20 and 76 mM KCl stimulation experiments from 1 , 825 proteins ( not counting isoforms and multiprotein identifications ) ( Fig 1C and S1 Table ) . A total of 1 , 917 phosphopeptides were significantly up-/down-regulated over time and detected in both the 20 mM and 76 mM KCl conditions . Significance across time was determined using a moderated F-statistic adjusted for multiple hypothesis testing . P < 0 . 05 was required for inclusion in the set of 1 , 917 phosphopeptides . The phosphoproteome was perturbed more by 76 mM KCl stimulation , when compared to 20 mM KCl , and at earlier time points . At 10 s , there were approximately double the number of significant changes resulting from 76 mM KCl stimulation compared to 20 mM KCl ( S2A–S2C Fig; a moderated t-statistic adjusted for multiple hypothesis testing was used to determine the significance of single time points , P < 0 . 05 ) . This supports the view of a graded response of the presynaptic phosphoproteome to different levels of stimulation . At 300 and 900 s , significant phospho-signalling persisted , but the difference in number of significantly regulated phosphorylation sites between 20 mM and 76 mM KCl stimulation was smaller ( S2A–S2C Fig ) . Thus , perturbation of phospho-signalling was found to be long-lasting as a consequence of acute stimulation . To provide a resource of broad utility , we also stimulated cultured hippocampal neurons with 76 mM KCl for 10 s and quantified changes in phospho-signalling using our bioinformatics approach ( Fig 1A ) . This resulted in the confident identification of 22 , 063 unique phosphopeptides ( Fig 1D and S1 Table ) , of which 7 , 070 were significantly regulated by the stimulation ( moderated t-statistic adjusted for multiple hypothesis testing , P < 0 . 05 ) . Only 4 . 6% of these activity-dependent sites have a known function or regulatory role , and only four of the top 100 largest magnitude significant changes have been explored up to now ( PhosphoSitePlus , [33] ) . A similarly small fraction of activity-independent phosphorylation sites ( 4 . 8% ) identified have a known function or regulatory role . This indicates a very large gap in the knowledge of signalling mechanisms dependent on neuronal activity . A total of 3 , 549 phosphopeptides were identified in both neurons and synaptosomes ( intersection ) . An intersection of 195 phosphopeptides from 123 proteins were significantly regulated after 10 s of 76 mM KCl stimulation in synaptosomes and neurons . These phosphopeptide signals were well correlated ( R2 = 0 . 51 , S2D Fig , S1 Data ) . To avoid comparison of phosphoproteins that occur in multiple subneuronal compartments , we focused on active zone scaffold and synaptic vesicle–associated proteins , which are highly specific to the nerve terminal and key components of the release machinery ( listed in S1 Table ) , thereby excluding postsynaptic proteins by default . This presynaptic-focused correlation analysis is shown in Fig 2A ( S1 Data ) . Active zone scaffold proteins ( bassoon; piccolo; RIM1; liprin-α3; protein rich in E , L , K , and S 1 [ELKS1]; corresponding to gene names: Bsn , Pclo , Rims1 , Ppfia3 , and Erc1 ) , synaptic vesicle endocytosis-specific clathrin uncoating protein auxilin 1 ( Dnajc6 ) , and presynaptic-specific cytoskeletal organizing protein tau ( Mapt ) were among those proteins with highly correlated activity-dependent phospho-signalling ( R2 = 0 . 56 , Fig 2A , S1 Data ) . The high correlation between the activity-induced phospho-signalling in synaptosomes and cultured hippocampal neurons indicates that signalling pathways in synaptosomes were well preserved . Eight of the top 100 largest changes from hippocampal neurons were sites from bassoon ( Fig 2B ) . Bassoon appears to be the major target of neuronal activity-dependent phospho-signalling and is targeted for phosphorylation and dephosphorylation along its entire lengthy sequence ( Fig 2B ) . In contrast , the postsynaptic density proteins and glutamate receptors had a relatively modest response to stimulation ( Fig 2C ) . A formal comparison of the number of significantly regulated phosphorylation sites against protein length revealed that bassoon did indeed have one of the highest numbers of regulated phosphorylation sites per amino acid residue . There was no correlation between number of regulated phosphorylation sites and length ( Fig 2D , S1 Data ) . The same lack of correlation was observed for the regulated phosphorylation sites from hippocampal neurons ( Fig 2D ) . Presynaptic substrates , such as RIM1 and microtubule-associated protein 1B ( MAP1B; gene names: Rims1 and Map1b ) featured among the high-magnitude changes ( S1 Table and S1 Data ) , indicating that activity-dependent phospho-signalling has a pronounced effect on a subset of presynaptic substrates . The vesicle-tethering protein synapsin 1 is the only presynaptic protein that has been substantially examined during the poststimulus period [28] . We identified 26 significantly regulated synapsin 1 phosphopeptides from synaptosomes . Synapsin 1 S566 , a CaMKIIα substrate site from in vitro studies [35] , was the most up-regulated site at 10 s ( Fig 3A ) . Another CaMKIIα substrate and activity-dependent site , S603 [12] , was also up-regulated at 10 s ( Fig 3A ) , and this was confirmed by western blot ( Fig 3B , S1 Data ) using 76 mM KCl stimulation . Phosphorylation of these D domain sites is expected to promote the dispersion of the recently identified synapsin liquid phase and associated vesicles [36] . Phosphorylation was generally up-regulated for the vesicle binding A domain and the E domain of synapsin 1 . The phospho-regulation patterns were complex for domains B and D that contain evolutionarily conserved low complexity and disordered sequences , as identified by the Pfam ( protein families ) database [34] ( Fig 3C ) . In agreement with a previous report [28] , we observed bidirectional regulation of S549 and S62+S67 in our synaptosome data for both 20 and 76 mM KCl stimulations ( Fig 3A ) . Another well-studied presynaptic substrate , dynamin 1 [27 , 37 , 38] , was down-regulated at major activity-dependent phosphosites S774 and S778 at 10 s ( Fig 3A and 3B ) . Thus , our synaptosome phosphoproteomics data reproduce known stimulus and poststimulus phospho-specific profiles , verifying the validity of our stimulation and repolarization paradigm . Furthermore , the majority of synapsin 1 phosphorylation sites had the same direction of regulation after 76 mM KCl stimulation in both synaptosomes and hippocampal neurons ( 16 of 19 significantly regulated at 10 s , Fig 3A ) . Western blotting confirmed up-regulated phosphorylation at S603 and down-regulated S62+S67 phosphorylation ( Fig 3A , boxed ) , which were highly correlated phosphorylation sites in Fig 2A ( red circles ) . To determine which groups of proteins and/or biological processes had similar activity-dependent temporal phospho-signalling , the 1 , 917 significantly regulated phosphopeptides from synaptosomes were k-means clustered across time for both concentrations of KCl ( Fig 4A and S1 Table ) . Phospho-regulated postsynaptic proteins were not filtered from the data prior to the clustering , or for any analysis in this work , except where stated ( Fig 2A ) . We rely on the knowledge that , although small intact postsynaptic compartments are present in synaptosome preparations , the predominant component is the isolated presynaptic terminal [22] . Six clusters were determined as optimal for enrichment analysis ( see S3A and S3B Fig , S1 Data ) . Clusters were summarized as ( 1 ) high-magnitude up-regulated , ( 2 ) high-magnitude down-regulated , ( 3 ) poststimulus down-regulated , ( 4 ) poststimulus up-regulated , ( 5 ) low-magnitude down-regulated , and ( 6 ) low-magnitude up-regulated , represented by a small stylized line graph adjacent to the heat map in Fig 4A . These line graphs are representative of the overall 20 mM and 76 mM KCl stimulated patterns . Fig 4B and 4C ( S1 Data ) show the sum of log2 ( stimulated intensity/control intensity ) for each cluster versus time for the 20 and 76 mM KCl stimulations and are provided to allow a comparison of the cluster trends using accurate line graphs . Cluster 4 phosphorylation was , overall , slightly down-regulated at 10 s and highly up-regulated at 300–900 s after 20 mM KCl ( Fig 4A–4C ) . After 76 mM KCl , the down-regulation at 10 s was greater in magnitude , and there was less up-regulation at 300–900 s . Cluster 2 phospho-regulation also exhibited an overall slight bidirectionality; a change in direction occurred for some phosphopeptides from 90 to 300 s after 20 mM KCl ( Fig 4A–4C ) . The pattern changed to down-regulation at all time points after 76 mM KCl . Cluster 1 and 2 had high-magnitude initial regulation that continued into the poststimulus , and the regulation extended further after 76 mM KCl ( Fig 4A–4C ) . Cluster 1 up-regulation outlasted cluster 2 down-regulation for at least 10 min of the time course after 20 mM KCl ( Fig 4A and 4B ) . Cluster 3 phosphorylation sites mainly failed to respond at 10 s but were down-regulated throughout the poststimulus , and this was independent of the strength of stimulus ( Fig 4A–4C ) . This implies a temporally controlled signalling mechanism that specifically changes the poststimulus balance of phosphorylation and dephosphorylation independent of the level of Ca2+ influx . Cluster 5 was similarly disconnected from the stimulus strength in the poststimulus but at a relatively low level of down-regulation . Low-magnitude up-regulated cluster 6 had more phosphopeptides up-regulated at 300 s after 76 mM KCl ( Fig 4A–4C ) . Thus , clustering enabled the identification of six specific patterns of phosphopeptide regulation , including high-magnitude up-regulation > 5 min post stimulus , and some were highly influenced by the stimulus strength . Each cluster was subsequently investigated for gene ontology enrichment . The cellular component term “presynaptic active zone” was most enriched in cluster 2 , followed by clusters 1 , 4 , and 5 . There was a lack of enrichment in cluster 3 ( Fig 5A , S1 Data ) . Clusters showing enrichment contained active zone proteins such as bassoon , piccolo , RIM1 , liprin-α3 , ELKS1 , and ELKS2 [5] . High-magnitude down-regulated cluster 2 was also enriched with the “synaptic vesicle” term ( Fig 5A ) and contained synapsin 1 and endocytic proteins known to be substrates of calcineurin [12 , 13] . Multiple members of the microtubule-binding collapsin response mediator protein family were specifically enriched in cluster 1 , resulting in inclusion of the “hydrolase activity , acting on carbon-nitrogen ( but not peptide ) bonds” molecular function term in Fig 5B ( S1 Data ) . The “protein serine/threonine kinase activity” molecular function term was enriched for clusters 1 and 4 . The “protein kinase binding” term was enriched for up-regulated clusters 1 and 6 and somewhat enriched for down-regulated clusters 2 and 5 ( Fig 5B ) . Poststimulus up-regulated cluster 4 was enriched with the molecular function term “microtubule binding” ( Fig 5B ) . Low-magnitude down-regulated cluster 5 was most enriched for “GTPase activator activity” . Low-magnitude up-regulated cluster 6 was the only cluster enriched for “voltage-gated potassium channel activity” ( Fig 5B ) and the biological process term “potassium ion transmembrane transport” ( Fig 5C , S1 Data ) . The term “neurotransmitter secretion” was enriched for clusters 1 , 2 , 4 , and 6 . Clusters 2 and 6 were enriched for “exocytosis” ( Fig 5C ) . Cluster 1 was enriched for the “microtubule cytoskeleton organization” term ( Fig 5C ) . Cluster 3 lacked significant enrichment for most terms , except “cytoskeletal organization” ( Fig 5C ) . In general , higher-magnitude clusters 1 , 2 , and 4 revealed phospho-signalling that targeted active zone scaffold components , protein kinases , the collapsin response mediator protein family , and vesicular and cytoskeletal-related proteins , which may influence neurotransmitter release , the synaptic vesicle cycle , and regulation of the cytoskeleton/microtubules . The enriched gene ontology terms lacked cluster specificity ( Fig 5A–5C ) . The enrichment analysis may have been undermined by proteins participating in multiple distinct signalling pathways . These proteins contained phosphorylation sites spread across multiple clusters . Piccolo , bassoon , and MAP1B ( gene names: Pclo , Bsn , and Map1b ) had phosphopeptides represented in all six clusters . Cytoplasmic linker–associated protein 2 ( Clasp2 ) as well as synapsin 1 and 3 were represented in five of six clusters . The number of regulated phosphorylation sites for each protein was plotted against the number of clusters of which each protein was a member in S4A Fig . This enabled the identification of proteins highly represented in greater than four clusters . Further separation of proteins was achieved by using the largest magnitude change of log2 intensity as a multiplier of cluster number ( S4B Fig , S1 Data ) . Tau ( Mapt ) , SNAP25-interacting protein ( Srcin1 ) , and RIM1 ( Rims1 ) were identified as proteins represented in four of six clusters , with relatively high-magnitude responses and numbers of significantly regulated phosphorylation sites ( S4A and S4B Fig , S1 Data ) . In contrast , spectrin beta chain ( 273 kDa , Sptbn1 ) had 14 regulated phosphopeptide signals that were present in only two clusters , 13 of which were in lower-magnitude cluster 5 ( S1 Table ) . As shown in Fig 2D ( blue-coloured gene names ) , the same proteins we singled out for representation in multiple clusters ( S5A–S5F Fig ) had higher numbers of regulated sites per length but were not required to be large proteins . Representation in many clusters suggests participation in multiple phospho-signalling pathways , which can potentially be independently modulated and impact multiple downstream functions . These highly phospho-regulated proteins are important presynaptic signal integrators and indicate that presynaptic protein functions can be subject to diverse phospho-signalling pathways . Since phosphorylation site functions do not relate directly to ontology terms at the gene level ( Fig 5A–5C ) , we investigated ontology at the level of phosphorylation sites . Publicly available curated descriptions of the regulatory role of specific phosphorylation sites are available from PhosphoSitePlus [33] . We determined the enrichment of these phospho-regulation terms for each cluster ( Fig 5D , see Materials and methods ) . Most clusters exhibited a different pattern of enrichment . However , the low number of phosphorylation sites associated with phospho-regulation terms ( <5% , see S1 Table ) limited the power of this analysis . Significant fold enrichment was determined in poststimulus up-regulated cluster 4 for “intracellular localization” and “cytoskeletal reorganization” terms ( P < 0 . 05 , Fig 5D ) . The process of cytoskeletal reorganization was collectively regulated by specific phosphorylation sites in tau ( Mapt ) , phosphatidylinositol 4-phosphate 5-kinase type-1 γ ( Pip5k1c ) , insulin receptor substrate p53 ( IRSp53; Baiap2 ) , Septin 7 ( Sept7 ) , and exocyst complex component 7 ( Exoc7 ) ( S1 Table ) . Thus , poststimulus phospho-regulation of the cytoskeleton was a significant process identified in the pattern of cluster 4 . The clusters of phosphopeptides in Fig 4A were first analyzed for protein kinase substrate motif enrichment using motif-x [40] . The CaMK protein kinase family RXXS motif was enriched for cluster 1 ( S6A Fig ) . Proline-directed motifs , with proline in the +1 position relative to the phosphorylation site , were mainly enriched for down-regulated clusters . The PXSP motif associated with MAPK substrates [41] was enriched in cluster 4 ( S6A Fig ) . These results indicate that particular families or classes of protein kinases may be associated with the clusters of regulated phosphorylation sites . However , this approach lacks the sophistication required to narrow the prediction to specific protein kinases . Ideally , we would identify and determine the contribution of the major protein kinases at specific time points . To achieve this , we developed a computational method named KinSwing . KinSwing is a statistical approach that integrates known protein kinase substrate motifs , probabilistic matching to substrate sequences , significance of phospho-regulation , and direction of phospho-regulation over time to infer the contribution of specific protein kinases to phospho-signalling ( Fig 6A , see Materials and methods ) . KinSwing does not require the data to be clustered a priori . Instead , KinSwing determines protein kinase contribution at each time point , which can be subsequently used for clustering . First , position weight matrices , representing the frequency of occurrence for each amino acid residue , at and adjacent to the phosphorylation site , were determined for 355 protein kinases using experimentally verified substrate sequences described in the PhosphoSitePlus database [33] . The probability of a significant motif match for each phosphopeptide in our study was then determined and used to identify substrate sequences . The substrates that either significantly increased or decreased in phosphorylation , at each time point , were assigned a positive or negative “swing” ( Fig 6A ) . An overall sum of “swings” was calculated , with consideration for the number of substrate sequence matches , such that the sum for each protein kinase was weighted and transformed into a z-score , enabling relative comparisons between protein kinases and across time ( see Materials and methods ) . The z-score will henceforth be referred to as the KinSwing score ( KS score ) . KinSwing aims to predict the positive or negative inferred activity for a specific protein kinase . Eighty-seven KS score profiles were clustered ( Fig 6B ) after removing protein kinases based on consensus sequences with limited and highly variable scores ( see Materials and methods and S1 Table ) . Clusters are represented as line graphs of averaged KS scores in Fig 6C ( S1 Data ) . It is important to appreciate that protein kinase activity should be considered in light of opposing protein phosphatase activity . Since substrate motifs for protein phosphatases lack specificity relative to protein kinase substrates and protein kinases have a more diverse impact on cellular pathways [42] , we take a protein kinase–centric view of activity . Also , KinSwing cannot resolve the activities of protein kinases and protein phosphatases at the subcellular level when applied to phosphoproteomics data from whole cells . KS profiles allowed the identification of the major regulatory protein kinases and their patterns of regulation . CaMKIIα ( CAMK2A ) clustered alone ( Fig 6B and 6C , cluster vii ) and was the only protein kinase predicted to have increased activity after 76 mM KCl stimulation , relative to 20 mM KCl stimulation . Many protein kinases contain regulatory phosphorylation sites , which could be utilized to validate activity inferred by KinSwing . There were 23 and 68 significantly perturbed protein kinase regulatory sites identified from synaptosomes and hippocampal neurons , respectively ( S1 Table , regulatory sites listed ) . The T286 autoactivation site and the T306 inhibitory site for CaMKIIα were up-regulated and down-regulated , respectively , at 10 s in hippocampal neurons but were not detected in synaptosomes using mass spectrometry ( S7A Fig ) . Western blotting demonstrated that T286 was indeed significantly up-regulated at 10–300 s in synaptosomes ( S7A Fig , S1 Data ) . In support of increased protein kinase activity , the CaMKIIα S275 putative autoactivation site [43] was significantly up-regulated at all time points in synaptosomes by mass spectrometry ( S7A Fig ) . CaMKIIα binds the postsynaptic density and to synaptic vesicles . In S7B Fig ( S1 Data ) , we compared the regulatory patterns and probabilities of confident CaMKIIα , PKAα , and PKCα substrate motif matches for the pre- and postsynaptic substrates in Fig 2C . Disks large-associated protein 4 ( Dlgap4 ) and SH3 and multiple ankyrin repeat domains protein 3 ( Shank3 ) had the largest up-regulation of those predicted as CaMKIIα substrates by KinSwing ( S7B Fig ) . Known CaMKIIα substrate glutamate receptor 1 S849 ( S831 in mature protein , gene name: Gria1 ) was weakly up-regulated in hippocampal neurons ( S1 Table ) but was above the probability threshold for prediction as a CaMKIIα substrate in the KinSwing analysis . Of the presynaptic proteins in Fig 2C , RIM1 , piccolo , and known substrate synapsin 1 ( S5A , S5D and S5E Fig ) were supported by probability as CaMKIIα substrates ( S7B Fig ) . Bassoon had many long-lasting up-regulated phosphorylation sites , which can be considered as putative CaMKIIα substrates ( Fig 2B , synaptosome data ) , but only S2845 and S1126 were supported by probability as CaMKIIα substrates ( P < 0 . 05 ) . Likely , many more substrates exist in nonlinear motifs , defying prediction . In these cases , the pattern of regulation may be useful for predicting kinase substrate relationships . Not all putative CaMKIIα substrates were up-regulated or regulated in the same direction on the same protein , indicating incorrect prediction or site-specific effects of CaMKIIα or opposing protein phosphatase activity that depended on localization or other biological factors . Nevertheless , correlation of protein kinase activity patterns , via KinSwing and regulatory sites , with substrate phosphorylation patterns allowed for the identification of approximately 20 putative CaMKIIα substrates . Clusters vii , viii , and ix were exclusively made up of CaMK and protein kinase A , G , and C ( AGC ) family members and were generally up-regulated at 10 s and 300 s after 20 mM KCl ( Fig 6B and 6C ) . The PKAα subunit ( PRKACA ) was the most up-regulated at 300 s after 20 mM KCl , but unlike CaMKIIα , its inferred activity was much reduced with increased stimulation strength ( 76 mM KCl ) . PKCα ( PRKCA ) in cluster ix underwent the most rapid reversal from positive to negative KS score from 10 s to 90 s after 20 mM KCl . Another pattern was that the PKA , CaMK-like ( CaMKL ) , sterile 20 ( STE20 ) , and ribosomal S6 kinase 1 ( RSK ) classes of protein kinases had positive inferred activity after 20 mM KCl stimulation that was diminished or ablated after 76 mM KCl stimulation ( S6B Fig ) . The PKC class separated into its own cluster ( S6B Fig ) and was not as affected by the stimulation level as PKCα individually ( Fig 6B ) . PKCζ ( PRKCZ ) had a stimulation level–independent negative KS score for all poststimulus time points ( Fig 6B and 6C cluster v ) , which most resembled the down-regulated phosphopeptide clusters 3 and 5 in Fig 4A . Thus , the inferred activity of some subclasses of the CaMK and AGC families were up-regulated at particular times and were dependent on the strength of the stimulus . The CDK , glycogen synthase kinase 3 ( GSK3 ) , MAPK , and CLK ( CGMC ) family members CDK1/2/5 , GSK3α/β , and MAPK1/3/8/14 in clusters i and ii ( Fig 6B and 6C ) had the most negative KS scores at 10 s , implying rapid deactivation of CMGC kinases and/or dephosphorylation of the putative CGMC substrates . GSK3β S9 up-regulation was detected at 10 s in hippocampal neurons , which inhibits GSK3β [44] . Protein kinase Bα ( AKT1 ) , which targets GSK3β S9 , also had corresponding up-regulated phosphorylation at S129 , which induces activity ( S1 Table ) in agreement with a positive AKT1 KS score at 10 s ( Fig 6B ) . However , GSK3β KS scores were negative at all times after 76 mM KCl , and this did not correlate with increased AKT1 KS scores . Higher protein phosphatase activation may have had a greater effect on GSK3β substrates than AKT1-mediated GSK3β inactivation after 76 mM KCl . At 300 s after 20 mM KCl , MAPK1 had the highest inferred activity of cluster ii ( Fig 6B ) . In contrast , at 300 s after 76 mM KCl stimulation , the positive KS score was ablated , indicating dominance of protein phosphatases in both stimulus and poststimulus periods following 76 mM KCl stimulation . KS profiles implicate cluster ii protein kinases , particularly MAPK1 in the poststimulus up-regulated cluster 4 pattern in Fig 4A . In summary , several patterns of inferred protein kinase activity emerged . CaMKIIα was active under all conditions , aligning with the high-magnitude up-regulated cluster 1 in Fig 4A . A subset of CMGC protein kinases , including MAPKs , CDKs , and GSK3α/β , had negative KS scores at 10 s ( Fig 6B ) , aligning with the high-magnitude down-regulated pattern of Fig 4A cluster 2 . Poststimulus activity of MAPKs , PKA , and other protein kinases could contribute to the recovery of cluster 2 substrates or the poststimulus up-regulated pattern of cluster 4 in Fig 4A . Many protein kinase contributions were ablated after stronger stimulus . Thus , our newly developed KinSwing analysis identified protein kinases potentially contributing to the patterns of regulated phospho-signalling . Using our activity-dependent phosphoproteome of hippocampal cultured neurons , we independently calculated KS scores for each protein kinase ( Fig 6B ) . The response at 10 s after 76 mM KCl stimulation was highly correlated with that of synaptosomes ( Fig 6D , R2 = 0 . 81 ) . Casein kinase 2α ( CK2α; CSNK2A1 ) was the only protein kinase with substantial deviation between synaptosomes and hippocampal neurons ( Fig 6D ) . CK2α inferred activity also deviated greatly between 20 mM and 76 mM stimulations , resulting in a unique cluster ( cluster iii , Fig 6B and 6C ) . Interpretation of the inferred CK2α activity requires further experiments . Overall , the high correlation supports the concept that the synaptosomes and hippocampal neurons had a similar signalling network . To guide our spatial understanding of phospho-signalling at the site of vesicle release and visualize activity-dependent protein targets in the stimulus and poststimulus periods , we generated curated protein interaction networks of the active zone and periphery for each of the clusters determined in Fig 4A . In Fig 7A and 7B , we mapped the protein networks of clusters 1 and 2 . Proteins were represented by their gene names , and the letter size was scaled by the highest up- or down-regulated change at any time point to emphasize strongly responding proteins . The blue letters indicate proteins with ≥3 represented phosphorylation sites . Active zone scaffold proteins , synapsins , and cytoskeletal and endocytic proteins were rapidly responding targets of activity-dependent phospho-signalling in clusters 1 and 2 ( Fig 7A and 7B ) , of which a subset were identified previously as activity-dependent [27] . Up-regulated cluster 1 also included strongly responding protein kinases . This included activity-inducing phosphorylation of CaMKIIα ( S7A Fig ) and MAPK1 ( Fig 7A ) . To allow a comparison of the proteins affected by each cluster , substrates identified in down-regulated cluster 3 ( S7C Fig ) , low-magnitude down-regulated cluster 5 ( S8A Fig ) , and low-magnitude up-regulated cluster 6 ( S8B Fig ) were also presented as presynaptic protein interaction networks . Each cluster was also mapped separately onto the same , unfiltered protein network in S1–S6 Files . In each case , the average poststimulus log2 intensities were used to scale the gene name letter size . Activity-dependent vesicle and mitochondrion transport proteins syntabulin ( Sybu ) [45] and syntaphilin ( Snph ) [46] featured in the cluster 3 network as possessing multiple down-regulated phosphorylation sites ( S7C Fig ) . Multiple voltage-gated Ca2+ channel subunits had down-regulated phosphorylation sites in cluster 5 ( S8A Fig ) , and down-regulated phosphorylation was frequent on many cytoskeletal proteins . Voltage-gated K+ channel subunits and synaptic vesicle and exocytosis proteins had up-regulated phosphorylation in cluster 6 ( S8B Fig ) , accounting for the enrichment of related gene ontology terms in Fig 5A–5C . The poststimulus up-regulation observed for cluster 4 is of particular interest because the pattern of regulation implies a late but significant change in phospho-regulated protein function ( Fig 5D ) . Cluster 4 substrates are displayed on the curated protein interaction network in Fig 8A . Letter size was mapped to the largest difference in magnitude from 10 s to 300 or 900 s after 20 mM KCl to identify the largest up-regulated phosphorylation relative to early time points . Motif analysis ( S6A Fig ) and KinSwing ( Fig 6B ) predicted that MAPK is likely to have phosphorylated cluster 4 substrates . In addition , a MAPK1 activation site was up-regulated at all poststimulus time points ( Fig 8B , bottom ) . Thus , we compared the probability of a confident match for cluster 4 substrates to the MAPK1 substrate motif , as a heat map in Fig 8B ( S1 Data ) , using the first step of the KinSwing process to generate probabilities ( Fig 6B ) . Alongside , we also compared GSK3β and CDK5 , which co-clustered with MAPK1 , and PKAα , which was the most up-regulated at 300 s after 20 mM KCl ( Figs 6B and 8B ) . Thus , we present the phospho-signalling network for cluster 4 and identify known and potential substrates of MAPK1 and other protein kinases . Phosphorylation of ( the protein expressed from ) Pip5k1c at S552 ( Fig 8B ) promotes its role in the formation of focal adhesions [47] . B-Raf ( Braf ) phosphorylation at S134 is known to be up-regulated in a feedback loop within the MAPK pathway [48] and was up-regulated in cluster 4 ( Fig 8B ) . Phosphorylation of two sites on IRSp53 ( Baiap2 ) by a CaMKL family protein kinase Par1b ( Mark2 ) , itself a cluster 4 member , may negatively regulate cell polarity and spreading [49] . The calcium-activated potassium channel subunit alpha-1 ( Kcnma1 ) was up-regulated from 90 s at S712 , which is known to cause reduced probability of channel opening and is expected to be a substrate of PKC [50] . These phosphorylation sites with known functions in diverse processes can now be put into the context of neuronal activity through determination of their patterns of phospho-signalling . In contrast to Fig 7A , in which all synapsins were prominent , only synapsin 1 had cluster 4 phosphorylation sites ( Fig 8A ) . Synapsin 1 domain B sites S62 and S67 ( Fig 8B ) and S549 in domain D ( Fig 3C ) are known MAPK substrates [12 , 28] . MAPK phosphorylation of synapsin 1 limits neurotransmitter release [18] . Active zone scaffold proteins were major targets within the rapidly responding clusters 1 and 2 ( Fig 7A and 7B ) and also within the poststimulus up-regulated phosphorylation of cluster 4 , having four proteins with at least three represented phosphopeptides ( Fig 8A ) . However , a different set of proteins and phosphorylation sites were prominent in cluster 4: liprin-α3 ( Ppfia3 ) , Munc13-1 ( Unc13a ) , and Ca2+- and calmodulin-dependent serine protein kinase-interacting protein 1 ( Caskin1 ) . This indicates that a poststimulus phospho-signalling pathway targets the active zone scaffold using a distinct subset of protein components and activity-dependent sites . Phosphorylation of S2845 on bassoon promotes 14-3-3 binding and changes bassoon molecular exchange rates with the proposed effect of dissociation from the cytomatrix at the active zone [51] . S2845 is an in vitro RSK family substrate [51] , and the RSK family had a pattern of inferred regulation that was consistent with cluster 4 ( S6B Fig ) . Dynamin 1 was poststimulus up-regulated at S857 , a phosphorylation site that is de-enriched at presynaptic terminals [38] . The up-regulation of dynamin 1 phosphorylation is inhibitory for synaptic vesicle supply [17] , which could promote a depression of vesicle release . The dynamin 1 phospho-dependent binding partner syndapin 1 ( Pacsin1 ) [52] was among a group of endosomal proteins that were prominent in cluster 4 ( Fig 8A ) but not cluster 1 and 2 ( Fig 7A and 7B ) . The phosphorylation of Munc18-1 ( Stxbp1 ) on Y473 has been shown to inhibit neurotransmitter release [53] and was one of three Munc18-1 sites with poststimulus up-regulation . Two other cluster 4 proteins were associated with the vesicle fusion machinery ( Fig 8A , Syt7 and Nsf ) . Thus , MAPK1 and other protein kinases are implicated as regulators of cluster 4 . MAPK1 may be upstream of protein kinases in cluster 4 , but this is currently only supported by motif prediction ( Fig 8B ) . Many substrate sites in Fig 8B have not been functionally investigated; however; the known roles of bassoon , dynamin 1 , synapsin 1 , and Munc18-1 implicate the poststimulus up-regulation of cluster 4 in influencing the synaptic vesicle cycle .
This first study of the temporal dynamics of the presynaptic phosphoproteome has led to a number of new insights and provides a valuable resource for future analyses . Firstly , it is now apparent that activity-dependent poststimulus phospho-signalling is complex and long-lasting and targets distinct protein domains and networks of phosphoproteins across time . Secondly , initial activity-dependent phospho-signalling in synaptosomes closely correlates with signalling in cultured hippocampal neurons , despite the more numerous compartments harvested with the latter . Thirdly , we have identified specific protein kinases that mediate the initial and poststimulus response , which we have analyzed at unprecedented depth . This achievement was facilitated by our development of a new computational method , KinSwing . These insights greatly advance our knowledge about the stimulus-coupled dynamics of phospho-signalling at the nerve terminal . A key feature of the KinSwing method was to enable a comparison of the inferred protein kinase activity at each time point and condition . The underlying protein phosphatase activity was also indicated . For example , MAPK1 was highly activated by the stimulus , but its substrates were mainly dephosphorylated prior to 300 s . Although KinSwing weights kinase activity by the number of substrates used to build a kinase model , the kinase substrate matching and the extent of representation of the kinome in the output is limited by the availability of protein kinase substrate data . This can be important when considering substrates that are targeted by closely related protein kinases with limited a priori specificity data [33] . In addition , KinSwing best leverages large-scale experimental data that are rich in repeated patterns of phospho-regulation . However , the value of KinSwing and similar approaches will scale well as the size of public phosphorylation site databases increases . Thus , KinSwing can be applied to any phosphoproteomics study , including any neuronal stimulation paradigm amenable to biochemical analysis , and demonstrates the value of using and developing probabilistic methods for future phosphoproteomics studies . KinSwing identified CaMKIIα as a protein kinase with a large change in inferred activity across the time course that scaled with stimulation strength and CaMKIIα activity correlated with the profile of high-magnitude up-regulated phosphorylation ( Fig 4A , cluster 1 ) . CaMKIIα activation is known to enhance neurotransmitter release [54] , but very few presynaptic substrates and mechanisms are known . The putative substrates identified here in the active zone scaffold , including bassoon , piccolo , liprin-α3 , and RIM1 , might participate in phospho-regulated mechanisms that promote vesicle release and presynaptic plasticity . Calcineurin is the primary candidate for the observed high-magnitude dephosphorylation ( Fig 4A , cluster 2 ) , since known substrates associated with synaptic vesicles [12 , 13] were enriched in this phospho-signalling pattern ( Fig 5A ) . Protein phosphatase 2A ( PP2A ) is also a candidate , since it can undergo Ca2+-dependent activation [55] . The high-magnitude up-regulation ( cluster 1 ) lasted longer than the high-magnitude dephosphorylation ( cluster 2 , Fig 4A–4C ) . This observation fits with the knowledge that calcineurin or PP2A are dependent on elevated calcium , but CaMKIIα becomes autoactivated [56] . Stronger stimulation increased the magnitude of dephosphorylation for the initially down-regulated patterns ( Fig 4A–4C , clusters 2 and 4 ) . This resulted in a delayed return to prestimulus levels and a dampened poststimulus up-regulation for the bidirectional pattern ( Fig 4A–4C , cluster 4 ) . Mechanistically , the former could be achieved by sustained inactivation of protein kinases ( cluster 2 ) , such as the inhibition of GSK3β [44] . For the latter , initial stronger down-regulation could have dampened up-regulation ( cluster 4 ) ; however , an alternative mechanism could be the increased activation of protein phosphatases . Phosphatases could have feasibly kept up-regulation low at 900 s after 76 mM KCl while MAPK1 activation was high ( Fig 8B ) . Protein phosphatase 1 ( PP1 ) acts downstream of calcineurin [57 , 58] . PP1 and PP2A account for the vast majority of basal phosphatase activity in presynaptic terminals [57] and are implicated in the regulation of neurotransmitter release [59–61] . Poststimulus up-regulated phosphorylation ( cluster 4 , Fig 8A ) was associated with the increased activity of MAPK1 ( Fig 6B and S6A Fig ) . KinSwing predicted that PKAα , CaMKL , STE20 , and RSK classes of protein kinases also contributed to poststimulus up-regulation . These predictions were supported by evidence from the literature [49–51] . Substrates of poststimulus up-regulation included synapsin 1 , Munc18-1 , and distinct components of active zone scaffold proteins . Endocytosis and endosomal proteins were also implicated ( Fig 8A ) . MAPK1/3 activation and targeting of synapsin 1 was found by others to have a negative effect on posttetanic enhancement [18] . In this context , the poststimulus up-regulation may be a homeostatic plasticity mechanism that dampens the effects of strong stimuli . The up-regulated phosphorylation we identified involves a possible MAPK-signalling pathway that has many more components beyond synapsin 1 ( Fig 8B ) . Activity-dependent phosphorylation in synaptosomes and hippocampal neurons was highly correlated for presynaptic proteins at the common 10-s time point ( Fig 2A ) . Active zone scaffold proteins and synapsins were more highly targeted than postsynaptic scaffold proteins and neurotransmitter receptors ( Fig 2C ) . This indicates that the active zone scaffold proteins may be more finely tuned to respond to Ca2+ influx , despite vastly greater postsynaptic abundance of Ca2+-sensitive proteins such as CaMKIIα . We have identified bassoon as the major presynaptic signalling hub and the most targeted protein for activity-dependent phospho-signalling within hippocampal neurons . Piccolo , RIM1 , microtubule-regulating MAP1B , and tau , as well as SNAP25-interacting protein , were also identified as candidate signal integrators . Thus , phospho-signalling in synaptosomes and intact neurons was similar , and several presynaptic proteins were identified as potential signal integrators , of which bassoon was most prominent . Our 10-s stimulation with KCl would have caused a large calcium influx , which likely exceeds that produced by electrophysiological stimulation . Thus , our results likely include some nonphysiological phospho-signalling , necessitating validation of our work with electrophysiological stimulation paradigms . However , the strength and duration of stimulation may be more important from a physiology prospective than the method of depolarization . A 10-s high-frequency ( ≥40 Hz ) electrical stimulation , combined with phospho-specific antibody detection , produced virtually identical phospho-signalling responses for the proteins dynamin 1 , Akt/PKB , and GSK3β [62] . High-frequency firing rates within the brain are not typical but occur naturally in specific neurons during tasks such as spatial navigation [63] . Obtaining meaningful biochemical measurements at low frequencies will be challenging . In this work , we could have missed regulated phosphorylation sites because of undersampling but also because the change in phosphorylation was too low to measure ( despite functional significance ) . In general , our lower potassium concentration produced robust phospho-signalling , particularly during the poststimulus ( S2B and S2C Fig ) , indicating that high potassium concentrations can potentially be avoided . Nevertheless , elevated potassium stimulation is highly physiologically relevant to understanding the molecular consequences of traumatic brain injury [64] and in the understanding of homeostatic plasticity mechanisms , since pretreatment with elevated potassium was shown to be neuroprotective in a model of excitotoxicity [65] . Our 10-s stimulation was relatively acute , compared to the previous sustained depolarization [27] , and allowed for the first examination of the poststimulus response . Stimulating synaptosomes at a high and low level increased our ability to discern Ca2+ concentration–dependent patterns of temporal regulation . Future work will require a combination of synaptic vesicle turnover and electrophysiological measurements in neurons to obtain a clearer view of the functional consequences of our stimulations and phosphoproteome dynamics . The value of information on activity-dependent phosphorylation is in knowing which protein domains and functions are mechanistically linked to neuronal activity , while ruling out phosphorylation that is independent of activity . Our studies revealed that the vast majority of phospho-regulated functions remain unidentified . The lack of association with activity indicates that neuroscience-relevant signalling mechanisms have not yet been a significant focus for research , which might be explained by known biases in research effort [66] . Here , we have highlighted the activity-dependent phosphorylation site data relevant to presynaptic mechanisms and neurotransmitter release . The sensitivity of presynaptic phospho-signalling , if translated into major functional changes , could have implications for the pharmacological targeting of protein kinase/phosphatase activity within neurons . Other important targets of phospho-signalling that regulate diverse cellular functions are contained within this resource for neuroscientists . Therefore , this resource , and our highly developed analysis approach , will allow others to further decipher the functional significance of synaptosomal and neuronal activity-dependent phosphorylation .
All experiments involving rats were conducted according to authorized procedures and with ethics approval from the Children’s Medical Research Institute/Children’s Hospital at Westmead Animal Ethics Committee ( projects C116 and C353 ) . Eight- to 20-wk-old Sprague-Dawley male rats were humanely killed by decapitation . The whole brain was extracted . P2 fraction synaptosomes were isolated and prepared as described previously [67] , with minor modifications . Briefly , the S1 fraction was centrifuged at 948g for 10 min . All centrifugation steps were performed at 4 °C . The supernatant was discarded , and the P2 fraction was resuspended in a Krebs-like buffer solution ( 118 mM NaCl , 4 . 7 mM KCl , 20 mM 4-[2-hydroxyethyl]-1-piperazineethanesulfonic acid [HEPES]/trisaminomethane [Tris] [pH 7 . 4] , 25 mM NaHCO3 , 1 . 18 mM MgSO4 , 1 . 2 mM CaCl2 , 0 . 1 mM Na2HPO4 , 1 . 85 g/L glucose ) , which had previously been bubbled with carbogen ( 95% O2 , 5% CO2 ) for 45 min to produce a final pH of 7 . 3–7 . 5 . The solution was centrifuged at 13 , 800g for 10 min , and the resulting pellet was resuspended in 10 mL Krebs-like solution and centrifuged at 948g for 10 min to obtain a P2 synaptosome pellet . The pellet was resuspended in 1 . 75 mL of Krebs-like solution , rested at 37 °C for 45 min , and then placed on ice until stimulated . The number of biological replicates is reported for each experiment . For the mass spectrometry and the western blotting , each biological replicate was produced using the brain tissue of an individual rat . Prior to stimulation , the synaptosomes were warmed to 37 °C for 5 min . To depolarize the synaptosomes , aliquots of 300 μL of synaptosomes were mixed with an equal volume of solution such that the final KCl concentration was 20 or 76 mM , and the sodium concentration was lowered by a similar amount such that the monovalent salt and osmotic concentration was constant . The depolarization continued for 10 s only before centrifugation at 13 , 500 rpm in a CM-50 MP ( ELMI , Riga , Latvia ) benchtop centrifuge . To repolarize the synaptosomes , they were resuspended in Krebs solution and incubated for up to 15 min . Samples were collected after 10 s of mock treatment or depolarization and at 90 s , 300 s , and 900 s in the repolarization solution . Prior to lysis , the synaptosomes were pelleted by centrifugation and the supernatant removed . The synaptosomes were lysed in a 300 μL solution and briefly agitated with a benchtop vortex . The lysis solution consisted of 2% SDS , 25 mM HEPES/Tris ( pH 7 . 4 ) , 1 mM ethylenediaminetetraacetic acid ( EDTA ) , 1 mM ethylene glycol-bis ( β-aminoethyl ether ) -N , N , N′ , N′-tetraacetic acid ( EGTA ) , 1x Roche Complete protease inhibitor cocktail , and 1x Calbiochem Phosphatase inhibitor cocktail II . The samples were incubated at 85 °C for 10 min to ensure inactivation of proteases and phosphatases . Synaptosomes were prepared as above . Synaptosomes were stimulated with 76 mM KCl or mock stimulated with 4 . 7 mM KCl for 10 s , centrifuged to remove the liquid , and then lysed , as above . In a third condition , synaptosomes were centrifuged , and then KCl was added to achieve an isotonic solution of 76 mM KCl , as in the stimulated condition , but without disturbing the pellet . After 45 s , the supernatant was removed , and the sample was lysed . Equal amounts of each sample were applied to SDS-PAGE and examined by western blot with anti-synapsin 1 pS603 ( see “SDS-PAGE and western blotting” section below ) . Samples were subjected to reduction with 10 mM dithiothreitol ( DTT ) for 45 min at 56 °C followed by alkylation with 20 mM iodoacetamide for 30 min in the dark at 23 °C . Another 10 mM DTT was added to quench excess iodoacetamide . The protein content of each sample was then precipitated using methanol/chloroform precipitation [68] . The dry protein pellets were redissolved in 90 μL 8 M urea and 10 μL 1 M triethylammonium bicarbonate ( TEAB ) and digested for 2 h at 23 °C using 0 . 15 U endoproteinase Lys-C per sample . The samples were then diluted to 1 M urea with 50 mM TEAB followed by digestion with 20 μg TrypZean trypsin ( Sigma-Aldrich , St . Louis , MI , United States ) per sample . After 4 h , another 20 μg was added and the samples digested for another 4 h ( both steps performed at 23 °C ) . Digested samples were then subjected to in-solution reductive dimethylation [69] essentially according to Boersema and colleagues [70] . Labelling was performed according to the below scheme ( note: control is mock stimulated for 10 s ) : Replicates 1 and 2: 20 mM KCl set 1: Control ( light ) , 10 s ( medium ) , 90 s ( heavy ) 20 mM KCl set 2: Control ( light ) , 300 s ( medium ) , 900 s ( heavy ) 76 mM KCl set 1: Control ( light ) , 10 s ( medium ) , 90 s ( heavy ) 76 mM KCl set 2: Control ( light ) , 300 s ( medium ) , 900 s ( heavy ) Replicates 3 and 4: 20 mM KCl set 1: Control ( medium ) , 10 s ( heavy ) , 90 s ( light ) 20 mM KCl set 2: Control ( medium ) , 300 s ( heavy ) , 900 s ( light ) 76 mM KCl set 1: Control ( medium ) , 10 s ( heavy ) , 90 s ( light ) 76 mM KCl set 2: Control ( medium ) , 300 s ( heavy ) , 900 s ( light ) Replicates 5 and 6: 20 mM KCl set 1: Control ( heavy ) , 10 s ( light ) , 90 s ( medium ) 20 mM KCl set 2: Control ( heavy ) , 300 s ( light ) , 900 s ( medium ) 76 mM KCl set 1: Control ( heavy ) , 10 s ( light ) , 90 s ( medium ) 76 mM KCl set 2: Control ( heavy ) , 300 s ( light ) , 900 s ( medium ) After dimethyl labelling , each set of light , medium , and heavy samples were mixed and acidified to 0 . 2% trifluoroacetic acid ( TFA ) . Insoluble material was removed by centrifugation . Samples were then subjected to phosphopeptide enrichment and fractionation by TiSH ( TiO2 , sequential elution from immobilized metal affinity chromatography [SIMAC] and hydrophilic interaction liquid chromatography [HILIC] ) [30] . While multiphosphorylated peptides from the SIMAC procedure [71] were analyzed directly by LC-MS/MS , monophosphorylated peptide fractions were fractionated using HILIC . Samples were dissolved in 2 μL dimethyl sulfoxide , followed by 18 μL H2O , 180 μL acetonitrile ( AcN ) , and 2 μL 10% TFA . Samples were then fractionated by HILIC using a 1 × 250-mm high-performance liquid chromatography ( HPLC ) column packed with 5-μm TSKGel Amide 80 resin ( Tosoh , Tokyo , Japan ) using a Thermo Scientific Ultimate 3000 HPLC system ( Thermo Scientific , Bremen , Germany ) operated at 50 μL/min . One-minute fractions were collected throughout the increasing aqueous solvent gradient going from 100% to 60% B for 35 min ( B solvent: 90% AcN , 0 . 1% TFA , A solvent: 0 . 1% TFA ) . These fractions were pooled based on the UV absorption chromatogram recorded by an Ultimate 3000 variable wavelength detector at 210 nm to produce 10 fractions containing similar amounts of peptide , which were dried by vacuum centrifugation . Each fraction was resuspended in 0 . 15 μL 100% formic acid ( FA ) followed by 2 . 4 μL 0 . 1% TFA and loaded in A solvent 0 . 1% FA directly onto a homemade 400 × 0 . 075-mm reversed phase capillary column packed with ReproSil Pur C18 AQ 1 . 9-μm resin ( Dr Maisch , Ammerbuch-Entringen , Germany ) using an Ultimate 3000 RSLC nano system ( Thermo Scientific , Bremen , Germany ) . Peptides were analyzed over a 125-min ( HILIC fractions ) or 180-min ( multiphosphorylated fraction ) run with a flow rate of 250 nL min− 1 , being eluted off the column using an increasing gradient from buffer A ( 0 . 1% FA ) to B ( 90% AcN , 0 . 1% FA ) . Eluted peptides were sprayed into an LTQ-Orbitrap Elite hybrid high-resolution mass spectrometer ( Thermo Scientific , Bremen , Germany ) via a Proxeon nano-electrospray source ( Thermo Scientific , Bremen , Germany ) operating at 2 . 3 kV . A full mass spectrometry scan of the m/z 300–1 , 800 range was acquired in the orbitrap at a resolution of 60 , 000 full width at half maximum ( FWHM ) and a target value of 1 × 106 ions . For each full scan , the 10 or seven ( first or second injection replicate , respectively ) most intense ions ( >+1 charge state ) were selected for higher-energy collision dissociation ( HCD ) and detected at a resolution of 15 , 000 FWHM . Settings for the HCD event were as follows: threshold for ion selection was 10 , 000 , target value of ions for HCD was 1 × 105 , maximum injection time was 500 ms at the mass spectrometry level and 300 or 500 ms ( first or second injection replicate , respectively ) at the MS/MS spectrometry level , activation time was 0 . 1 ms , isolation window was 1 . 8 Da , first fixed mass was 120 , and normalized collision energy was 35 or 30 ( first or second injection replicate , respectively ) . The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium [72] via the PRIDE [73] partner repository with the dataset identifier PXD010007 . Raw data files from the LTQ-Orbitrap Elite instrument were then processed and peak lists generated using MaxQuant version 1 . 5 . 2 . 8 [74] . Database searching of the resulting peak lists was performed using the Andromeda search engine [75] built into MaxQuant against the rat UniProt Reference Proteome database ( containing all UniProtKB/Swiss-Prot+TrEMBL rat proteins including isoforms ) ( downloaded 11 March 2015 , containing 29 , 390 entries ) . The database search was performed with trypsin/P digestion , allowing a maximum of two missed cleavages . Carbamidomethylation ( C ) was set as a fixed modification , whereas oxidation ( M ) , acetylation ( protein N-terminal ) , pyro-glutamate ( N-terminal E and Q ) , deamidation ( N and Q ) , and phosphorylation ( S , T , and Y ) were variable modifications . A maximum of five modifications per peptide was allowed , and the maximum peptide mass was 6 , 000 Da . The minimum peptide length was seven amino acids , and the default protein contaminants were included in the search . The first search mass spectrometry tolerance was 20 ppm , the main search mass spectrometry tolerance was 4 . 5 ppm , and the MS/MS tolerance was 20 ppm . Peptide-spectrum match ( PSM ) , protein , and the modification site false-discovery rate were set to 1% . The minimum score for modified peptides was 40 . Re-quantify and “match between runs” options were enabled ( matching time window 0 . 7 min , alignment time window 20 min ) , and second peptide and dependent peptide searching was enabled and disabled , respectively . Synaptosomes lysate was generated from mock and 76 mM KCl stimulated synaptosomes at the time points described above . Samples were subjected to reduction with 10 mM tris ( 2-carboxyethyl ) phosphine ( TCEP ) for 5 min at 85 °C followed by alkylation with 30 mM iodoacetamide for 40 min in the dark at 23 °C . Another 10 mM TCEP was added to quench excess iodoacetamide . Protein content of each sample was precipitated using methanol/chloroform [68] . Dry protein pellets were redissolved in 10 μL 8 M urea and digested for 3 h at 25 °C by adding 5 μg endoproteinase Lys-C ( 1 μg/μL in 2 . 5 mM HEPES/NaOH [pH 8 . 0] ) per sample . Promega Sequencing Grade Trypsin ( 8 μg ) in 80 μL 50 mM HEPES/NaOH ( pH 8 . 0 ) was subsequently added and samples incubated for 16 h at 21 °C . Finally , 15 μg TrypZean trypsin was added to each sample , followed by a 6 h incubation . The peptides were subjected to labelling with TMT10plex according to the vendor’s instructions . Note that although TMT10plex was used , all quantitative information was discarded , since only a gene list was extracted from this data . The sample was subjected to phosphopeptide enrichment . The TiO2 flow-through was desalted using a Waters Sep-Pak Plus C18 cartridge according to the manufacturer’s instructions and retained for LC-MS/MS . Note that no phosphopeptide enrichment analysis of these samples was done . This particular TMT10plex labelled sample was used as an experimental by-product to determine the protein content of synaptosomes . An aliquot of the TiO2 flow-through , containing approximately 2 μg of peptide , was dried completely . The peptide was resuspended in 0 . 3 μL 100% FA followed by 2 . 8 μL 0 . 1% TFA and loaded in A solvent 0 . 1% FA directly onto a homemade 350 × 0 . 075-mm reversed phase capillary column with integrated emitter ( prepared in-house using a P-2000 laser-based micropipette puller [Sutter Instrument , Novato , CA , US] packed with ReproSil Pur C18 AQ 1 . 9-μm resin [Dr Maisch , Ammerbuch-Entringen , Germany] using an Ultimate 3000 RSLC nano system [Thermo Scientific , Bremen , Germany] ) . The peptides were eluted off the column and introduced into a Q Exactive Plus mass spectrometer ( Thermo Scientific , Bremen , Germany ) via Nanospray Flex nano-electrospray source ( Thermo Scientific , Bremen , Germany ) operating at 2 . 3 kV . Peptides were analyzed over a 240-min run selecting the 12 most intense ions for HCD MS/MS . The settings for the HCD MS/MS event were as follows: minimum automatic gain control target was 5 , 500 , intensity threshold was 50 , 000 , the maximum injection time was 110 ms , isolation window was 1 . 4 Da , and normalized collision energy was 34 . Peptide match was preferred , and dynamic exclusion was 20 s . Raw LC-MS/MS data files were processed and peak lists generated using MaxQuant version 1 . 5 . 8 . 3 [74] . Database searching of the resulting peak lists was performed using the Andromeda search engine [75] built into MaxQuant against the rat UniProt Reference Proteome database ( containing all UniProtKB/Swiss-Prot+TrEMBL rat proteins including isoforms ) ( downloaded 11 March 2015 , containing 29 , 390 entries ) . The database search was performed with trypsin/P digestion , allowing a maximum of two missed cleavages . Carbamidomethylation ( C ) was set as fixed modification , whereas oxidation ( M ) , acetylation ( protein N-terminal ) , pyro-glutamate ( N-terminal Q ) , and deamidation ( N and Q ) were variable modifications . A maximum of five modifications per peptide were allowed , and the maximum peptide mass allowed was 6 , 000 Da . The minimum peptide length was seven amino acids , and the default protein contaminants were included in the search . The first search mass spectrometry tolerance was 20 ppm , the main search mass spectrometry tolerance was 4 . 5 ppm , and the MS/MS tolerance was 20 ppm . PSM and protein , as well as the site false-discovery rate , was set to 1% , and the minimum score for modified peptides was 40 . The re-quantify , match between runs , second peptides , and dependent peptide searching options were disabled . The list of identified proteins from the MaxQuant “proteinGroups . txt” output file was used as the background protein list for the “Enrichment analysis” section below . To map and quantify individual phosphosites , we utilized the MaxQuant “evidence . txt” ( ProteomeXchange dataset identifier PXD010007 ) output file containing all identified PSMs for all further analyses . To remove redundancy due to sequences being present as both Swiss-Prot and TrEMBL entries , instances of PSMs to duplicated protein sequences in UniProtKB were removed using a custom Perl script . Only Swiss-Prot entries were kept where exact sequence duplication was observed ( two instances: P03926;Q7HKW2 and P05508;Q7HKW3 ) . Then , to determine the amino acid position of phosphosites relative to the full-length canonical protein sequence , a “mapping file” was created ( S2 Table ) , using a custom Perl script . A total of 22 , 643 unique phosphopeptides were present in the “evidence . txt” file . A total of 11 , 055 phosphopeptides mapped to a single unique protein accession . Of those phosphopeptides , 10 , 530 contained more than one protein accession , and 478 mapped to more than one unique gene identifier . Using a custom Perl script , 1 , 058 were excluded from downstream analyses because they did not contain a localized phosphosite or were contaminant peptides ( “CON__” ) or those identified in the reverse database by MaxQuant processing for a false-discovery rate of identification ( “REV_” ) . To map the phosphosite positions , we developed a “best” evidence algorithm according to the following rules: For phosphopeptides that mapped to multiple UniprotKB protein accessions , the accession with the highest “protein existence ( PE ) ” value was kept as the best evidence ( http://www . uniprot . org/help/protein_existence ) . Where the protein accession was an isoform ( therefore lacking PE information ) , the PE value was taken from the parent protein . When the PE value was equal , a Swiss-Prot ( sp ) entry was taken over a TrEMBL ( tr ) entry . If both entries were Swiss-Prot , the nonisoform was selected . If both entries were isoforms , the lowest-numbered isoform was selected . If the phosphopeptide mapped to multiple unique genes , we determined the “best” annotation for each unique gene ( steps 1–5 ) and recorded this as a multigene-mapped phosphopeptide . For each of the annotations and phosphosites , the centred sequence of amino acids mapping to each phosphosite was retrieved from the UniProtKB fasta file ( i . e . , phosphopeptides containing more than one phosphosite would also have more than one centred sequence ) . Each phosphopeptide identified was then given an identifier , which was composed of the “BEST_PROTEIN” and “BEST_PHOSPHO_POSITION” ( e . g . , D4A3S7:15 , see S2 Table ) and was the basis for downstream statistical analysis of phosphorylation . To determine the significance of phosphorylation changes over time , the mapping file ( S2 Table ) and “evidence . txt” ( ProteomeXchange dataset identifier PXD010007 ) were imported into R version 3 . 4 . 0 ( R Foundation for Statistical Computing , Vienna , Austria ) for further analysis . Each entry ( phosphopeptide ) in the “evidence . txt” file was merged to its corresponding identifier from the mapping file . The MaxQuant normalized intensity for each unique identifier for the 20 and 76 mM KCl 10-s , 90-s , 300-s , and 900-s time point samples was divided by the intensity of the mock-stimulated ( 4 . 7 mM KCl 10 s ) sample to provide a ratio , which was log2 transformed . Where merged phosphopeptides did not contain a minimum phosphosite localization probability score of 0 . 75 in any of the experiments conducted or were not detected in at least three of the six replicates in either the 20 mM or 76 mM KCl conditions , these phosphopeptides were excluded from further analysis . These criteria retained 6 , 489 phosphopeptides for the 20 mM KCl treatment and 6 , 215 unique phosphopeptides for the 76 mM KCl treatment . For the combined 20 mM and 76 mM KCl analysis , this was the intersect of phosphopeptides passing these criteria in both 20 mM and 76 mM KCl treatments , a total of 5 , 715 . For 20 mM and 76 mM KCl data , these data were normalized independently for separate analyses as well as together ( the intersection ) for comparative analyses . Initial inspection indicated requirement for normalization and correction of nonbiological sources of variation . For downstream normalization and analysis , we followed the statistical preprocessing method we previously implemented for analysis of phosphoproteomics data [31] , and more detail is provided in Waardenberg , 2017 [32] . Data were first rescaled using quantile normalization , assuming that data were missing at random [76] , followed by missing value imputation using the k-nearest neighbour approach ( k = 10 ) [77] , and data were iteratively reweighted using the surrogate variable analysis [78] . Consistent with previous findings [31] , this improved separation of biological samples , where the biological variable of interest ( time ) was clearly clustering together and separating across the first principle components ( Fig 1B ) . For statistical comparison of 76 mM or 20 mM KCl data , we fit a linear model with Bayes variance shrinkage using limma [79] and tested for differential phosphorylation at each time point by computing moderated t-statistics ( 10 s versus control; 90 s versus control; 300 s versus control; 900 s versus control ) and the F-statistic for change at any time point and performed pairwise comparisons for each time point . P values were corrected for multiple hypothesis testing using the Benjamini and Hochberg method [80] . For 76 mM versus 20 mM KCl , we tested for differential phosphorylation between these concentrations at each time point using the intersected data ( as per above ) . Unless reported otherwise , a corrected P value less than or equal to 0 . 05 was reported as significant . All data and statistics for each comparative analysis are available in S1 Table . One-day-old Wistar rat pups were humanely killed by decapitation . The brain was removed and dissected on ice to isolate the hippocampus . The hippocampi from eight pups were incubated in 4 mL dissection buffer ( Hank’s Balanced Salt Solution supplemented with penicillin/streptomycin , 1 mM sodium pyruvate , 20 mM HEPES/Tris [pH 7 . 4] ) and 0 . 45% glucose ( Thermo Scientific , North Ryde , NSW , Australia ) with 120 μL 43 . 6 mg/mL ( ≥104 units ) papain solution ( Worthington Biochemical Corporation , Lakewood , NJ , US ) at 37 °C for 20 min . The buffer was aspirated , and the hippocampi were washed three times in 4 mL Neurobasal medium supplemented with 5% foetal bovine serum , 4 mM GlutaMax , penicillin/streptomycin , and B27 ( Thermo Scientific , North Ryde , NSW , Australia ) ( “NM5 medium” ) followed by trituration using glass pipettes with an increasingly smaller tip diameter . The dissociated neurons were suspended in 50 mL NM5 medium and seeded into poly-D-lysine-coated six-well cell culture plates ( 4-mL suspension per well ) . The following day , the medium was replaced with Neurobasal medium supplemented with 4 mM GlutaMax , penicillin/streptomycin , and B27 ( “NM0 medium” ) . Until 21 d in vitro ( DIV21 ) , half of the medium was replaced with fresh NM0 twice per week . At DIV21 , the neurons were mock stimulated ( 4 . 7 mM KCl ) or depolarized ( 76 mM KCl ) for 10 s . At the 10-s time point , immediately after aspirating the stimulation buffer , the neurons were lysed by adding 500 μL 20 mM HEPES/NaOH ( pH 8 . 0 ) , 2% SDS , 2 mM EGTA , 2 mM EDTA with 1x Roche Complete protease inhibitor cocktail and 1x Calbiochem Phosphatase inhibitor cocktail II . The lysate was transferred to a 2-mL microcentrifuge tube and frozen on dry ice . At the end of the experiment , the lysates were heated to 90 °C for 5 min to inactivate all proteases and phosphatases . All samples were subjected to reduction with 10 mM TCEP for 5 min at 85 °C followed by alkylation with 30 mM iodoacetamide for 40 min in the dark at 23 °C . The protein content of each sample was precipitated using MeOH/chloroform [68] . Dry protein pellets were redissolved in 20 μL 9 M urea and diluted to 1 M urea with 35 mM HEPES/NaOH ( pH 8 . 0 ) . TrypZean trypsin ( 8 μg; Sigma-Aldrich , Castle Hill , NSW , Australia ) was subsequently added to each sample , followed by incubation for 16 h at 21 °C . The samples were dried by vacuum centrifugation and resuspended in 30 μL 50 mM HEPES/NaOH ( pH 8 . 0 ) . The peptide samples were subjected to quantitative labelling with TMT10plex according to the vendor’s instructions and the following scheme . Three tags were used for an experiment that was discarded . The discarded experiment was a 20 mM KCl treatment . The experiment was discarded on the basis that well-known activity- and Ca2+ influx–dependent protein kinase substrate sites S603 and S566 in synapsin 1 were not up-regulated , and Ca2+ influx activity–dependent dynamin 1 phosphatase sites S774 and S778 were not down-regulated ( all raw data are available via ProteomeXchange with identifier PXD010007 ) . TMT-126: 4 . 7 mM KCl , replicate 1 TMT-127N: Discarded condition TMT-128C: Not used TMT-128N: 76 mM KCl , replicate 1 TMT-128C: 4 . 7 mM KCl , replicate 2 TMT-129N: Discarded condition TMT-129C: 76 mM KCl , replicate 2 TMT-130N: 4 . 7 mM KCl , replicate 3 TMT-130C: Discarded condition TMT-131: 76 mM KCl , replicate 3 After 60 min of incubation at 21 °C , 5% hydroxylamine was added to an end concentration of 0 . 25% , followed by incubation for 20 min at 21 °C . The samples were then mixed and subjected to phosphopeptide enrichment and fractionation using TiO2 and HILIC as described for the reductive dimethylation experiment but without performing the SIMAC procedure . Based on the UV absorption chromatogram , the HILIC fractions were pooled into 14 fractions containing similar amounts of peptide , which were dried by vacuum centrifugation prior to LC-MS analysis . Each fraction was resuspended in 0 . 3 μL 100% FA followed by 2 . 8 μL 0 . 1% TFA and loaded in A solvent 0 . 1% FA directly onto a homemade 450 × 0 . 075-mm reversed phase capillary column with integrated emitter ( prepared in-house using a P-2000 laser-based micropipette puller [Sutter Instrument , Novato , CA , US] ) packed with ReproSil Pur C18 AQ 1 . 9-μm resin ( Dr Maisch , Ammerbuch-Entringen , Germany ) using an Ultimate 3000 RSLC nano HPLC system ( Thermo Scientific , Bremen , Germany ) . Peptides were analyzed using a 180-min LC-MS run with a flow rate of 250 nL min− 1 being eluted off the column using an increasing gradient from buffer A ( 0 . 1% FA ) to B ( 90% AcN , 0 . 1% FA ) . Eluted peptides were introduced into an Q Exactive Plus mass spectrometer ( Thermo Scientific , Bremen , Germany ) via a Nanospray Flex ion source ( Thermo Scientific , Bremen , Germany ) operating at 2 . 3 kV . A full mass spectrometry scan of the m/z 375–1 , 500 range was acquired at a resolution of 70 , 000 FWHM , a target value of 3 × 106 ions , and a maximum injection time of 100 ms ( mass spectrometry ) . For each full scan , the 12 most intense ions ( charge state +2 to +7 ) were selected for HCD and detected at a resolution of 35 , 000 FWHM . The settings for the HCD MS/MS event were as follows: Minimum automatic gain control target was set to 2E5 , intensity threshold was 1 . 0E5 , the maximum injection time was 120 ms , isolation window was 1 . 0 Da , first fixed mass was 120 , and normalized collision energy was 34 ( stepped collision energy ) . Peptide match was off , “Exclude isotopes” was on , “dynamic exclusion” was 12 s , and the “If idle” setting was set to “pick others” . Raw LC-MS/MS data files were processed , and peak lists were generated using MaxQuant version 1 . 5 . 2 . 8 [74] . Database searching of the resulting peak lists was performed using the Andromeda search engine [75] built into MaxQuant against the rat UniProt Reference Proteome database ( containing all UniProtKB/Swiss-Prot+TrEMBL rat proteins including isoforms ) ( downloaded 11 March 2015 , containing 29 , 390 entries ) . The database search was performed with trypsin digestion , allowing a maximum of two missed cleavages . Carbamidomethylation ( C ) was set as fixed modification , whereas oxidation ( M ) , acetylation ( protein N-terminal ) , pyro-glutamate ( N-terminal Q ) , deamidation ( N and Q ) , and phosphorylation ( S , T , and Y ) were variable modifications . A maximum of five modifications per peptide was allowed , and the maximum peptide mass allowed was 6 , 000 Da . The minimum peptide length was seven amino acids , and the default protein contaminants were included in the search . The first search mass spectrometry tolerance was 20 ppm , the main search mass spectrometry tolerance was 4 . 5 ppm , and the MS/MS tolerance was 20 ppm . PSM and protein , as well as the site false-discovery rate , was set to 1% , and the minimum score for modified peptides was 40 . The “re-quantify” and “match between runs” options were disabled . For TMT-derivatized peptides , 47 , 674 unique phosphopeptides were present in the “evidence . txt” file . Following the same filtering procedure for the dimethylation data , 22 , 063 unique phosphopeptides remained for statistical analysis . Data were processed essentially as described for the dimethylation dataset . A log2 ratio was computed for each of the three paired replicates ( 76 mM KCl treated intensity/4 . 7 mM KCl control intensity ) using the corrected reporter ion intensities provided in the MaxQuant file . Missing value imputation was not conducted . Thus , phosphopeptides with no , or zero , intensity values were discarded . Log2 ratios were then corrected via surrogate variable analysis , which included the discarded conditions ( see above ) to ensure that all sources of variation were modelled . Limma analysis was used to calculate moderated t-statistics . P values were corrected for multiple hypothesis testing using the Benjamini and Hochberg method [80] . All data and statistics for each comparison ( dimethylated and TMT ) are available in S1 Table . The 1 , 917 unique phosphopeptides from the 20 mM and 76 mM KCl stimulated dimethylation dataset were imported into Perseus 1 . 5 . 8 . 5 for further analysis . Data were k-means clustered by row ( each phosphopeptide ) with 100 starts and 100 iterations . Six clusters ( k ) was determined to be optimal for enrichment analysis ( S3A and S3B Fig ) . Thus , very small and very large clusters were avoided for enrichment analysis . Clusters 1 and 2 , which consisted of the highest magnitude changes , contained 57 and 44 unique UniProt accessions , respectively . Clusters 3 , 4 , 5 , and 6 contained 130 , 129 , 451 , and 368 unique accessions . The entire set of 1 , 917 unique phosphopeptides contained 865 unique accessions . Each of these sets were imported into DAVID ( database for annotation , visualization , and integrated discovery ) [81] for gene ontology enrichment analysis . The protein content of our P2 synaptosomes was determined by LC-MS/MS analysis in a separate experiment ( see section “Preparation of samples , mass spectrometry , and data processing for determination of the P2 synaptosomes proteome used in the enrichment analysis” above ) . The identified proteins contained 3 , 866 unique accessions . These accessions were combined with the 865 unique proteins accessions from phosphopeptides identified from synaptosomes and 4 , 947 unique protein accessions from phosphopeptides identified from hippocampal neurons . The intersection of these sets was a list of 7 , 588 nonredundant protein accessions , which was used as the background . Gene ontology enrichment was queried for “molecular function” , “biological process” , and “cellular component” using “MF_Direct” , “BP_Direct” , and CC_Direct” , since these sets provide direct annotations for each accession [82] . The default threshold P > 0 . 1 was used to extract an initial list of enriched terms , which was also required to have at least four accessions , prior to further filtering . For comparison of clusters , the top five most probable enriched terms/domains from each cluster with P < 0 . 05 were arranged into a heat map of probabilities . Columns ( clusters ) in the heat map were allowed to exceed the initial limit of five terms if they were highly ranked in another cluster , but each row was required to contain a term with at least P < 0 . 01 . The P values for each term were hierarchically clustered by row ( each term ) using Perseus 1 . 5 . 8 . 5 . Phosphorylation site regulation ontology term enrichment was performed as follows . Terms were extracted from the “ON_FUNCTION” and “ON_PROCESS” columns of the “Regulatory_sites . gz” file , downloaded 26 May 2018 from PhosphoSitePlus [33] . The terms are available in S1 Table . Only terms with at least nine matching phosphorylation sites within the set of 1 , 917 significantly regulated phosphorylation sites from synaptosomes were considered . There were 83 , 82 , 201 , 191 , 797 , and 563 unique phosphopeptides for clusters 1 , 2 , 3 , 4 , 5 , and 6 , respectively . This included possible redundant multisite phosphorylation assignments , which were kept . The number of occurrences for each term per cluster was determined . Fold enrichment was calculated by considering the ratio between foreground and background terms as a 2 × 2 contingency table of ( 1 ) the number of terms matched within a cluster , ( 2 ) the size of the cluster , ( 3 ) the number of term matches for all 1 , 917 peptides , and ( 4 ) 1 , 917 as the total background . For each term , significance of the fold enrichment was determined by applying a Fisher’s exact test to the 2 × 2 contingency table ( P < 0 . 05 ) . All clusters with no term matches were assumed nonenriched and given a P value of 1 . The centred peptide sequences for each of the six clusters were examined for enriched protein kinase substrate motifs using motif-x [40] with the following settings: central character , S; width , 15; occurrences , 20; significance , 0 . 000001; and background , IPI Rat Proteome . Curated substrate sequences for all human protein kinases , which consisted of in vitro and in vivo evidence , were downloaded from PhosphoSitePlus ( 13 January 2016 ) [33] . Protein kinase family and class information was downloaded from kinase . com ( 26 July 2017 ) [83] . A human substrate dataset was used , since it was much larger than the rat dataset , including increased coverage of the kinome . Tyrosine protein kinases and their substrates were not considered and were removed from the experimental and curated datasets . All analyses were performed in R version 3 . 4 . 0 ( R Foundation for Statistical Computing , Vienna , Austria ) . For building protein kinase substrate models , we considered the set of n centred amino acid sequences ( 13 amino acids in length , from PhosphoSitePlus ) , S , for each protein kinase , as a set of identity matrices , Ka , p , for each amino acid , a , where a ∈ {1 , 2 … 20} , at each sequence position , p , where 1 indicates a match for each amino acid , a , at position p for the i-th amino acid sequence , and 0 indicates no match for amino acid residue , a , at position p for the i-th amino acid sequence: Ka , p={ 1 , ifSi , p=a0 , ifSi , p≠a and determined the position probability matrix , Ma , p , for each protein kinase , K , as a log likelihood ratio of the average frequency of each amino acid , a , at each position , p , for the set of sequences , S , divided by the background frequencies of amino acids a , Ba: Ma , p=log ( ( 1n∑i=1nKi ) +CBa+C ) A pseudocount , C , of 0 . 00001 was added to the position probability matrix to avoid log zero divisions and Ba = 0 . 05 assuming equal distribution of amino acids . For each input sequence identified in our phosphoproteome ( 13 amino acids in length , centred on each phosphosite ) , we then score its match to protein kinase , K , as the sum of the corresponding entry , Map , of matrix , Ma , p , for each amino acid residue , a , at position , p , of sequence length i: Sscore=∑i=1nf ( a , p ) where f ( a , p ) =Map∈Ma , p We then determined the probability , P ( Sscore|R , N ) , where the probability of Sscore is conditional on a reference distribution of N randomly sampled sequences , of 1 , 000 , and R sequences determined to have a test statistic less than or equal to Sscore: P ( score ) =R+1N+1 where R=∑n=1NI ( ( Sscore ) n*≥ ( Sscore ) i ) To build protein kinase substrate networks , we consider the set of protein kinase substrate predictions P ( Sscore|R , N ) ≤ 0 . 05 as edges . Sipk then represents the product of two logic functions: ( 1 ) Sik , directionality , based on the log2 fold change , i , of each predicted substrate of protein kinase , k; and ( 2 ) Spk , significance , p , of phosphorylation change for each predicted substrate of protein kinase , k: Sipk=Sik∙Spk , Sipk≠0 where i is given and is the log fold change: Sik=f ( i ) ={ −1 , i<01 , i≥0 and p is given and is the significance of log fold change , i: Spk=f ( p ) ={ 1 , p<0 . 050 , p≥0 . 05 Sipk therefore represents the state of the kinase substrate network , where −1 represents protein kinase substrates significantly differentially phosphorylated and negatively regulated , and +1 represents protein kinase substrates significantly differentially phosphorylated and positively regulated . For each protein kinase , K , the sum of the absolute values for each protein kinase , Ck , therefore represents the overall connectivity of the protein kinase , K ( i . e . , how many substrates were associated with a protein kinase for each condition and time point ) : Ck=∑Sipk≠0[Sipk] Posk represents the positive proportion of edges for protein kinase , K , Posk=∑Sipk≥0[Sipk]Ck Negk represents the negative proportion of edges for protein kinase , K , Negk=∑Sipk<0[Sipk]Ck swingk is an overall score of the proportion of positive and negative edges for protein kinase , K , weighted for the number of edges , Ck , and the number of substrate sequences , Sn . A pseudocount , c , of 1 , is added to Posk and Negk to avoid log zero divisions , thereby weighting for protein kinases with the greatest swing and greatest evidence of swing based on the number of predicted substrates and the number of substrates used to infer the protein kinase position weight matrix: swingk=log2 ( Posk+cNegk+c ) *log2 ( Ck ) *log2 ( Sn ) To permit comparison of kinase activity scores across time , swingk was finally transformed into a z-score , Z ( swingk ) , where μ is the mean of all swing scores , and σ is the standard deviation of all swing scores: Z ( swingk ) =swingk−μσ The raw output was filtered such that each reported z-score , Z ( swingk ) , which we name the KS score , required ( 1 ) at least 19 experimentally validated protein kinase substrate matches and ( 2 ) ≥1 time point with a KS score greater than the corresponding standard deviation for that score ( such that the KS score deviated from zero ) . See S1 Table for KS scores and standard deviations for all protein kinases examined across time . KS scores were also calculated for groups of protein kinases that were annotated according to classes , and these results are also in S1 Table . The remaining scores were clustered using the Perseus 1 . 5 . 8 . 5 hierarchical clustering function . During the course of this work , methods with similarities to KinSwing were published , i . e . , KARP ( kinase activity ranking using phosphoproteomics data ) [84] , IKAP ( inference of kinase activities from phosphoproteomics ) [85] , and iKAP ( in silico kinome activity profiling ) [86] . KARP has a fundamentally different output , resulting in a ranking of protein kinases on a strictly positive scale , not inferring overall directionality of activity . Although iKAP is concerned with quantification of directionality of kinase substrates , they utilize a chi-squared test for testing the proportion of positive and negative phosphosite edges within a sample condition , rather than transforming to a statistic , which is valuable for comparison across multiple conditions and time courses , as we have done . IKAP uses an optimization procedure based on precise kinase substrate database matching followed by correlation , which assumes that kinase substrate networks act in a linear manner . Future studies that determine precision based on experimental prediction would further guide the development of these algorithms . Network edges were determined by submission of a nonredundant list of gene names for the 1 , 917 significantly regulated phosphopeptides to STRING [87] using Rattus norvegicus as the organism . Node and edge data from STRING were subsequently imported into Cytoscape version 3 . 5 . 1 [88] . The following R . norvegicus gene names used by STRING were manually changed to match the gene names used by DAVID and/or UniProt: ENSRNOG00000018712 , Camk2a; Dmxl2-ps1 , Dmxl2; LOC687090 , Tprg1l; Sec311 , Exoc1 . Within Cytoscape , the network was filtered by discarding proteins without specific cellular component annotations . Nodes were required to have the following DAVID “CC_Direct” cellular component annotations , which were grouped into nine categories: ( i ) cytoskeleton and actin cytoskeleton , ( ii ) endosome , ( iii ) microtubule and microtubule organizing centre , ( iv ) synaptic vesicle , ( v ) active zone scaffold , ( vi ) voltage-gated potassium channel , ( vii ) voltage-gated calcium channel , ( viii ) soluble N-ethylmaleimide-sensitive factor attachment protein receptor ( SNARE ) complex , ( ix ) clathrin-coated pit , and ( x ) presynaptic membrane and membrane raft . Protein kinases , protein phosphatases , and subunits/regulators of phosphatases were also retained ( identified by UniProt gene ontology annotation as of 31 July 2017 ) . The cellular component data were curated to improve relevance to synaptic biology . The following manual changes were made to cellular component annotation or protein kinase/phosphatase annotation based on UniProt annotations as of 31 July 2017 or references provided: Syn1 changed from “active zone scaffold” to “synaptic vesicle” , Syn2 changed from “SNARE complex” to “synaptic vesicle” , Camk2a changed from “presynaptic membrane” to “synaptic vesicle” , Caskin1 annotated with “active zone scaffold” [5] , Stxbp1 annotated with “SNARE complex” , Nefh annotated with "cytoskeleton" , Sh3gl2 annotated with "synaptic vesicle" , Syt9 annotated with "synaptic vesicle" , Rab12 annotated with “endosome” , Rab27a annotated with "synaptic vesicle" , Rab8b annotated with "synaptic vesicle" , Sgsm1 annotated with "synaptic vesicle" , Ppfia1 annotated with "active zone" , Exoc1 annotated with "synaptic vesicle" , Exoc7 changed from “microtubule” to “synaptic vesicle” , Dbnl annotated with "synaptic vesicle" , Slc30a3 annotated with "synaptic vesicle" , C2cd2l annotated with "presynaptic membrane" , Dnajc6 annotated with "clathrin-coated pit" , Doc2a annotated with "synaptic vesicle" , Dnm3 changed from “microtubule” to “synaptic vesicle” , Cadps annotated with “synaptic vesicle" , Synj1 changed from “microtubule” to synaptic vesicle , Ap3b2 and Ap3d1 annotated with “clathrin-coated pit” , Dlg4 “synaptic vesicle” annotation was deleted , Sptbn1 annotated with “actin cytoskeleton” , Sptbn2 annotated with “actin cytoskeleton” , Sptbn4 annotated with “actin cytoskeleton” , Dyrk1b annotated with “protein kinase” , Itsn1 changed from cytoskeleton to “clathrin-coated pit” , and Dpysl3 , Dpysl4 , and Dpysl5 annotated with “microtubule” . Each of the 11 groups were separately arranged in Cytoscape 3 . 5 . 1 using the “grid layout” . Arbitrary adjustments in the layout of each grid were made to improve clarity; thus , no conclusions should be drawn from the distance between nodes . Membrane proteins were arbitrarily arranged in a line . A subset of inositol kinases and phosphatases were moved to the membrane ( Pip5k1c , Inpp4a , and Pi4k2a ) . The network was made up of a total of 267 nodes and 907 edges . The edges were coloured red if the “experimentally_determined_interaction” value from STRING was ≥0 . 3 . Proteins were represented by gene names . Proteins with three or more responding phosphorylation sites had gene names presented in blue letters . Letter size for each node/gene name was scaled linearly ( from an arbitrary minimum value ) by using the quantitative phosphopeptide data . For clusters 1 and 2 , the absolute value of the largest log2 ( stimulated intensity/control intensity ) after either 20 mM or 76 mM KCl stimulation was used to scale the letters . For cluster 4 , the largest up-regulation from 10 s to 300 s or from 10 s to 900 s after 20 mM KCl was used to scale the letter size . For clusters 3 , 5 , and 6 , the average poststimulus log2 ( stimulated intensity/control intensity ) , using all 20 mM and 76 mM KCl values , was used in scaling letters . Networks for specific clusters were filtered to remove proteins that were not significantly phospho-regulated within the cluster . Unfiltered networks for each cluster can be found in S1–S6 Files . Samples were separated via SDS-PAGE and transferred to 0 . 22-μm-pore-size nitrocellulose membrane using the Trans-Blot Turbo Transfer System ( both Bio-Rad , Gladesville , NSW , Australia ) . Blots were blocked with a solution of phosphate-buffered saline ( PBS ) with 5% ( w/v ) bovine serum albumin , 0 . 5% PVP-40 , 0 . 1% ( v/v ) TWEEN-20 ( PBST ) for 1–16 h at 4 °C . Blots were incubated with the following primary antibodies and their respective concentrations diluted into PBST: 1:2 , 000 Anti-synapsin phospho-Ser-603 rabbit polyclonal ( 612-401-C95 , Rockland Immunochemicals , Pottstown , PA , US ) 1:2 , 000 Anti-synapsin phospho-Ser-62+Ser-67 rabbit polyclonal ( TA309241 , Origene , Rockland , MD , US ) 1:1 , 000 Anti-dynamin phospho-Serine-774 sheep polyclonal [17] 1:2 , 000 Anti-CaMKII phospho-Thr-286 rabbit monoclonal ( D21E4 , Cell Signalling Technology , Beverly , MA , US ) 1:50 , 000 Anti-β-actin horseradish peroxidase–conjugated mouse monoclonal ( A3854 , Sigma-Aldrich Castle Hill , NSW , Australia ) Primary antibody incubation was performed for 1 h at 22 °C , with agitation . Blots were then washed 3 times for 5 min each with PBST and incubated with the following secondary antibodies at their respective concentrations diluted into PBST: 1:20 , 000 Anti-Rabbit IgG horseradish peroxidase–conjugated pig polyclonal ( P0217 , Dako , Glostrup , Denmark ) 1:20 , 000 Anti-sheep IgG horseradish peroxidase–conjugated rabbit polyclonal ( P0163 , Dako , Glostrup , Denmark ) Secondary antibody incubation was performed for 1 h at 22 °C with agitation and was not required for the anti-β-actin antibody . Blots were washed 3 times for 5 min per wash with PBST and then incubated for 5 min with SuperSignal West Pico Chemiluminescent Substrate ( Thermo Fisher Scientific , West Ryde , NSW , Australia ) solution composed of 1:1 peroxide:luminol enhancer solution . Blots were transferred to a transparent plastic pouch , gently strained of excess substrate , and then imaged using an ImageQuant LAS-4000 Camera ( Fujifilm , Brookvale , NSW , Australia ) . Image files were processed and densitometry was performed using the associated Fujifilm Multi Gauge software .
|
Neurobiological processes are altered by linking neuronal activity to regulated changes in protein phosphorylation levels that influence protein function . Although some of the major targets of activity-dependent phospho-signalling have been identified , a large number of substrates remain unknown . Here , we have screened systematically for these substrates and extended the list from hundreds to thousands of phosphorylation sites , thereby providing a new depth of understanding . We monitored phospho-signalling for 15 min after the stimulation , which to our knowledge had not been attempted at a large scale . We focused on presynaptic protein substrates of phospho-signalling by isolating the presynaptic terminal . We also stimulated hippocampal neurons but did not monitor the poststimulus . Although the phospho-signalling is immensely complex , the findings could be simplified through data exploration . We identified distinct patterns of presynaptic phospho-regulation across the time course that may constitute co-regulated protein networks . In addition , we found a subset of proteins that had many more phosphorylation sites than the average and high-magnitude responses , implying major signalling or functional roles for these proteins . We also determined the likely protein kinases with the strongest responses to the stimulus at different times using KinSwing , a computational tool that we developed . This resource reveals a new depth of activity-dependent phospho-signalling and identifies major signalling targets , major protein kinases , and co-regulated phosphoprotein networks .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"vesicles",
"nervous",
"system",
"enzymes",
"protein",
"interaction",
"networks",
"enzymology",
"electrophysiology",
"neuroscience",
"phosphatases",
"synaptic",
"vesicles",
"synaptosomes",
"network",
"analysis",
"cellular",
"structures",
"and",
"organelles",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"proteins",
"protein",
"kinases",
"proteomics",
"biochemistry",
"methods",
"&",
"resources",
"cytoskeletal",
"proteins",
"cellular",
"neuroscience",
"anatomy",
"post-translational",
"modification",
"synapses",
"cell",
"biology",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"neurophysiology"
] |
2019
|
The temporal profile of activity-dependent presynaptic phospho-signalling reveals long-lasting patterns of poststimulus regulation
|
In recent years different types of structural variants ( SVs ) have been discovered in the human genome and their functional impact has become increasingly clear . Inversions , however , are poorly characterized and more difficult to study , especially those mediated by inverted repeats or segmental duplications . Here , we describe the results of a simple and fast inverse PCR ( iPCR ) protocol for high-throughput genotyping of a wide variety of inversions using a small amount of DNA . In particular , we analyzed 22 inversions predicted in humans ranging from 5 . 1 kb to 226 kb and mediated by inverted repeat sequences of 1 . 6–24 kb . First , we validated 17 of the 22 inversions in a panel of nine HapMap individuals from different populations , and we genotyped them in 68 additional individuals of European origin , with correct genetic transmission in ∼12 mother-father-child trios . Global inversion minor allele frequency varied between 1% and 49% and inversion genotypes were consistent with Hardy-Weinberg equilibrium . By analyzing the nucleotide variation and the haplotypes in these regions , we found that only four inversions have linked tag-SNPs and that in many cases there are multiple shared SNPs between standard and inverted chromosomes , suggesting an unexpected high degree of inversion recurrence during human evolution . iPCR was also used to check 16 of these inversions in four chimpanzees and two gorillas , and 10 showed both orientations either within or between species , providing additional support for their multiple origin . Finally , we have identified several inversions that include genes in the inverted or breakpoint regions , and at least one disrupts a potential coding gene . Thus , these results represent a significant advance in our understanding of inversion polymorphism in human populations and challenge the common view of a single origin of inversions , with important implications for inversion analysis in SNP-based studies .
Different types of structural variants ( SVs ) , including deletions , duplications , insertions , inversions , and translocations , have been recently discovered in the human genome and there is growing evidence of their importance in human diseases , complex traits , and evolution [1]–[4] . Specifically , inversions , that involve a change of orientation of DNA , have been a model in evolutionary biology for almost 90 years and were shown to have adaptive effects in the genus Drosophila and other organisms [5] , [6] . In humans , inversions have been associated to genetic diseases , such as Hemophilia A [7] , and complex disorders , like asthma [8] . In addition , inversions could confer susceptibility to other rearrangements with negative phenotypic consequences , like the emerin gene deletion [9] , [10] , olfactory receptor-gene cluster translocation [11] , and microdeletion syndromes such as those in 3q29 and 17q21 . 31 [12] , [13] . Finally , human inversions could also have important evolutionary consequences , as in the case of the 17q21 . 31 region that has been related to increased female fertility and positive selection [14] . The majority of SV studies have been based on microarrays , which are very powerful to identify unbalanced changes like copy number variants ( CNVs ) . Inversions , on the other hand , are very difficult to study because they do not usually result in gain or loss of DNA . Therefore , the knowledge of human inversions has lagged behind . This is reflected in the fact that although several hundred inversions have been reported in humans [4] , few ( <15 ) inversions have been characterized in greater detail [4] , [11] , [12] , [14]–[22] . These and other studies [23]–[25] have shown that inversions are generated by two main processes: breaks in relatively simple regions that are joined in opposite orientation by non-homologous mechanisms or non-allelic homologous recombination ( NAHR ) between inverted repeats ( IRs ) ( either repetitive elements or segmental duplications ( SDs ) ) . However , very little is known about the frequency and distribution of the inversions in human populations , with most studies limited to a handful of individuals . The main exceptions are the 17q21 . 31 inversion , in which the global distribution has been estimated by linkage with SNPs [14] , six large inversions studied by FISH in 27 individuals of three populations [12] , the worldwide genotyping of the 8p23 inversion distribution based on SNP data and genetic substructure [21] , and the recent analysis of eight simple inversions in 42 human samples of diverse origins , including one inversion genotyped in 57 populations [20] . Traditional methods for inversion analysis include G-banding karyotyping [26] , FISH [11] , [12] , Southern blot hybridization [9] , or pulsed-field gel electrophoresis [19] . With the development of sequencing techniques , it has been possible to identify inversions at a genome-wide level by sequence comparison [18] , [27] , [28] or paired-end sequencing and mapping ( PEM ) [20] , [23] , [25] , [29]–[32] . However , these techniques have important limitations . In most cases the presence of IRs at the inversion breakpoints hinders their detection using single reads or paired-end sequencing of short fragments , and PEM has a very high rate of false positives in inversion prediction [33] , [34] . Moreover , all the above methods have a low throughput and can be applied only to a few individuals , providing just a partial picture of the distribution of the inversions . As an alternative , computational algorithms have been developed to take advantage of available nucleotide variation data in multiple individuals to predict inversions and estimate inversion genotypes [21] , [35]– , although it is unclear how accurate these predictions are across populations . In addition , these algorithms work only for relatively large and frequent inversions . Thus , it is always necessary to validate inversion predictions using independent techniques and genotype them in a large number of individuals . PCR amplification offers more possibilities for high-throughput analysis and different PCR-based techniques have been used to validate inversions . Regular or long-range PCR [20] , [25] , [27] , [38] , [39] are limited by the size of the fragments to amplify and work poorly for fragments above 10 kb . Therefore , their applicability is reduced to inversions generated by simple breaks or small IRs at their breakpoints . Haplotype-fusion PCR is a very promising technique to study inversions caused by duplicated sequences of almost any kind [40] , [41] , although it has not yet been used extensively and reproducibly to genotype inversions . Inverse PCR ( iPCR ) [42] is based on creating circular molecules of DNA by restriction enzyme digestion and self-ligation , followed by amplification across a self-ligated site ( Figure 1 ) . Thus , with iPCR there is no need to amplify across the breakpoints and it is possible to analyze inversions mediated by medium-long IRs . iPCR has been used to sequence the flanking regions of known sequences [43] , sequence breakpoints of translocations [44] , [45] , or generate long insert pairs [46] . In addition , an iPCR assay has been applied to genotype inversions mediated by 9 . 5 kb SDs causing hemophilia A in multiple patients [47]–[50] and in prenatal diagnosis [51] . However , iPCR limits for different types of inversions have not yet been tested . Therefore , optimization of the iPCR technique could open the possibility of high-throughput validation and genotyping of inversions in a simple manner . Here , we have tested different reagents and conditions to optimize the iPCR method and developed a high-throughput iPCR protocol that allows us to genotype a wide-variety of inversions mediated by IRs in a large number of individuals in just one day . As an example of the potential use of this method , we have analyzed 22 inversions predicted in the human genome by PEM and determined the frequency and genetic transmission of 17 of them in 70 CEU individuals . In addition , we have checked the association between inversions and nucleotide variation using the HapMap [52] and 1000 Genomes Project ( 1000GP ) [53] SNP data , the putative ancestral orientation , and the possible gene effects of the validated inversions . All this information has been deposited in the new InvFEST human polymorphic inversion database [54] .
Inversion validation was based on the analysis of PEM data from fosmids of nine different individuals [23] with our inversion prediction algorithm GRIAL to confirm the presence of an inversion signature and refine the breakpoint locations [54] . For simplicity , we always refer to the orientation matching the NCBI36 ( HG18 ) human reference genome as standard ( Std ) and the opposite orientation as inverted ( Inv ) arrangement . We selected 31 inversion candidates present in at least one of the nine individuals with ≥2 fosmids supporting both the Std and Inv orientations , and IRs of <40 kb in the predicted ranges of the two breakpoints . The next step was finding restriction enzymes that cut at both sides of the inversion breakpoints , but not within the IRs , and that generate fragments of less than ∼70 kb . This way we identified 22 potential inversions for iPCR analysis distributed in 11 chromosomes and with a size of 5 . 1–226 kb ( Table 1 and Table S1 ) . For 13 inversions , the region of sequence exchange could be defined more precisely from at least one completely sequenced BAC or fosmid showing the inverted conformation [23] , [24] . In all cases the refined breakpoints were located either within or at the ends of the IRs in a region of high sequence identity between them [54] . Finally , we compared the selected regions in HG18 with those in the GRCh37 ( HG19 ) assembly or additional genome patches [55] and in three cases we found important changes between them , which were analyzed in more detail ( see below ) . Figure 1 shows the general iPCR design for inversions mediated by IRs between regions A and B ( breakpoint 1 , BP1 ) and C and D ( breakpoint 2 , BP2 ) . Previous studies had shown that iPCR could be used to validate inversions mediated by IRs with self-ligating circular fragments of up to 21 . 5 kb [47] . Therefore , to apply this technique to the largest range of inversions possible , we first tested the optimal DNA concentration and the effects of different reagents in the circularization efficiency of long DNA fragments . The analysis was based on the previous work of Wo et al . [56] , and we selected two inversions , HsInv0340 and HsInv0286 , which generate circular molecules of approximately 30–40 kb in the two orientations when cut with the staggered-end enzyme BamHI or the blunt-end enzyme SwaI . Consistent with previous observations [56] , for the inversion analyzed with a staggered-end enzyme ( HsInv0340 ) , we found a continuous increase in iPCR amplification with DNA dilution during ligation , with the most diluted DNA ( 0 . 31 ng/µl ) showing a 6 . 5–8 . 2-fold improvement in circularization efficiency for both sizes of DNA circles ( Figure 2A ) . For the inversion analyzed by iPCR with blunt ends ( HsInv0286 ) , we observed a similar trend for the two fragment sizes , although amplification was too low for accurate quantification . Next , using one of the optimal DNA dilutions ( 0 . 62 ng/µl ) , we found that the addition of different reagents did not produce an improvement in the self-ligation step of staggered-end fragments . In blunt-end fragments there was an increase in amplification when PEG , glycerol and glycogen were used ( Figure 2B ) . In particular , glycogen showed a 6 . 4–15 . 7-fold increase in the self-ligation of 33 kb and 41 kb fragments . Also , we quantified the possibility that the iPCR target region is produced by ligation between two different fragments using as a control the AD product , which could be generated only by the formation of concatamers . However , AD amplification was nonexistent from blunt-end fragments and very low and constant throughout the dilutions and the reagents tested for staggered-end fragments ( Figure 2 ) . Based on the results obtained in the optimization experiments , we developed iPCR assays to validate the two breakpoints of the previously selected inversions , except for the three regions not correctly assembled in the HG18 sequence . For the remaining 19 inversions , the restriction enzyme target site and the primers were selected outside of the refined breakpoint regions involved in the inversion ( Table S1 ) , although in three cases ( HsInv0124 , HsInv0397 , and HsInv0272 ) one of the primers and restriction sites were located in a divergent region between the two IRs . Overall , we used seven different enzymes and the size of the self-ligation fragments varied between 2 . 3 and 73 kb ( Table 1 and Table S1 ) . As a first validation step , for all inversions we did multiplex iPCRs of both potential orientations of the two breakpoints ( AB or CD fragments and AC or BD fragments ) in the nine individuals of diverse origins used to predict the inversions by PEM [23] . Figure 3 shows the amplification results of two inversions as an example . We did not observe the expected bands in undigested or unligated genomic DNA controls in any inversion . For many of the iPCR assays we also tested the amplification of the AD fragment concatamer , and no or very little amplification was found . The iPCR genotypes from both breakpoints were the same in 16 of the 19 inversions ( Table S2 ) , but in the other three there were some discrepancies . In HsInv1051 there is no BD amplification in the only individual heterozygous for the inversion according to the fosmids and the AC iPCR ( NA19240; Table S2 ) , although this case is particularly challenging because it corresponds to the largest self-annealing circle we have tried to amplify ( 73 kb ) . Similarly , in HsInv0397 there is one individual ( NA12156 ) with different genotype for BP1 ( Std/Inv ) and BP2 ( Std/Std ) . However , according to the fosmid information the individual should be heterozygous for the inversion ( Table S2 ) . In HsInv0241 three samples appear as Std/Inv in BP1 and Inv/Inv in BP2 , and the same results were found with a new set of primers for the CD region . By genotyping known SNPs within the HsInv0241 inverted region through sequencing and restriction enzyme digestion , we showed that an allele is missing in BP2 in some of the supposedly Inv homozygous individuals ( data not shown ) . Therefore , for the last two inversions it is likely that there is an unknown structural variant or restriction site polymorphism ( either a new restriction site or the elimination of the restriction site used in the iPCR ) that prevents BP2 amplification in one orientation , and that the assay showing the two chromosome arrangements represents the correct genotype . Furthermore , to assess the reproducibility of this technique , for inversions HsInv0344 and HsInv0389 we showed that the same iPCR results were obtained with two different restriction enzymes ( BamHI or EcoRI for HsInv0344 and NsiI or BglII for HsInv0389 ) . Thus , with a good design and the appropriate controls , inversion genotyping by iPCR is quite robust . With regard to the inversion validation , in 17 of the 19 inversions the two genomic orientations were found in the panel of nine humans , but in four of them the Inv arrangement was present only in one of the original Yoruban ( YRI ) individuals used to predict the inversion ( Table S2 ) . In the remaining two inversions ( HsInv0272 and HsInv0526 ) , iPCR of both breakpoints only identified the Std arrangement in all individuals and are likely false positives in the PEM analysis . To check the validity of iPCR results , we compared them with those of the fosmid PEM [23] . To do this , for each individual we considered only the fosmids whose ends map uniquely and outside the IRs at the breakpoints , and those mapping within the IRs were discarded since they are often not informative . For the 17 validated inversions , the genotypes obtained with iPCR and the fosmid PEM data are consistent for all individuals ( Table S2 ) . The false positive inversion predictions HsInv0272 and HsInv0526 were supported just by a few fosmids and all of them have one read mapping in a particular region within the inverted SDs . Sequence analysis has shown that in both cases the incorrect fosmid mappings are caused by variation across individuals in the nucleotide divergence between the two SD copies . For three of the selected candidates in HG18 , HsInv0306 , HsInv0414 , and HsInv0710 , the genomic regions show a different and more complex organization in subsequent releases of the human genome sequence . Therefore , we re-evaluated the inversion predictions taking into account the new data and the preliminary iPCR results . In HsInv0414 , the whole region was duplicated in the HG19 assembly and a 50-kb gap was created , which makes the design and the interpretation of iPCRs very difficult and the region was not analyzed any further . HsInv0710 and HsInv0306 are two overlapping inversion predictions of different length supported by many discordant fosmids . This region has been updated with a new sequence patch ( GL949743 . 1 ) , which has 75 kb of extra sequence that transforms the 15-kb inverted SDs found in HG18 and HG19 into two SD blocks of 109 kb and 95 kb ( Figure S1 ) . To determine if the two inversion candidates are still valid in this context , we re-mapped a set of 1725 concordant and discordant fosmid paired-end reads with mappings spanning the region of interest +/−50 kb ( see Materials and Methods ) . We obtained a total of 20 discordant paired-end reads , and most of the fosmids originally supporting the inversion predictions mapped in highly identical regions within the new SD blocks and were not informative . For HsInv0710 , only one of the 53 fosmids still supports the inversion , although it maps within the inverted SDs with just slightly higher score in the discordant than the concordant orientation . For HsInv0306 , 19 of the 65 fosmids continue to map as discordant in orientation in the HG19 patch ( Figure S1 ) . However , a similar amount of fosmids from all the individuals also support the reference orientation and the 19 apparently discordant fosmids are explained by a ∼16 kb polymorphic deletion of part of SD2 ( Figure S1 ) . Thus , there is not reliable PEM evidence that these inversions exist . Due to the size of the new SDs it was not possible to interrogate the presence of the inversion by iPCR . Nevertheless , we designed several iPCR primers to confirm the organization of the genomic region ( Figure S1 ) . For HsInv0710 , six of the nine individuals are heterozygous for AB and BD ( the other three being homozygous for AB ) and the fragments AC and CD were never amplified . Similarly , for HsInv0306 , the nine individuals showed AB and AC amplification . These results support the existence of big SDs and indicate that the sequence of the new patch is probably correct . In order to genotype inversions in multiple individuals , we set up the iPCR protocol in a plate format . For each of the 17 true polymorphic inversions , we analyzed one breakpoint in 77 individuals , including 68 additional CEPH samples with ancestry from northern and western Europe ( CEU ) and 12 father-mother-child trios . The genotyping worked very well , with on average more than 98% valid genotypes for iPCRs with staggered-end enzymes ( Table S3 ) . For six inversions , we repeated 40–100% of the individuals in independent iPCRs testing the other breakpoint and in all cases the results were the same . The only exception was inversion HsInv0241 , in which both breakpoints were tested for all individuals , and as in the previous experiments , 23 individuals ( 30% ) showed different results in BP1 ( Std/Std or Std/Inv ) and BP2 ( nothing or Inv/Inv ) . As mentioned before , these results are consistent with a missing Std allele in the iPCR of BP2 in these individuals and we considered only the genotypes obtained from AB and BD amplifications . In addition , for inversion HsInv0286 , which involved blunt ends , there was no clear amplification in some DNAs and we could obtain reliable genotypes for 48 CEU individuals ( Table S3 ) . Inversion frequency and heterozygosity on the CEU population varied considerably ( Table 2 ) . Of the 17 inversions , two ( HsInv0832 and HsInv1051 ) were not present in any of the CEU individuals , three ( HsInv0209 , HsInv0340 , and HsInv0341 ) showed very low frequency ( <5% ) , and the rest had a minor allele frequency ( MAF ) between 10–47% ( Table 2 ) . Likewise , the observed heterozygosities ( H ) ranged from 0 to 0 . 55 , and all polymorphic inversions were in Hardy-Weinberg equilibrium ( Table 2 ) . As a control , we also analyzed the genetic transmission of the Std and Inv allele in the available family trios , and in all cases it fits perfectly the expected inheritance pattern ( Table 2 ) . For inversions HsInv0209 , HsInv0340 , HsInv0341 , HsInv0832 and HsInv1051 we could not analyze the inheritance for the Inv allele in the CEU population . Thus , we checked the families of the YRI individuals with the inversion and in all of them the Inv allele was correctly transmitted ( Table S3 ) , showing that the inversion genotypes obtained by iPCR behave as a normal genetic variant . Inversions are known to inhibit recombination and generate genetic substructure with high linkage disequilibrium ( LD ) of the SNPs within the inverted region [5] , [6] . Therefore , we examined the patterns of nucleotide variation within the inverted region and in the 10 kb flanking the 14 inversions with at least two inverted alleles in the genotyped CEU individuals using the SNP data from HapMap ( 46 unrelated individuals ) [52] and 1000GP ( 28 unrelated individuals ) [53] . The HapMap data revealed perfect LD ( r2 = 1 ) between the inversion HsInv0286 and HsInv0347 and several SNPs ( Table 3 and Table S4 ) . In addition , inversion HsInv0396 presented very high LD values ( r2 = 0 . 9 ) , with three SNPs . The 1000GP data allowed us to detect more SNPs with perfect LD for the three inversions mentioned above and HsInv0114 ( Table 3 and Table S4 ) . As expected , most of these tag SNPs were located inside the inverted region , although some of them were located outside as well ( Table 3 ) . Genotype data were also used to calculate the minimum number of shared SNPs between the two arrangements without phase estimation . To minimize the effect of SNP genotyping errors , for the 1000GP data we performed these calculations using all genotypes and only the most reliable ones based on the genotype likelihoods . Surprisingly , for six inversions ( HsInv0124 , HsInv0341 , HsInv0344 , HsInv0389 , HsInv0393 , HsInv0403 ) a high proportion of the HapMap and 1000GP SNPs located within the inversion were shared between the two arrangements ( Table 3 ) . That is , according to the genotype information , these nucleotide variants were polymorphic both in Std and Inv chromosomes , and the shared SNPs were distributed over the whole inverted region . In contrast , a much lower proportion of shared SNPs between arrangements was observed for the four inversions ( HsInv0114 , HsInv0286 , HsInv0347 , HsInv0396 ) for which SNPs in high LD were identified ( Table 3 ) . The remaining inversions did not present such a clear pattern , mainly due to either a very low frequency of the inversion ( HsInv0278 , HsInv0209 ) or a low number of SNPs within the region ( HsInv0241 , HsInv0397 ) . Very similar results were observed when using only the filtered most-reliable 1000GP SNP genotypes ( Table S5 ) . Next , by inferring haplotypes using the PHASE program [57] , [58] , we were able to calculate the nucleotide diversity ( π ) in Std and Inv chromosomes using the 1000GP data . For most inversions , π values were similar in both arrangements , although for HsInv0278 and HsInv0396 , π of Inv chromosomes was more than five times lower ( Table 3 ) . In addition , we used Fst to measure genetic differentiation between arrangements ( Table 3 ) . In general , the most significant Fst values were observed for the inversions with tag SNPs . An exception was HsInv0389 , in which the differentiation is due to the fact that there were two sets of very different haplotypes , that were mainly Std or Inv , although some haplotypes were shared between arrangements . On the contrary , Fst values of the inversions with shared SNPs tended to be low or even zero , consistent with the absence or very small differentiation between orientations . Figure 4 shows the distribution of fixed and shared SNPs along the inverted region . To illustrate the relationship between phased haplotypes within the inverted region , Median-Joining networks and Neighbor-Joining trees based on HapMap and 1000GP SNPs were constructed for all inversions ( Figure 5 and Figure S2 ) . The 14 inversions could be classified into three main groups . For the four inversions with fixed SNPs between arrangements ( HsInv0114 , HsInv0286 , HsInv0347 , HsInv0396 ) , the networks and trees showed that the Std and Inv haplotypes formed separate clusters ( Figure 5A and 5B ) , suggesting that the inversion arose from a unique event . In the case of HsInv0124 , the haplotype network derived from the 1000GP data shows many recombination events , including possible gene conversion between arrangements . However , the network of the inverted and flanking regions ( +/−10 kb ) from the HapMap data ( not shown ) has two clearly separated clusters of Std and Inv haplotypes . Similarly , in HsInv0209 , HsInv0278 or HsInv0397 there were few individuals or a limited number of SNPs in the inverted region to draw reliable conclusions , but a unique origin of the inversion could not be discarded according to existing evidence . In HsInv0278 , there are only two close HapMap SNPs within the inverted region giving rise to three haplotypes shared between Std and Inv . However , in the 1000GP data , the analyzed Inv chromosomes correspond to the same HapMap haplotype and cluster together , which indicates that we are only observing part of the variation . On the other hand , Std and Inv haplotypes of four inversions ( HsInv0341 , HsInv0344 , HsInv0389 , HsInv0393 ) were dispersed on the trees and networks ( Figure 5C and 5D ) . This is consistent with the high number of shared SNPs found between arrangements for all these inversions and suggests that at least some of them have appeared recurrently several times . For example , in HsInv0341 there are three inverted chromosomes included in two highly divergent haplotype clusters of Std chromosomes ( Figure 5C ) , suggesting that the inversion has occurred independently twice . Something similar could have happened in HsInv0393 , in which we observed two main HapMap haplotypes that include mostly chromosomes of one orientation and a few of the other one , although in the 1000GP data Inv haplotypes clustered with Std ones are missed due to the lack of individuals . In HsInv0389 , which was previously found to be recurrent across mammals [59] , there are two highly divergent Std and Inv haplotype clusters , probably derived from an initial inversion event . Within each of these groups , a few haplotypes are found with the opposite conformation suggesting two additional inversion and re-inversion events ( Figure 5D ) . Similar complex networks with divergent clusters of mixed Std and Inv haplotypes could be observed in HsInv0344 , which are compatible with two or three inversions and re-inversions . Lastly , HsInv0241 and HsInv0403 are small inversions that present few HapMap SNPs within the inversion . For the 1000GP set , there were two different haplotypes shared between Std and Inv chromosomes . However , these two haplotypes are just defined by a small number of SNPs separated by 1–2 kb and could be explained either by inversion recurrence or a long gene conversion tract . The iPCR technique is especially sensitive to the loss and generation of additional restriction sites , and given the low quality of most non-human primate reference genome sequences , it is not easy to transfer the assays to other species . However , we examined the orientation of the 17 human polymorphic inversions in non-human primates by iPCR . In most cases , the human iPCR assay could be applied to the other species , but for five regions we had to develop a completely new assay or modify the existing one ( Table S1 and Table S6 ) . Using DNAs from four chimpanzees and two gorillas ( including a father-son pair in each species ) , we obtained reliable results from at least one breakpoint in one species for 16 inversions ( Table 2 and Figure S3 ) , and there was information from both breakpoints for 11 inversions in chimpanzees and five inversions in gorillas . Of those 16 inversions , nine ( 56 . 3% ) were polymorphic in at least one of the non-human primate species , including HsInv0403 that was polymorphic in the three species analyzed ( Table 2 ) . In addition , HsInv0832 showed different orientations in chimpanzees and gorillas . Thus , this provides further support for the existence of several independent origins for these inversions . To complement the above results , we checked the orientation of these regions in current primate genome sequences by dot plot and Blast . Chimpanzee genome orientation agreed perfectly with that obtained in the iPCR , whereas in gorilla there were several discrepancies that could be due to the quality of the current assembly and the difficulty of assembling regions flanked by IRs ( Table 2 ) . By combining the iPCR information with the position of the Denisovan , chimpanzee and rhesus sequences in the haplotype trees ( Figure S2 ) , we could estimate the ancestral orientation for 14 of the human polymorphic inversions ( Table 2 ) . In general , there was a good correspondence between the seemingly oldest arrangement and a higher frequency and increased levels of nucleotide variation in the CEU population . However , there were some interesting exceptions to be studied in more detail . For example , in HsInv0114 the Inv orientation was found in both outgroups and showed a >60% frequency in CEU populations , although both the haplotype tree and the nucleotide variation values suggest that the Std orientation might be ancestral . Similarly , in HsInv0278 the supposedly ancestral Inv orientation showed a lower frequency and less nucleotide diversity than Std chromosomes . Finally , for HsInv0340 the Inv orientation is missing in the CEU population , although it is ancestral according to the gorilla iPCR results . To investigate the possible functional consequences of the inversions we also analyzed their association with genes annotated in the HG18 and HG19 genomes ( RefSeq and UCSC genes , http://genome . ucsc . edu ) , especially those genes located within the inverted sequence or spanning the predicted breakpoints ( Table 2 ) . In total , seven inversions do not contain genes in the inverted or breakpoint regions , and two just change the orientation of genes within the inversion . One of these ( HsInv0389 ) inverts the FLNA and EMD genes and has been associated to a deletion causing Emery-Dreifuss muscular dystrophy [9] , [10] . Here , we provide an easy method to identify the inversion carriers . In the other eight inversions , the predicted breakpoint regions overlap with described genes ( Table 2 ) , although the expected consequences on the genes are variable . In three cases ( HsInv0124 , HsInv0344 , and HsInv0393 ) , the exonic sequences of genes located within the breakpoint region are 100% identical in the two SDs implicated in the inversion . Therefore , if there is an exchange between the SDs , the sequence of the mRNA should not be affected . In four inversions , the SDs contain entirely or in part two genes of the same family . These include two copies of tRNA-Val and tRNA-Leu in HsInv0278 and the AQP12A and AQP12B genes in HsInv0241 , in which as before the exchange between them should not affect the genes . For example , HsInv0241 exchanges the first exon of AQP12A and AQP12B , where all the differences between the two copies are located , and apparently just alters the relative position of the genes . For HsInv0209 , there are two inverted SDs of 7 kb and 94% average identity that include completely the KRTAP5-10 and KRTAP5-11 genes ( 85% identity ) . Due to the lack of Inv arrangement sequence data , the inversion breakpoints have not been refined within the duplications . However , the most likely location of the breakpoints is a 782 bp region with 99 . 9% identity between SDs , which is >2 . 4 kb away from the 3′ ends of both genes . Finally , for HsInv0396 , almost the entire PABPC1L2B and PABPC1L2A genes are included within the possible breakpoints and contain three single nucleotide variants , although they do not produce an amino acid change and appear to be polymorphic between different copies of the same duplication . Thus , again the inversion will just change the relative orientation of the two gene copies by exchanging the last part of the 3′UTR , which is the most divergent between them . The best example of a gene disruption caused by an inversion is that of CCDC144B , which spans BP1 of HsInv1051 . This gene spreads over 87 . 8 kb , with several coding exons at both sides of the SD implicated in the inversion , and the inversion moves the two first exons 200 kb away from the rest . CCDC144B is part of a family with two other members , CCDC144A and CCDC144C , with >99 . 1% identity and very similar exon-intron structure . Nevertheless , whereas CCDC144A encodes a protein of 1365 amino acids , CCDC144B and CCDC144C have different frameshift changes that reduce their coding capacity to 725 and 647 amino acids , respectively . The possible function of these proteins is not clear and it has been considered that both genes could be pseudogenes , although the expression of CCDC144B is supported by multiple mRNAs and ESTs . Interestingly , the HsInv1051 inversion is not present in the CEU population and so far has been found in a single YRI family , which suggests that it has a relatively low frequency maybe related to its effect on the gene . When we looked in detail to the effect of the breakpoints of all the inversions in other mRNAs , spliced ESTs and ENCODE/GENCODE ( version 14 ) gene annotations , there are several non-coding RNAs that could be affected . One clear case is that of HsInv0340 , that truncates the putative LINC00395 RNA supported by three spliced ESTs . However , further work is needed to explore the possible functionality of this RNA .
Information on human inversions has been typically scarce due to the technical difficulties in their experimental validation and genotyping . Here , we describe an optimized protocol to genotype a big fraction of inversions in a fast and high-throughput fashion . Although the iPCR method had been used before to study individual inversions [47] , we have demonstrated that it can be scaled easily to a large number of inversions with high reproducibility , with just minor problems in a few of the assays . In addition , the elimination of intermediate purification steps allowed us to use much smaller amounts of DNA per sample ( <100 ng in front of 1 µg [47] ) . Finally , through the optimization process we have shown that with appropriate high-molecular weight DNA , iPCR can be used for large DNA fragments of more than 64 kb , which is three times bigger than in previous studies [47] . This makes it possible to analyze inversions mediated by long IRs ( up to ∼25 kb in our case ) , which are very difficult to genotype by other methods . Therefore , the iPCR fills a gap in the study of inversions and increases considerably the range of inversions that can be genotyped . Moreover , iPCR could also be useful in the analysis of other structural variants , like translocations , or complex genomic regions in which the exact organization is not clear . One of the main limitations of the iPCR is the availability of restriction sites in the regions of interest that generate fragments of a size that can be efficiently recircularized ( <70 kb ) . In the case of inversions mediated by IRs , this means having restriction sites exclusively outside the repeats and within the inverted region , which could be quite difficult when the inversion is small or has large SDs . In particular , we have seen that the iPCR protocol works very consistently for restriction enzymes with staggered ends , but the efficiency decreases considerably for blunt-end enzymes . Thus , in order to increase the number of inversions that can be analyzed , future improvements of the technique should be directed to increase the efficiency of the recircularization , especially for big fragments and fragments with blunt ends , and the possibility of making directed cuts in the regions of interest . In addition , the iPCR design relies on a good sequence assembly , and compared with other genotyping techniques , it is more sensitive to changes on the overall organization of the region . For example , the iPCR amplification could be affected by indels or structural variants , which could create new restriction sites or modify the size of the resulting fragments , or other types of restriction site polymorphisms . Therefore , for complex and highly-variable regions it might be difficult to interpret the results , and it is important to check the consistency between the two breakpoints . We have observed differences between the results of the two breakpoints in a few individuals for three of the 17 validated inversions , although in all cases we were able to deduce that an allele was missing in one of the assays . Thanks to the iPCR , we have confirmed the organization of a complex region that was incorrectly assembled in the human genome and invalidated two inversion predictions caused by sequence differences between SDs . Moreover , we have validated 17 polymorphic inversions , showed that all of them behave as normal genetic variants , and obtained a first estimate of their frequency in the CEU population . This represents one of the biggest studies of human inversions both in the number of inversions and individuals analyzed . Only one of these inversions , HsInv0389 , had been previously detected at a frequency of 18% in a sample of 108 chromosomes from individuals of European descent [9] , which is very similar to the frequency we obtained ( 17 . 6% ) . Interestingly , two inversions ( HsInv0832 and HsInv1051 ) were not found in CEU individuals and their frequency could differ between populations , as it has been shown previously by SNP inference for the inversions in 17q21 . 31 [14] or in 8p23 [21] . Future studies of more individuals from other populations would provide a clearer picture of the worldwide distribution of these inversions . Traditionally , especially from studies in Drosophila , it has been assumed that inversions have a unique origin and are monophyletic [60] . Consistent with this , our analysis of the association between the inversion and SNP genotypes has shown that several of them have probably a unique origin and are labeled by tag SNPs , at least in the CEU sample . However , for four other inversions the same region appears to have gone through several inversion and re-inversion events in the human lineage . These inversions are characterized by: i ) the absence of fixed SNPs and a very high amount of shared SNPs between arrangements , reaching in some case the totality of the SNPs present in the Inv arrangement ( Figure 4 ) ; and ii ) Std and Inv haplotypes spread on the networks and trees with sometimes more than one shared haplotype ( Figure 5 ) . There are two additional inversions that show this same pattern , but the number of SNPs affected is too small to reach any conclusion . To make sure that these results are not affected by genotyping errors , all the individuals indicating inversion recurrence were genotyped at least twice for both breakpoints , especially when recurrence was based only on a few individuals , and their identity was confirmed by microsatellites . In addition , for the four inversions showing the highest number of shared SNPs in the 1000GP data , ∼20% of them were re-sequenced and confirmed . Finally , it is worth mentioning that , in most cases , inversion recurrence was based on independent SNP data from HapMap and 1000GP , and it was supported by both simple genotype data and haplotype ( phased ) data . In particular , the HapMap haplotypes were inferred using trio information , which minimizes phase errors , and were consistent with those obtained for the 1000GP data . Assuming that the occurrence of identical independent mutations across several positions is very unlikely , the only other mechanism besides inversion recurrence that could explain the level of genetic flux observed between arrangements is recombination ( either gene conversion or double crossovers ) . Current recombination estimate in humans is ∼23 crossovers per cell ( approximately one per chromosome arm ) [61] , which given the size of the inversions and the phenomenon of crossover interference [62] makes the possibility of double crossovers within the inverted region very small . In addition , gene conversion tracts in mammals are usually short , extending for only a few hundred base pairs , and are rarely longer than 1 kb [63] ( although some studies suggest the existence of gene conversion events up to 22 or 53 kb; see [61] , [64] ) . Similarly , a recent high-resolution recombination map in Drosophila melanogaster found an average gene conversion tract length of 518 bp , with a 95% confidence interval in most chromosomes of less than 800 bp [65] . In fact , there is evidence of small gene conversion tracts that explain the limited number of shared SNPs between arrangements in some of the monophyletic inversions in this and other studies [21] , [64] . However , the existence of many shared SNPs and identical haplotype blocks in Std and Inv chromosomes along the whole inverted region ( Figure 4 ) , with sizes between 5 . 6 and 37 . 6 kb , contrasts with the pattern observed in inversions mediated by non-homologous mechanisms ( David Vicente and Mario Cáceres , unpublished data ) and strongly suggests the recurrence of the inversions . Unfortunately , it is not possible to check if the breakpoints of the diverse inversion events are different since they all occur in regions of high sequence identity . Moreover , when a few chimpanzees and gorillas were analyzed , we found that nine inversions were polymorphic in at least one species and another one showed different orientations between them , suggesting that they have occurred independently in these lineages . An alternative possibility is that these inversions were shared polymorphisms from the common ancestor . However , preliminary estimates of the age of the inversions in humans indicate that they are less than 350 , 000 years old , and the more than 6 million years of divergence between these three species makes this explanation unlikely . Together with the mammal recurrent inversion HsInv0389 [59] , human inversions with multiple origins in primates include the six inversions showing a high number of shared SNPs between arrangements in humans , plus five that seemed unique or were not polymorphic in the CEU population . Therefore , this brings the total number of inversions showing signs of recurrence within humans or between different species to 11 of the 16 that could be analyzed ( 69% ) . Previously , recurrence of SVs was known for those causing genomic disorders , like the inversion causing hemophilia [66] . For polymorphic inversions found in natural populations , the recurrence of inversions mediated by NAHR between SDs had already been postulated by comparison of different lineages in mammals [59] or primates [21] , [67] , and there was some evidence that it could occur in humans as well [12] , [68] . In addition , experimentally it was suggested that some inversions could appear repeatedly in human cells [69] , [70] , although these results should be confirmed with independent techniques . Nevertheless , inversion recurrence had never been demonstrated to the extent shown here , with in some cases signs of several inversion and re-inversion events of the same region both within and between species . Interestingly , the possible recurrent inversions are a representative sample of all the analyzed inversions in terms of size and breakpoint features . Thus , high levels of recurrence could be a characteristic of inversions mediated by large IRs , and it would be interesting to see if even more inversions will show a polyphyletic origin when additional human populations or individuals of other species are analyzed . The observed high incidence of inversion recurrence has two important consequences . First , for many of the inversions mediated by IRs , which are probably a large fraction , genotypes could not be inferred from SNP information but must be resolved experimentally . Second , genome-wide association studies based on SNP genotyping would miss the phenotypic effects of most inversions . In general , according to our bioinformatic predictions , the functional consequences of the studied inversions on the flanking genes is expected to be small , with 41% in which there is not any gene around the breakpoints and many others that do not affect the gene mRNA . However , there are two inversions that disrupt a possible coding gene and a long non-coding RNA , and additional experiments are needed to analyze the effects on the expression of these genes across multiple tissues . Also , it would be interesting to check other possible effects of the inversions on more remote genes . The availability of reliable assays to genotype inversions would allow us to make associations of inversions and gene-expression levels . In addition , it makes possible to genotype groups of human samples with particular phenotypes . Therefore , this study not only gives us a better idea of the distribution and evolutionary history of inversions in human populations , but opens the possibility to determine their functional consequences in the near future .
We used a total of 83 human samples from the HapMap project [52] ( with the exception of NA15510 ) , including 70 CEPH individuals with ancestry from northern and western Europe ( CEU ) , 10 Yoruban individuals ( YRI ) , and two individuals from China ( CHB ) and Japan ( JPT ) ( Table S3 ) . The 70 CEU samples corresponded to 46 independent individuals , with 12 parent-child pairs , and 12 complete trios . Genomic DNA was obtained from Epstein-Barr virus-transformed B-lymphoblastoid cell lines of each individual ( Coriell Cell Repository , Camden , NJ , USA ) . DNA was extracted from a 40-ml cell culture grown according to the recommended procedures using standard phenol-chloroform extraction protocol with modifications to obtain high molecular weight DNA [71] . Identity of all the DNAs extracted was confirmed using the MSK microsatellite kit ( Coriell Cell Repository , Camden , NJ , USA ) . In addition , DNA samples from four chimpanzees and two gorillas were also used . DNAs from both gorillas and chimpanzee N457/03 were isolated from frontal cortex brain tissue obtained from the Banc de Teixits Animals de Catalunya ( BTAC , Bellaterra , Barcelona , Spain ) . The remaining chimpanzee DNAs were extracted from Epstein-Barr virus-transformed B-lymphoblastoid cell lines generated from blood of three individuals from the Barcelona Zoo . All procedures involving the use of human and non-human primate samples were approved by the Animal and Human Experimentation Ethics Committee ( CEEAH ) of the Universitat Autònoma de Barcelona . Inversions to test were selected from the analysis of the fosmid PEM data from Kidd et al . [23] using our own inversion predicting algorithm GRIAL , as described in the InvFEST database [54] . To identify and refine the IRs possibly involved in the rearrangement , the sequence of the predicted inversion in the HG18 genome version was self-aligned with BLAST at the NCBI website [72] . In addition , whenever possible , the position of the breakpoint regions was delimited by multiple sequence alignment of the IRs from the Inv and Std arrangement using available human genome and whole fosmid sequences [54] . The breakpoint boundaries were defined by three or more contiguous paralogous sequence variants ( PSVs ) that get exchanged between the two copies of the IRs in the individuals with the inversion . When necessary , re-mapping of the fosmid paired-end reads to HG18 , HG19 or individual genome patches was carried out using the SMALT v . 0 . 6 . 1 program ( http://www . sanger . ac . uk/resources/software/smalt/ ) . Reads were mapped independently and for each one we kept track of the top 10 mappings with >90% identity , Smith-Waterman scores above 300 , and differing <5% from the hit with highest score . Fosmids with two mapped reads were classified into three categories: concordant , discordant in orientation , and ambiguous . Concordant pairs map with the expected orientation ( +/− ) . Discordant pairs are candidates to target an inversion breakpoint and one of the reads shows the opposite orientation ( +/+ or −/− ) . We consider a pair to be ambiguous when one or both reads map to a copy of a highly-identical IR region ( defined using MEGABLAST as sequences spanning ≥1 kb and ≥97% identity ) and it has both concordant and discordant alternative mappings . For the iPCR assays , we selected restriction enzymes with target sites outside the IRs and at each side of the inversion breakpoints that create fragments between 2 . 3–73 kb using NEBcutter [73] ( Table 1 and Table S1 ) . Then , for each inversion , at least four primers ( Table S6 ) pointing outwards close to the restriction site at each side of the fragment were designed with Primer3plus [74] to amplify the AB and CD fragments ( Std orientation ) and the AC and BD fragments ( Inv orientation ) ( Figure 1 ) . Both primers and restriction sites were checked against NCBI dbSNP Build 137 to avoid differences in the assay due to genetic variation between individuals . In addition , to ensure PCR specificity , no more than one primer in each pair was located in repetitive regions and no reliable amplification was predicted in the human genome by the NCBI Primer-BLAST tool [72] . To optimize the iPCR process , 2600 ng of NA18517 genomic DNA were digested overnight in a 130-µl volume reaction containing 40 U of BamHI ( Roche ) or SwaI ( New England Biolabs ) at 37°C and 25°C , respectively , followed by heat inactivation of the enzymes at 65°C for 15 min . Digested DNA was self-ligated for 3 . 5 hours at 25°C in six reactions ( 400 ng each ) using 400 U of T4 DNA ligase ( New England Biolabs ) in different final volumes to obtain DNA concentrations of 10 , 5 , 2 . 5 , 1 . 25 , 0 . 62 and 0 . 31 ng/µl . The ligase was inactivated at 65°C for 10 min , and the final volume of all reactions was standardized to 1280 µl with 1× ligation buffer . Then , DNA was purified once with phenol-chloroform and chloroform-isoamyl alcohol , precipitated with 2 . 5 volumes of ethanol and 0 . 1 volumes of sodium acetate 3M pH 5 . 2 , and resuspended in 50 µl of water . To assay the effect of different chemical reagents , the process was the same , with the exception that 640 µl ligation reactions containing 0 . 62 ng/µl of DNA were carried out with 2% polyethylene glycol ( PEG ) , 10% glycerol , 1 µg/ml glycogen or 1 µg/ml bovine serum albumin ( BSA ) . Two ligation reactions without any added reagent were used as control . Self-ligation efficiency at different DNA concentrations or chemical reagents tested during iPCR optimization was compared by quantitative real-time PCR using the iTaq SYBR Green Supermix with Rox ( BioRad ) in an ABI Prism 7500 Real-Time PCR System ( Applied Biosystems ) . NA18507 circularized DNA in each ligation was amplified with iPCR primers specific for the two inversion arrangements , using the CLIP2 gene as a reference to control for DNA differences ( Table S6 ) . Amplification conditions were an initial denaturation step of 2 min 45 s at 95°C , followed by 40 cycles of 95°C for 15 s and 60°C for 50 s , and a final dissociation step . For each condition , PCR amplifications were done in triplicate from approximately 13 ng of purified DNA . Real-time PCR results were analyzed using the Sequence Detector and Dissociation Curve programs ( Applied Biosystems ) . Relative ratio quantification was calculated by the Pfaffl method [75] using the Ct values and efficiency of each primer obtained from the standard curve . For high-throughput iPCR , reactions were performed in a 96-well plate with slightly different conditions for fragments with staggered and blunt ends . Typically , 100–150 ng of genomic DNA were digested under conditions recommended by the manufacturer overnight in a 25 µl reaction with 3 U of staggered-end restriction enzymes ( EcoRI , HindIII , SacI , BamHI , Roche; NsiI , BglII , New England Biolabs ) . Alternatively , 500 ng of genomic DNA were digested overnight in a 25 µl reaction with 10 U of the blunt-end enzyme SwaI ( New England Biolabs ) . Enzymes were inactivated at 65°C for 15–20 min ( for BglII 20 min at 85°C ) . Self-ligation of 20 µl of digested DNA was performed for 3 hours at 25°C in a total volume of 175 µl with 120 U of T4 DNA ligase ( New England Biolabs ) . For blunt ends , self-ligation was done in 100 µl with 1 µg/ml of glycogen and 400 U of T4 DNA ligase ( New England Biolabs ) . Ligation reactions were inactivated for 10 min at 65°C . Circular DNA molecules were amplified directly without any further purification in 25 µl PCR reactions with 10 µl ( ∼5–7 ng and 40 ng of DNA for staggered- and blunt-ends , respectively ) of the digestion and ligation mix ( after vigorous vortexing 20–30 s to 3500 rpm ) , 1 . 5 U of Taq DNA polymerase ( Biotherm ) , 0 . 4 µM of each primer ( for multiplex PCR , 0 . 8 µM common primer and 0 . 4 µM unique primers ) , 0 . 2 mM dNTPs , and 1× Taq DNA polymerase buffer without MgCl2 . Amplification was carried out by 5 min at 95–98°C , 33–35 cycles at 95°C for 30 s , 55–61°C for 30 s , and 72°C for 30 s , and a final extension at 72°C for 5 min . PCR products were analyzed by electrophoresis on ethidium bromide-stained 1 . 5–2% agarose gels . It is important to note that other Taq DNA polymerases failed to amplify the unpurified ligation products , probably due to buffer incompatibilities . Inversion allele frequency , heterozygosity , and Hardy-Weinberg ( HW ) equilibrium were calculated for the CEU population sample considering only unrelated individuals with Arlequin v3 . 1 [76] . For chr . X inversions , HW equilibrium and heterorozygosity were calculated from female genotypes . To explore the nucleotide variation associated to the inversions , SNP data in a region from −10 kb to +10 kb from the IRs ( excluding all SNPs within the IRs ) were retrieved from HapMap [52] and 1000GP phase 1 [53] . These two databases were complementary since for the 1000GP fewer CEU individuals were available than for HapMap , although with more polymorphisms . Given the low sequence coverage of the 1000GP data , we performed two separate analyses with all or with only the most reliable SNP genotypes . To obtain a high quality 1000GP genotype set , we compared the likelihood of the given and the next most likely genotype , and discarded the genotypes in which it was not at least 10 times larger . Pairwise LD between polymorphisms was quantified by the r2 statistic using Haploview v4 . 1 [77] . To avoid phasing errors , shared polymorphisms between Inv and Std chromosomes from genotype data were conservatively estimated based on the presence of polymorphic SNPs in Std/Std and Inv/Inv homozygotes or Std/Inv heterozygotes homozygous for both alleles of a SNP or for an allele polymorphic in one orientation and not in the other one . Inference of the haplotype phase from HapMap and the 1000GP SNP data was carried out with PHASE v2 . 1 [57] , [58] adding the inversion genotypes at the locations of the two breakpoints and using the available trio information when possible . We considered the inferred Std and Inv haplotypes as two sub-populations and differences between them were evaluated by means of Fst values with Arlequin v3 . 1 . To investigate the relationships between haplotypes , we generated Median-Joining networks with Network 4 . 611 [78] . Neighbor-Joining trees were built with the Phylip v3 . 69 package [79] using the available alignments from the chimpanzee and rhesus genomes ( Ensembl Release 66 ) or Denisova hominin genome [80] as outgroups . Measures of nucleotide diversity ( π ) were calculated with DnaSP version 5 . 10 . 1 [81] .
|
Inversions have been an evolutionary biology model for almost a century , and recently the discovery of a high amount of structural variation in multiple organisms , including humans , has renewed the interest in them . Since early on , it was shown that they were adaptive and that they were involved in human diseases . However , in humans , the study of inversions has lagged behind due to important limitations in the experimental methods to analyze them . Here , we have optimized a technique for high-throughput validation and genotyping of inversions mediated by inverted repeats . By genotyping 17 of these inversions in a diverse sample of human individuals , including many of European origin and several non-human primates , we have carried out the most complete genotyping effort of human inversions to date . The results of our study indicate that a high proportion of these inversions are recurrent and have occurred multiple times during evolution . This represents an example of the plasticity of the genome and opens a new paradigm in the study of inversions , challenging the common view that inversions have a unique origin .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genomics",
"genetic",
"polymorphism",
"genome",
"evolution",
"genetics",
"population",
"genetics",
"comparative",
"genomics",
"biology",
"human",
"genetics",
"evolutionary",
"biology",
"population",
"biology",
"structural",
"genomics"
] |
2014
|
Validation and Genotyping of Multiple Human Polymorphic Inversions Mediated by Inverted Repeats Reveals a High Degree of Recurrence
|
The vascular wilt fungi Verticillium dahliae and V . albo-atrum infect over 200 plant species , causing billions of dollars in annual crop losses . The characteristic wilt symptoms are a result of colonization and proliferation of the pathogens in the xylem vessels , which undergo fluctuations in osmolarity . To gain insights into the mechanisms that confer the organisms' pathogenicity and enable them to proliferate in the unique ecological niche of the plant vascular system , we sequenced the genomes of V . dahliae and V . albo-atrum and compared them to each other , and to the genome of Fusarium oxysporum , another fungal wilt pathogen . Our analyses identified a set of proteins that are shared among all three wilt pathogens , and present in few other fungal species . One of these is a homolog of a bacterial glucosyltransferase that synthesizes virulence-related osmoregulated periplasmic glucans in bacteria . Pathogenicity tests of the corresponding V . dahliae glucosyltransferase gene deletion mutants indicate that the gene is required for full virulence in the Australian tobacco species Nicotiana benthamiana . Compared to other fungi , the two sequenced Verticillium genomes encode more pectin-degrading enzymes and other carbohydrate-active enzymes , suggesting an extraordinary capacity to degrade plant pectin barricades . The high level of synteny between the two Verticillium assemblies highlighted four flexible genomic islands in V . dahliae that are enriched for transposable elements , and contain duplicated genes and genes that are important in signaling/transcriptional regulation and iron/lipid metabolism . Coupled with an enhanced capacity to degrade plant materials , these genomic islands may contribute to the expanded genetic diversity and virulence of V . dahliae , the primary causal agent of Verticillium wilts . Significantly , our study reveals insights into the genetic mechanisms of niche adaptation of fungal wilt pathogens , advances our understanding of the evolution and development of their pathogenesis , and sheds light on potential avenues for the development of novel disease management strategies to combat destructive wilt diseases .
Vascular wilts caused by fungal pathogens are widespread and very destructive plant diseases , causing enormous economic losses . The survival structures produced by wilt pathogens may remain viable in the soil for more than 20 years [1] , making them a major constraint on agricultural production . Control of wilt diseases is also complicated by the scarcity of sources of disease resistant host germplasm , and the soil and vascular habitats of the pathogens . Wilts caused by Verticillium species are among the most devastating of these types of diseases . The primary causal agent , V . dahliae ( Vd ) , can cause diseases on over 200 plant species , including numerous economically important food crops , ornamental flowers , trees , and shrubs [2]–[4] . The list of hosts for V . dahliae is continually expanding , as new hosts in diverse ecological niches succumb to the pathogen [5] . Like other vascular pathogens , Vd enters and colonizes the plant vascular ( xylem ) system , disrupting water transport , and causing the characteristic symptoms of wilting , and often vascular discoloration ( Fig . 1C , D and E ) , and death of aerial tissues [2]–[4] . A diverse arsenal of carbohydrate active enzymes , including cellulases and pectin degrading enzymes , may be important for each major phase of the infection pathway . These enzymes may be necessary during penetration of the plant roots to gain access to the plant xylem and to breach the plant defense structures ( tyloses and pectin gels ) released into xylem vessels in response to infection [6]–[8] , and finally at the end of colonization , for the production of large numbers of survival structures in the plant tissue . Additionally , colonization of the xylem vessels requires the wilt pathogens to be adapted so that they may thrive in the xylem fluid , which undergoes diurnal fluctuations in osmolarity , and contains only low amounts of sugars , organic and amino acids , and inorganic ions [9] . We have sequenced the genomes of two closely related species of Verticillium , Vd and V . albo-atrum ( Vaa ) ( Fig . 2 ) . Their shared features include the formation of small , hyaline asexual spores for dispersal , absence of a sexual state , hemibiotrophic life style , and induction of wilting symptoms in a variety of different plants . More importantly , despite the phylogenetic relatedness , these two wilt pathogens differ significantly in host range , and the types of melanized survival structures they form to allow them to persist in the soil . Vd forms microsclerotia ( long-lived survival structures ) that are small clusters of melanized , thick-walled cells ( Fig . 1 A and F ) , whereas Vaa produces melanized hyphae that are referred to as dark resting mycelia ( Fig . 1 B ) . The microsclerotia produced by Verticillium dahliae can survive in the soil in the absence of a susceptible host plant , and under inhospitable conditions for more than 20 years [1] , which may have conferred it a competitive edge relative to Vaa by enabling it to disperse and persist in regions inhospitable to Vaa . In addition , both pathogens are generally not host-specific , but individual strains of Vd or Vaa may be differentially virulent on different plant species [5] or show cultivar specificity within a single plant species [10] , [11] . However , Vaa is limited to the more narrow range of hosts in temperate climates , while Vd is well known to have a very broad host range , and to infect over 200 plant hosts from temperate to subtropical climates [2] . Fusarium oxysporum ( Fo ) is another economically important wilt pathogen that infects over 100 plant species in diverse ecological niches worldwide [12] , [13] . Both Verticillium and Fo belong to the subclass Hypocreomycetideae of ascomycete fungi , but are in different phylogenetic lineages ( Fig . 2 ) . Fo shares with Vd and Vaa the ecological niche of the plant vascular system and causes nearly identical disease symptoms , yet differs significantly as it produces wilt symptoms much more quickly , and individual strains of Fo exhibit a high degree of host specialization . Within the Fo species complex over 120 specialized forms ( formae speciales; f . spp . ) have been described based on their specificity to various host species [14] , [15] . The recent Fusarium comparative genomics study revealed that Fo's lineage-specific chromosomes contribute to this strict host-specificity [12] . The comparative study presented here exposes the unique genomic profile of the Verticillium species , characterized by an enhanced capacity for degrading plant pectins . The comparison of the two Verticillium wilt pathogen genomes with that of Fo , the only other fungal wilt pathogen for which the complete genomic sequence is available [12] , also reveals a conserved set of proteins that potentially sustain niche adaptation . Our study also uncovers genomic regions ( genomic islands ) in Vd that are repeat-rich , and may confer enhanced genetic diversity to this primary causal agent of Verticillium wilt . Taken together , this study provides key insights into niche adaptation of wilt pathogens , lays out a foundation for future functional studies , and sheds light into potential directions for development of novel management strategies for controlling wilt diseases .
The whole genome shotgun assemblies of Vd strain VdLs . 17 ( 7 . 5× ) and Vaa strain VaMs . 102 ( 4× ) were generated using Sanger sequencing technology , and assembled using Arachne [16] ( Table S1 in Supporting Information and Methods ) . The current genome assembly of VdLs . 17 comprises 52 sequence scaffolds with a total length of 33 . 8 Mb , and an N50 scaffold length of 1 . 27 Mb ( that is , 50% of all bases are contained in scaffolds of at least 1 . 27 Mb ) . More than 95% of the sequence had quality scores of at least 40 ( 1 error/104 bases ) ( Table 1 ) . An optical map of the Vd chromosomes was constructed using the restriction enzyme AflII . The resulting ∼300× physical coverage map consists of 8 linkage groups , with an estimated genome size of 35 Mb . More than 99 . 7% of the assembled scaffolds aligned to the optical map ( Table S2 in Supporting Information ) , confirming the completeness and accuracy of the genome assembly . The map data enabled anchoring of the genome assembly to the linkage groups , and further allowed analyses of structural variation in the genome . Only 89% of the Vd reads were placed in the current assembly . Interestingly , when all the Vd reads were BLASTed against the assembled genome of Vd , over 97% of the non-ribosomal reads could be mapped onto the assembly . Even though only 4× sequence coverage was generated for Vaa VaMs . 102 ( Table 1 and Table S1 ) , we were able to deliver an assembly of 30 . 3 Mb in 26 sequence scaffolds , with an N50 scaffold length of 2 . 31 Mb ( Table 1 ) . The long continuity in this low coverage assembly was achieved through increasing the coverage with sequence from fosmid clones , and using Arachne-assisted assembly [16] . The assembly also benefited from the low repeat content of the Vaa genome , as less than 200 kb of the Vaa genome can be classified as repeats , whereas over 1 . 68 Mb of the Vd assembly were repetitive sequences ( See Methods ) . Even at low coverage , the current Vaa assembly captures almost all of the genome , as more than 95 . 5% of all sequence reads could be assembled , which is much higher compared to the assembled Vd reads . The genome of Vd strain VdLs . 17 contains 10 , 535 predicted protein-encoding genes ( Table 1 ) , covering approximately 44% of the genome . Strain VaMs . 102 of Vaa contains 10 , 221 predicted protein-encoding genes , covering approximately 41% of the genome . Among the annotated genes , 8699 of the proteins share 1∶1 orthologs between these two genomes , while 1357 and 1102 are specific to Vd and Vaa genomes respectively ( Web File 1 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html , and Fig . S1 ) . The two Verticillium genomes encode numerous carbohydrate-active enzymes , secreted proteins and transcription factors similar to those of other plant pathogenic fungi such as Fusarium spp . and Magnaporthe grisea [17] ( Web Files 2 and 3 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html , Table S3 , and Fig . S2 ) . However , certain gene families , including those among the carbohydrate-active enzymes and secreted protein families , were significantly expanded in the two Verticillium genomes ( Table 2 ) . Such expansion provides a unique genomic signature of these two plant pathogens that live on a wide range of plant material and have an endophytic-like growth phase , living within the host plant for a long time before the disease state becomes evident . The arsenal of potentially secreted proteins ( i . e . the secretome ) of plant pathogens includes key pathogenicity molecules that are generally referred to as effectors . These effectors are molecules that are secreted by pathogens during host colonization and that modulate host biochemistry and physiology , including defense responses , to facilitate host colonization [18] , [19] , [20] . A combination of web-based software tools for the prediction of subcellular localization [21] and signal peptide motifs [22] revealed similar numbers of potentially secreted proteins encoded in each of the Verticillium genomes ( Fig . S2; Web File 3 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) , namely 780 and 759 for Vd and Vaa , respectively . These numbers are comparable to those predicted in other fungi [17] . While 574 of these genes are conserved between the two species , 206 genes are specific to Vd and 185 are specific to Vaa . Since many fungal effectors are small cysteine-rich proteins [23]–[25] , all hypothetical proteins in the Vd and Vaa secretomes were classified based on their size and number of cysteine residues ( Fig . S3 ) . In total , 246 ( conserved ) hypothetical proteins can be designated as small ( <400 amino acids ) , cysteine-rich ( ≥4 cysteine residues ) proteins; 127 for Vd and 119 for Vaa , respectively . More than 60% of these predicted effectors are between 150 and 300 amino acids in size with 4–12 cysteines , typical for fungal effector proteins ( Fig . S3 ) . However , in neither Verticillium genome did we identify orthologs of well-characterized effectors reported in Fo [26] , Phytophthora infestans [27] , [28] , or Cladosporium fulvum [24] , [29] , with the exception of homologs of the C . fulvum LysM effector Ecp6 [30] , [31] and the C . fulvum virulence factor Ecp2 [32] . Under the selection conferred by a constant arms race between pathogens and their hosts , secreted proteins — especially effector proteins — are very diverse in pathogenic fungi . The elucidation of the roles of these potentially secreted proteins in Verticillium species therefore represents a challenging but potentially fertile ground for future functional studies . We observed the expansion of some families of relatively conserved , secreted proteins known to play significant roles in pathogenesis , including LysM effectors [25] , [31] , [33] , and the necrosis and ethylene-inducing-like protein ( NLP ) genes [34]–[36] . In contrast to C . fulvum and other Mycosphaerellaceae fungi that contain three LysM effector genes , the genomes of Vd and Vaa contain 7 and 6 LysM effector genes . Furthermore , while most fungal genomes contain two to three NLP genes , Vd and Vaa have eight and seven NLP gene homologs , respectively . It has previously been shown that NLPs display cytotoxic activity towards dicotyledonous and not towards monocotyledonous plant cells [35] , [36] The expansion of the NLP family , also reported in the Fo genome , may therefore contribute to the broad host range among dicotyledonous plant hosts . Alternatively , some NLP family members may have diverged to exert completely different functions . Furthermore , both Vd and Vaa genomes encode four copies of a gene encoding a cysteine-rich , fungal-specific extracellular EGF-like ( CFEM ) domain , and some proteins containing this domain are proposed to play an important role in virulence and as effectors [37] , [38] . The primary cell wall of dicotyledonous plants consists mainly of cellulose microfibrils embedded in a matrix of hemicelluloses , pectic polysaccharides , and glycoproteins [39] . Degradation of structurally complex pectin molecules requires numerous sugar-cleaving enzymes [40] , [41] . For comparison of the carbohydrate-active enzymes from Verticillium species with those of other fungi ( Table 2 , Web File 2 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) , the boundaries of the carbohydrate active modules and associated carbohydrate-binding modules of the proteins encoded by each fungus in the comparison were determined , and classified using tools available at the Carbohydrate-Active-EnZymes database [42] . These comparisons revealed that despite the overall similar representation of Vd and Vaa carbohydrate active enzymes to those of other ascomycetes , polysaccharide lyase ( PL ) gene families that directly degrade pectin constituents are particularly expanded in Vd and Vaa ( Table 2 , Web File2 , http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) . Among all sequenced fungal genomes , the Verticillium genomes encode the highest number and most diverse types of polysaccharide lyases to cleave different forms of pectins , including pectate lyases in the PL1 , PL3 , PL9 families , and rhamnogalacturonan lyases in the PL4 and PL11 families ( Fig . 3 ) . Interestingly , the PL11 family is present only in the wilt pathogens Vd , Vaa and Fo ( Table 2 , and Web File 2 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) . In addition to the significant expansion of polysaccharide lyase families ( Table 2 ) many other enzymes , such as d-4 , 5 unsaturated α-glucuronyl hydrolase GH88 and GH105 families of enzymes ( Fig . 3 ) that degrade the products generated by polysaccharide lyases [43] , are also expanded in Verticillium . Such enhanced pectinolytic machinery illustrates the enhanced capacity of these species to degrade plant cell walls . Additionally , as pectins are released into the xylem vessels by infected plants and may form a barrier to prevent pathogen movement [7] , [8] , the pectin-degrading enzymes may contribute directly to the advancement of the Verticillium wilt pathogens within plant xylem vessels . The conserved carbohydrate-binding module 1 ( CBM1 ) , generally referred to as a fungal-type cellulose binding domain , is usually appended to a diverse group of fungal enzymes . Through the conserved cysteine motif Cx10Cx5Cx9C , CBM1 anchors the enzyme's catalytic region to insoluble cellulose [44] , enabling attachment to plant cell walls , and likely increasing enzyme efficiency . There are 30 CBM1-appended proteins in Vd ( Table 2 , Web File 2 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html , and Fig . S4 ) . The majority ( 26 ) of these have orthologs in Vaa syntenic regions , reflecting their shared functional importance . Twenty-eight of the 30 Vd CBM1 proteins contain a signal peptide ( Fig . S5 ) , indicating that their enzymatic functions are extracellularly localized . The Verticillium genomes as well as two saprobes , the white rot fungus Phanerochaete chrysosporium and the dung fungus Podospora anserina , encode the highest number of CBM1-containing proteins among the reported fungal genomes ( Table 2 ) [45] . The putative enzymatic functions shared between the Vd and P . anserina CBM1-containing proteins are also remarkably similar , judging by the catalytic domains to which the CBM1s are appended ( Fig . S5 ) . For instance , in Vd there are a total of 18 CBM1-containing proteins encoding glycoside hydrolases , similar to 19 enzymes encoded in the P . anserina genome , and double the number ( 9 ) found in the Fo genome . P . anserina is an efficient saprobe . The shared profile of the CBM1-appended enzymes in P . anserina and the Verticillium species may indicate that the Verticillium species are also highly effective at utilizing diverse substrates for nutritional purposes . The enzymes may even contribute to saprophytic growth of Verticillium species after their emergence from the plant vascular system , and consequently to resting structure production . There is one notable difference between the CBM1-appended proteins from wilt pathogens and those of the saprophytic fungi , namely that the wilt pathogens uniquely encode CBM1-containing polysaccharide lyases ( from polysaccharide lyase families PL1 and PL3 ) . There are three CBM1-containing polysaccharide lyases in Vd , Vaa and one ( from polysaccharide lyase family PL1 ) in Fo , but none in P . anserina ( Fig . S5 ) . Interestingly , P . chrysosporium , which also possesses 30 CBM1-containing proteins , also lacks representative enzymes from polysaccharide lyase families PL1 and PL3 [45] . Therefore , conservation ( for PL1 ) of the CBM1-appended polysaccharide lyase proteins among the wilt pathogens may indicate an important adaptation for the utilization of pectin from the cell walls of live plants , or from gels released into the xylem during infection . Aside from their association with enzymes that degrade plant polysaccharides , an important role of CBM1 domains as elicitors of plant defense responses has been demonstrated experimentally in Phytophthora parasitica [46] , and in root colonization by Trichoderma reesei [47] . Through our comparative study we have identified four such candidates among the Vd CBM1-containing proteins ( Fig . S5 ) for future functional characterization . The two sequenced Verticillium genomes are highly similar . On average , more than 90% of the sequence in any given 10 kb window can be unambiguously aligned to the other genome with an average 92% nucleotide sequence identity . This level of relatedness enabled the generation of large-scale alignments between Vd and Vaa genomes ( Fig . 4 , columns a and b , respectively ) , and the determination of synteny with high confidence ( Methods , Supporting Information ) . However , the genome size of Vd is 2 . 6 Mb larger than that of Vaa assembly ( Table 1 ) . The colinearity of the syntenic maps revealed four regions of about 300 kb each in the genome of Vd , on chromosomes 3 and 4 , that have no synteny with the Vaa genome ( Fig . 4 , circled in red ) , and contribute to the larger genome size in Vd . These four regions are hereafter referred to as Vd lineage-specific ( LS ) regions 1 to 4 for their unique presence in the Vd genome . Nucleic acid hybridizations using probes from four different genes ( one from each of the four LS regions ) revealed substantial genetic variation among the Vd strains tested ( Fig . S6 ) . The four LS regions are repeat-rich ( Fig . 4 column c ) , and the enriched repetitive sequences include DNA transposons , and LINE-like and long terminal repeat ( LTR ) retroelements based on manual curation ( SG Amyotte et al , manuscript in preparation ) . Over 50% of all of the identifiable transposable elements in the Vd genome are found in the LS regions , contributing to an increased repetitive DNA content in the Vd genome ( 8-fold increase compared to that of the Vaa genome assembly ) . The skewed distribution of transposable elements in the LS regions is evident in the distribution of Pfam domains characteristic of the DNA transposon DDE superfamily endonucleases , and the retrotransposon RVE integrases ( Fig . S7 ) . Among the transposable elements in the LS regions are five different LTR ( VdLTRE1–5 ) , and we observed full-length , and actively transcribed copies of these elements in the Vd genome . Homologous sequences similar to elements VdLTRE1–4 were also found in Vaa . However , no significant matches to VdLTRE5 were detected in the Vaa genome assembly or Vaa unassembled sequence reads . In addition , within the Vd genome VdLTRE5 was present only in LS region 3 , suggestive of its recent invasion into the genome . Localized genomic dynamics is also reflected by the presence of genes that were duplicated either singly or in clusters within the four LS regions , with the cluster of seven genes in LS region 1 ( VDAG_02357 . 1 to VDAG_02363 . 1 ) defined as ancestral based on structure and sequence conservation ( Figs . S8 A and B ) . For example , the nucleotide sequences of VDAG_04863 . 1 , VDAG_09199 . 1 , VDAG_09219 . 1 are divergent from that of VDAG_02359 . 1 in the cluster VDAG_02363 . 1-VDAG_02357 . 1 ( Fig . S8B ) . Moreover , while the complete cluster of the seven adjacent genes ( VDAG_02363 . 1-VDAG_02357 . 1 ) is present in LS region 1 , a cluster of four or more highly similar genes appears once in each of the other LS regions ( Figs . S8A ) . The duplication could be the result of the enrichment of repetitive DNA ( Fig . 4 column c ) in these regions that provide localized sequence homology for intra- and interchromosomal recombination [48] , [49] . The presence of such genomic islands and their contribution to genome innovation through duplication , diversification and differential gene loss were also reported in Aspergillus fumigatus [50] . Interestingly , the LS regions are flanked by extensive ( 1 to 5 kb ) AT-rich sequences ( Fig . 4 column c ) , a characteristic of sequences which may have undergone Repeat-Induced Point ( RIP ) -like mutation . RIP has been regarded as a genome defense mechanism in which duplicated DNA sequences are irreversibly altered by G∶C to A∶T transitions , and most notably has been observed following meiosis [51] . Single homologs of the gene encoding the DNA methyltranferase ( DMT ) RID , identified as part of the RIP machinery in N . crassa [52] , were present in Vd ( VDAG_01783 . 1 ) and Vaa ( VDBG_01766 . 1 ) . RIPCAL analyses [53] detected RIP-like mutations among copies of VdLTREs , 2 , 3 , and 4 ( Fig . S9 , and data not shown ) but not in VdLTRE5 , further confirming a different evolutionary history for these elements . The conservation of VdLTREs 1–4 in the two Verticillium genomes , as well as the detectable signatures of RIPed sequences among the elements suggest that VdLTREs 1–4 elements were present in the ancestral species from which Vd and Vaa evolved , and that sexual reproduction existed in the pathogens' history ( SG Amyotte et al , manuscript in preparation ) . However , VdLTRE5 would appear to have integrated into the Vd genome after the divergence of these species , and when sexual reproduction was no longer functional in the Vd lineage . Interestingly , single but different mating type loci , the MAT1-1 and MAT1-2 idiomorphs , were identified in the Vaa and Vd genomes respectively ( Fig . S10 ) , and although both mating type loci ( MAT1-2 and MAT1-1 ) have been observed in Vd isolates tested [54] , [55] , a sexual phase has never been reported for either Vd or Vaa . The genetic flexibility achieved through the LS regions may provide capacity for Vd to rapidly adapt to different host niches . For instance , among the LS encoding genes , we identified two homologs ( VDAG_04894 . 1 and VDAG_04836 . 1 ) of the vdt1 gene ( GenBank Accession AB045985 ) , associated with host range specificity in Vd [56] . Overall , the four LS regions contain 354 predicted protein-encoding genes . Rather than essential ( “housekeeping” ) gene functions , the genes encoded in LS regions are known to play roles in iron and lipid metabolism , environmental stress responses , and potentially secondary metabolism ( Web File 4 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) , as well as pathotype specificity ( VDAG_04894 . 1 and VDAG_04836 . 1 ) . When compared to the core sequences of the VdLs . 17 genome , gene families including those of bZIP transcription factors , ferric reductases , and phospholipases , are significantly enriched in the LS region ( P<0 . 05 ) ( odds ratio analyses [57]; Table S4 in Supporting Information ) . Of the 354 predicted proteins , 25 ( 7% ) were predicted as secreted ( Web File 4 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) , a number that is not significantly different from the overall representation of secreted proteins in Vd ( 7 . 4% ) or Vaa ( 7 . 4% ) . To further validate the potential functional importance of the genes encoded in these LS regions , we analyzed EST sequences generated from three different experimental conditions , and found evidence for expression of 1 , 372 genes . Among those , 23% of the genes encoded in the LS regions were transcribed under the tested conditions , significantly higher ( P = 4e-6 ) than the 12 . 2% for genes located outside of the LS regions ( Fig . 4 , and Methods , Web File 5 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) . Even though the EST data only provide evidence for the functional importance of a small proportion ( 12% ) of the genes in the genome , the randomness of the sampling reinforces the idea that the LS regions observed in this study do not simply serve as the sink of “junk DNA” , but instead encode genes that may be functionally important . Among the expanded gene families , ferric reductase transmembrane proteins have important roles in cell differentiation through production of reactive oxygen species ( ROS ) [58] , [59] , and may influence pathogenic or symbiotic relationships between fungi and their host plants [60] , [61] . In addition to the orthologous NADPH oxidase subfamilies ( NoxA to NoxC ) that are present in both Verticillium genomes , and are shared among fungi , the Vd genome possesses four additional copies of ferric reductase-like proteins that form a distinctive clade ( Fig . S11A ) , suggesting a potentially important role of iron metabolism similar to those suggested for other plant host-pathogen interactions [62] . As a further indication of the importance of iron metabolism in Vd , an iron-binding ferritin ( VDAG_02389 . 1 ) with a potential role in iron sequestration [63] was uniquely present within LS region 1 . Among all four sequenced Fusarium genomes , homologs of this protein are present only in F . oxysporum ( FOXG_16665 , FOXG_16728 ) [12] , and both are located on Fol chromosome 15 , one of the four horizontally acquired chromosomes that are required for pathogenicity on tomato [12] . Members of the expanded basic-leucine zipper ( bZIP ) transcription factor family contain leucine zipper regions that mediate sequence-specific DNA-binding , and are predicted to have a nuclear localization ( Web File 4 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) . Phylogenetically , four of the six bZIP TFs encoded in the Vd LS regions form a distinct clade when compared to those encoded in other regions of the genome ( Fig . S11B ) . With the exception of the gene VDAG_09148 . 1 , which was under positive evolutionary selection , purifying selection for retention of gene function is operating on the other bZIP TFs encoded in the LS regions ( Fig . S12 ) . Apart from the bZIP factors , Vd carries an expanded family of phospholipases which includes a homolog of a patatin-like phospholipase ( PLP; VDAG_02397 . 1; Fig . S11C ) that catalyzes the nonspecific hydrolysis of various lipids , including phospholipids , glycolipids , sulfolipids , and mono- and diacylglycerols [64] . In addition to supplying energy for pathogen growth , lipid metabolism also produces signaling molecules that play crucial roles in intra- and inter-cellular signaling [65] , [66] , [67] . The expansion of the above regulatory factors , both TFs and phospholipases , may contribute/regulate pathogenic traits required for Verticillium wilt development [68] , [69] , [70] . LS region 1 encodes a sequence homologous to the high-osmolarity-glycerol response protein ( Hog1p ) , a well known kinase involved in osmoregulation in Saccharomyces cerevisiae [71] . In yeast this protein is nuclear-localized , and mediates the up-regulation of nearly 600 genes [72] . Almost all ascomycete genomes have a single HOG1 homolog . However , in addition to the HOG1 ortholog ( VDAG_08982 . 1 and VDBG_04396 . 1 in the core genomes of Vd and Vaa , respectively ) , Vd encodes an extra HOG1 sequence ( VDAG_02354 . 1 ) nestled in LS region 1 between LINE-like retroelement sequences . The functional importance of this extra HOG1 homolog is suggested by its expression in both the nutritionally rich complete medium , and during nitrogen-starvation , with a 2 . 5-fold increase in expression level during growth under the nitrogen-starved conditions ( Web file 5 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html , and Methods ) . The two Vd HOG1 homologs have different intron-exon structures , and the phylogenetic analysis of HOG1 from representative ascomycete fungi suggests that VDAG_02354 . 1 is not a duplicate of VDAG_08982 . 1 ( Fig . S13 ) . Overall , the LS regions provide some genetic flexibility , and genes encoded in the LS regions play important roles in signaling/transcriptional regulation and iron/lipid metabolism , processes that are important in host-fungal interactions and pathogenesis . Coupled with the enhanced arsenal of plant cell wall-degrading enzymes in Verticillium genomes , these genomic islands may contribute to the increased genetic diversity of Vd . As specific colonizers of plant xylem vessels , the major water transport system , the wilt pathogens must develop auxiliary osmoregulatory mechanisms to maintain osmotic stability and adapt to this unique ecological niche . Among the broad diversity of the fungal kingdom , fungal species from only four genera are reported to be able to colonize this particular ecological niche and induce wilts . These wilts , all notoriously destructive , include Fusarium wilt caused by members of the F . oxysporum species complex , Verticillium wilt caused by Verticillium spp . , wilt of oak trees caused by Ceratocystis fagacearum , and Dutch elm disease caused by Ophiostoma ulmi and O . novo-ulmi [1] . With a specific interest in identifying potential wilt pathogenicity-related genes , or those that may confer the ability to colonize the plant xylem , BLASTp searches were conducted using BlastMatrix [73] to identify proteins that were common to three sequenced fungal wilt pathogens ( Vd , Vaa , and Fo ) , but absent from the proteomes of F . solani , F . graminearum , and F . verticilloides . We identified 14 such candidates ( Table 3 ) . Extraordinarily , one of the genes identified in the search for wilt-specific proteins encodes a glucan glucosyltransferase closely related to bacterial enzymes involved in production of osmoregulated periplasmic glucans . When exposed to low-osmolarity conditions , Gram-negative bacteria use osmoregulated periplasmic glucans to adjust the osmolarity of their periplasmic space to prevent swelling and rupturing of the cytoplasmic membrane [74] . One of these related bacterial proteins includes the Erwinia chrysanthemi opgH protein , which is required for the production of osmoregulated periplasmic glucans and pathogenicity [75] . Although homologs of the glucan glucosyltransferase gene are widely distributed and well conserved among proteobacteria [74] , [75] , the only eukaryotic counterparts we have identified are those in the sequenced wilt pathogens ( VDAG_02071 , VDBG_03162 and FOXG_02706 ) , and in a fungal pathogen of insects , Metarhizium anisopliae , strain ARSEF 23 [76] . Searches of the NCBI database using BLAST searches ( See Methods ) revealed no homologs in any other fungal genomes or eukaryotic sequences . Phylogenetic analysis of the four fungal glucosyltransferases and representative bacterial OPGH sequences showed that the fungal glucosyltransferases clustered together with 100% bootstrap support , and are most closely related to those of proteobacteria in the order Rhizobiales ( bootstrap value 65%; Fig . 5 ) , supporting a model of horizontal gene transfer . In support of a potential mechanism for horizontal transfer , genetic transformation of Vaa can occur when Vaa and the Rhizobiales bacterium Agrobacterium tumefaciens are co-cultivated at plant wound sites [77] . Interestingly , Metarhizium anisopliae is known to colonize plant roots [76] , and the shared ecological niche and evolutionary lineage of plant pathogenic or endophytic fungi and Metarhizium spp . could potentially have been enabling factors in the acquisition of the glucosyltransferase in these fungal genera . In the alignments of Vd and Vaa genome assemblies , the two Verticillium glucosyltransferases are located in a highly conserved syntenic block ( scaffold 3 on chromosome 3 of the Vd genome ) . If horizontal gene transfer had occurred , it must have happened before the divergence of these two Verticillium species . As for the Fo glucosyltransferase , that gene is located in the core part of the Fo genome ( scaffold 3 on chromosome 8 of Fo assembly ) , in an approx 7 kb region where the synteny breaks down between Fo and its closely related sister species F . verticillioides . The absence of this gene from F . verticillioides and other sequenced Fusarium genomes ( F . graminearum , and F . solani ) suggests that a horizontal gene transfer event may have occurred only in the F . oxysporum lineage , and independently from transfer of the gene homolog into Verticillium . In further support of a model for independent transfer into F . oxysporum , analysis of the 20 kb sequences flanking the predicted F . oxysporum glucosyltransferase ( open reading frame FOXG_02706 ) did not reveal any conservation with sequences flanking the Verticillium spp . glucosyltransferases . To assess the role in V . dahliae of the glucosyltransferase homolog ( VDAG_02071 ) , knock-out transformants were generated in the wild-type strain VdLs . 17 ( Fig . S14 ) . No aberrant phenotype was observed during axenic growth ( See Methods ) , and no significant difference in pathogenicity between either of the knockout strains of Vd , and the wild type VdLs . 17 was observed on lettuce ( Lactuca sativa , plant introduction 251246; P>0 . 05 ) , under soilless pathogenicity assay test conditions ( Table S5 ) . However , a clear difference between the knockout and wild-type strains was observed during pathogenicity tests on Nicotiana benthamiana . At about ten days post-inoculation the first symptoms were observed on plants inoculated with the wild-type strain , and after 12 days unmistakable wilting was observed and the disease rapidly progressed . Upon inoculation with the knock-out transformants disease occurred more slowly , with plants showing less stunting and wilting than those inoculated with the wild-type strain ( Fig . 6; Fig . S15 ) . Thus , the gene is clearly a virulence factor that determines fungal aggressiveness in this host species . As the first sequenced Verticillium species , the analysis of the V . dahliae and V . albo-atrum assemblies provides a genomic profile of the genus , characterized by an impressive arsenal of proteins with pectinolytic activity , and enzymes containing plant cell wall attachment modules . The significant expansion of polysaccharide lyases in the Vd and Vaa genomes is especially revealing considering that pectin gels are usually released by the hosts into the xylem in response to the wilt infections [7] , [78] . Despite these obstructions in the xylem and the fact that pit membranes can effectively prevent the passage of large molecules to adjacent vessels [79] , Vd hyphae are still able to systemically colonize the plant xylem within 1–4 days following inoculation [2] , [80] . Such rapid establishment may rely on the presence of a diverse set of polysaccharide lyases that are able to rapidly breach the barriers around the pit membranes [81] , and the pectin gels around tyloses [7] . Indeed , pectin degrading enzymes have long been suggested to contribute to virulence in Verticillium spp . -host interactions [2] , and although disruption of single pectinase genes in the vascular wilt fungus Fo did not perturb virulence [78] , this lack of effect is undoubtedly due to the functional redundancy of these genes . Both Vd and Fo are able to attack a very broad range of plant species , but different mechanisms are employed to accomplish this . As a species complex , Fo causes wilts of over 120 plant species [15] . However , individual formae speciales of the fungus generally have host ranges restricted to a single family , or even genus of plant [82] . The recent comparative analysis of Fusarium genomes has clearly illustrated that such host specificity is conferred by a few lineage-specific chromosomes which encode genes conferring host specificity in the F . oxysporum species complex , and can be transmitted horizontally [12] . In contrast to such strict host-specificity , Vd is well known for its ability to rapidly adapt to new hosts , and the numbers of plant hosts reported to be susceptible to Vd continues to expand worldwide [2] , [4] . While the machinery that enables Vd and Vaa to interact with the live plant or decaying plant material does not itself appear to contribute to major differences in pathogenicity between the two species , one of the key differences between these two genomes is the existence of more than 1 Mb of structurally flexible sequences within the Vd genome . These flexible “genomic islands” encode important regulatory genes and may enable Vd to rapidly adapt to new niches , as illustrated by the spread of Verticillium wilt on lettuce in California in the 1990s [83] . Overall , the comparative genomic study reported here provides a strong foundation for future studies , such as functional investigations of polysaccharide lyases , and genes encoded in the LS regions .
To highlight the phylogenetic relationships of the vascular wilt pathogens Verticillium dahliae , V . albo-atrum and Fusarium oxysporum and other fungi relevant to this study a neighbor joining tree was constructed using nuclear ribosomal large subunit ( 28S ) DNA sequences . The sequences were retrieved for each species , and a 531 character alignment was analyzed using neighbor joining as implemented in PAUP v . 4 . 0b [84] . Sequences used in the alignments included the following GenBank accessions: Aspergillus niger ( EF661191 ) , Fusarium oxysporum f . sp . lycopersici ( EU214564 ) , F . verticillioides ( AB363766 ) , F . graminearum ( FJ755253 ) , Hypocrea jecorina ( AF510497 ) , Magnaporthe grisea ( AB026819 ) , Neurospora crassa ( FJ360521 ) , V . albo-atrum ( EF543839 ) , and V . dahliae ( DQ470945 ) . The Podospora anserina 28S sequence was retrieved from the P . anserina genome sequencing project ( http://podospora . igmors . u-psud . fr/ ) . The phylogenetic topology obtained was consistent with the one based on larger studies comprising a representative sample of the Sordariomycetes [85] . The fungal strains VdLs . 17 ( ATCC accession MYA-4575 ) and VaMS . 102 ( ATCC accession MYA-4576 ) were isolated from lettuce in California ( CA ) , USA and alfalfa in Pennsylvania ( PA ) , USA [5] , [86] , respectively . Other strains used in this study include VdLs . 16 ( lettuce isolate; CA , 1996 ) ; VdBob . 70 ( cauliflower isolate; CA , 1990 ) ; VdLe . 88 , ( tomato isolate; CA , 1996 ) ; VaaMs . 107 ( alfalfa isolate; PA , 1986 ) ; VdLe . 112 ( tomato isolate; CA , 1997 ) ; VdSm . 113 ( eggplant isolate; CA , 1997 ) ; VdLs . 439 ( lettuce isolate; CA , 2001 ) ; VdLs . 446 ( lettuce isolate; CA , 2001 ) [86]; VdSo . 925 ( spinach isolate , the Netherlands , 2003 ) ; VdSo . 936 ( spinach isolate , Washington State , 2003 ) ; VdLe . 1087 ( tomato isolate , CA , 1970 ) . Unless specified otherwise , cultures of these fungi were maintained on potato dextrose agar ( PDA ) or potato dextrose broth ( PDB ) media at 25°C prior to use . Cultures were maintained long term in closed vials on PDA , or as −80°C stocks in 20% glycerol . Protoplasts of strains VdLs . 17 and VaMS . 102 were produced by overnight incubation in a 5% ( w/v ) Glucanex ( Sigma ) enzyme mixture with buffer ( 0 . 8 M sorbitol , 1 M sodium citrate , and 10 mM EDTA ) , pH 5 . 8 . An Omniprep kit ( GBiosciences ) was used to extract DNA from protoplasts derived from strain VdLs . 17 conidia harvested from PDA plates , and mycelia of strain VaMs . 102 from PDB shake cultures . The PDA or PDB was supplemented with streptomycin ( 50 µg/ml ) , kanamycin ( 50 µg/ml ) and tetracycline ( 50 µg/ml ) for the culture of these fungi . Three EST libraries were produced from strain VdLs . 17 cultured in complete medium , root extract medium , or low nitrogen medium . Complete and low nitrogen media were prepared as described previously [87] , [88] . Root extract medium was prepared by the addition , to 100 ml basal medium [88] , of 5 ml supernatant from a mixture of water and ground root tissue ( 4 . 5 g of ground root tissue per 10 ml of water ) of lettuce cultivar Salinas ( Pybas Seeds , Salinas , CA ) . Three shake ( 150 RPM ) cultures of 100 ml of CM were each inoculated with 1×107 conidia/ml of strain VdLs . 17 , and maintained at 25°C in the dark . At 24 hrs , each of the cultures was centrifuged , washed with water , and resuspended in 100 ml of complete , low nitrogen or root extract medium . After an additional incubation period of 24 hrs , total RNA was extracted from each fungal culture with Trizol reagent ( Invitrogen ) . The cDNA populations were prepared using a SMART cDNA Library Construction Kit ( Clontech ) , normalized using a Trimmer kit ( Evrogen ) according to the manufacturer's instructions , and cloned into pCR2 . 1 ( Invitrogen ) . Whole genome shotgun assemblies of V . dahliae strain VdLs . 17 ( 7 . 5× ) and V . albo-atrum strain VaMs . 102 ( 4× ) were generated with Sanger technology at the Broad Institute using the approach outlined in Table S2 , and assembled using Arachne [16] . To compensate for the lack of genetic mapping information , an optical map [89] of VdLs . 17 was constructed ( Genome Center of Wisconsin , Madison , WI ) . Optical mapping is a single-molecule approach for the construction of ordered restriction maps . It uses large ( 250–3 , 000 kb ) , randomly sheared genomic DNA molecules as the substrate for restriction map construction . By determining the presence of sequence-specific restriction enzyme cut sites and the distances between them , restriction maps of large DNA fragments can be created . Such maps provide a useful backbone for the alignment and verification of sequence data . The VdLs . 17 optical map was constructed using the restriction enzyme AflII and aligned with in silico restriction maps of the genome assembly . The correspondence of the restriction enzyme cutting sites and the predicted fragment lengths have been used to order and orient the scaffolds to the optical map . The Vd optical map corresponds to ∼300× physical coverage and consists of 8 linkage maps with an estimated genome size of 35 Mb . Alignments were made between optical maps and the in silico maps of the sequence scaffolds using map aligner software developed at the Broad Institute . The assembled sequence scaffolds were ordered and oriented , and gaps were estimated ( Table S2 ) . The optical linkage group maps for V . dahliae strain VdLs . 17 can be accessed at http://www . broad . mit . edu/annotation/genome/verticillium_dahliae/maps/Index . html . Protein-encoding genes were annotated using a combination of manually curated genes , in addition to EST BLAST alignments , and ab initio gene predictions made by FGENESH , FGENESH+ ( http://linux1 . softberry . com ) , and GENEID ( http://genome . crg . es/software/geneid ) . Additionally , protein-encoding genes were predicted based on BLASTs of known genes available in public databases . BLAST matches with E values<1e-10 were considered to be usable BLAST evidence . HMMER [90] searches were also performed using the Pfam library to find Pfam domains on six-frame translations of the genomic sequences . Initially , subcellular localizations for all Vd and Vaa proteins were predicted using the WoLF PSORT software ( http://wolfpsort . org; [21] ) , resulting in identification of 1383 putative extracellular Vd proteins and 1 , 310 putative extracellular Vaa proteins . Only proteins containing a signal peptide and a signal peptide cleavage site , but lacking transmembrane domains , were selected . To this end , signal peptides and signal peptide cleavage sites were predicted in the set of putative extracellular proteins using the SignalP3 . 0 program [22] , where a final SignalP D-Score cut-off of 0 . 500 was used to increase specificity while retaining sensitivity . Subsequently , all proteins with signal peptides ( 1040 and 966 for Vd and Vaa respectively ) were analyzed for the presence of transmembrane domains using the web servers Phobius [91] and TMHMM ( version 2 . 0; [92] ) . Both servers identified differential , partially overlapping , sets of proteins with putative transmembrane domains . On average Phobius detected 22% more proteins with transmembrane domains than did TMHMM , and about 75% of the predictions were shared between the servers . For further analyses , all proteins with putative transmembrane domains as predicted by either of the two servers were removed from the dataset . For functional classification of the secretomes of both Verticillium species we used a number of resources , including Broad Institute automatic annotations , and Psi-BlastP [93] hits to proteins in the nr database , the Uniprot knowledge database uniref90 , and the Swissprot classified protein database . Furthermore , domain-calling analyses were performed using the Pfam database ( release version 23 ) and HMMER [90] . Subsequently , all results were parsed through BioPerl ( version 1 . 5 ) . All proteins lacking significant BLAST hits ( E-value<1e-10 ) in any of the databases and for which no significant Pfam domain was called by HMMER ( E-value<1E-01 ) were annotated as hypothetical proteins , as were proteins for which an orthologous non-informative hit was found in the genome of the other Verticillium species . Proteins with no significant BLAST hit , but for which a particular Pfam domain was called by HMMER , were annotated as Pfam domain 1-containing proteins . All proteins with significant , though non-informative , hits in any of the BLAST analyses , and no Pfam domain call by HMMER were classified as conserved hypothetical proteins ( hits were considered non-informative whenever their function could not be deduced from the hits , e . g . hits were to proteins with unknown function ) . Finally , all proteins with informative hits with or without recognized Pfam domain were annotated manually , and classified according to their potential function . The carbohydrate-active enzyme catalogs of VdLs . 17 and VaMs . 102 were compared with the corresponding catalogs from Aspergillus niger CBS 513 . 88 , Neurospora crassa OR74A , Magnaporthe grisea 70-15 , Gibberella zeae PH-1 ( ana . Fusarium graminearum ) , Gibberella moniliformis 7600 ( Fusarium verticillioides ) , Fusarium oxysporum f . sp . lycopersici 4286 , Podospora anserina DSM 980 , and Hypocrea jecorina ( Trichoderma reesei ) . The boundaries of the carbohydrate-active modules and associated carbohydrate-binding modules of the proteins encoded by each fungus in the comparison were determined using the BLAST and HMM-based routines of the Carbohydrate-Active-EnZymes database ( [42]; http://www . cazy . org/ ) . The display of the modular structure of the proteins was subsequently done using Flymod ( Lombard , Coutinho and Henrissat , unpublished ) . For CBM1 identification , a total of 37 CBM1 domains were initially identified for Vd using a 31 amino acid sequence of the VDAG_07210 . 1 CBM1 in low stringency ( E-value = 10 ) tBLASTn searches of the Verticillium group database . Thirty of the CBM1 domains resided in a predicted gene model . Further gene annotation corrections were performed manually , and cataloged for VDAG_07289 . 1 , VDAG_01694 . 1 , and VDAG_08156 . 1 , incorporating the CBM1 in the revised gene prediction models ( Verticillium Group , Broad Institute ) . Then the presence of the CBM1 module in each of the 30 predicted proteins was confirmed by searching with the protein sequences against a Pfam library ( http://motif . genome . jp/ ) , with cut-off E-values of ≤1e-6 . The program WoLF PSORT ( http://wolfpsort . org/; [21] ) was used to predict subcellular locations of the CBM1-containing proteins . Sequence alignment of amino acids in the CBM1 domains of Vd was performed using DNASIS MAX v2 . 9 ( MiraiBio , Hitachi Software ) . Comparative searches of CBM1-containing proteins in Fo and Vaa were conducted using BLASTp with each of the identified predicted CBM1-containing proteins from Vd as a query . Only those searches having an E-value cut-off <1e-12 and >50% alignment were recorded as hits , and only the first hit was selected for comparison . For additional analysis , the set of 30 P . anserina CBM1-containing predicted proteins , and 13 Fo CBM1-containing protein sequences were downloaded from the carbohydrate-active enzyme database ( http://www . cazy . org/geno/geno_eukarya . html ) and the Fusarium group database ( Broad Institute ) , respectively , and verified in motif searches with an E value cutoff of ≤1e-6 ( http://motif . genome . jp/ ) . Repeat sequences were detected using Cross_match [94] which searches the genome sequence against itself , filtering for alignments longer than 200 bp with greater than 60% sequence similarity . Full-length transposable elements were annotated using a combination of computational predictions based on BLAST analysis for transposase genes , and manual inspection using the DNASTAR-based GENEQUEST program ( http://www . dnastar . com ) to identify class I element open reading frames and terminal repeats . The genome distribution of repeated sequences was characterized using the sensitive mode of RepeatMasker version open-3 . 0 . 8 , with Cross_Match version 0 . 990329 , RepBase Update 9 . 04 , RM database version 20040702 . For the analyses of repeat-induced point mutations ( RIP ) in VdLTREs 2 , 3 , 4 and 5 , sequences of at least 500 bp in length corresponding to each type of element , were identified in BLASTn searches of the Vd genome ( E value cutoff <1e-5 ) with the type elements . Then the RIPCAL software tool for the automated analysis of RIP [53] was used for all retrieved sequences . For Vaa , BLASTn searches of the genome with the Vd type elements revealed a total of only 45 hits . AT-rich DNA sequences at the junctions of the lineage-specific ( LS ) regions of VdLs . 17 were identified by manual inspections of these regions . The AT-rich sequences included: supercontig 4: 311311–316103 ( LS region 1 , 4797 bp ) ; two sequences in LS region 2 in supercontig 8: 3028301–3029894 ( 1 , 593 bp ) and 3412477–3413645 ( 1 , 169 bp ) separated by approximately 383 kb; Supercontig 23: 219657–220756 ( LS region 3 , 1100 bp ) ; Supercontig 25: 19431–22306 ( LS region 4 , 2875 bp ) . The ferric reductase transmembrane domain-containing proteins were identified in feature searches of the Fusarium and Verticillium group databases ( Broad Institute ) . Additional ferric reductase proteins and those of the NOX classes from other fungi were from Aguirre et al . [58] , and GenBank sequences for each of the accessions were used for phylogenetic analyses . These sequences included: Saccharomyces cerevisiae , NP_014458; Candida albicans , EAK96678; N . crassa , XP_329210; Fusarium graminearum , XP_391371; M . grisea , EAA57330; and M . grisea , EAA56588 . The Claviceps purpurea , CAP12327 and M . grisea , XP_368494 transmembrane domains were added to the analysis since these NADPH oxidases are virulence factors in the respective pathogens [60] , [61] . Vd and Vaa ferric reductase proteins not included in the tree ( Fig . S11A ) were VDAG_06992 , VDAG_07588 , VDBG_05342 , VDBG_05649 , and VDBG_06458 . The protein sequences of patatin-like phospholipases were obtained from Vd , Vaa , and Fo databases by querying the databases using BLASTp with phospholipases predicted by the Broad annotation pipeline . Inspections of the protein alignments of patatin-like phospholipases from the Verticillium group database ( Broad Institute ) revealed major differences in the length and composition of these proteins . Therefore , domains common to patatin-like phospholipases were identified in the Vd and Vaa sequences using motif searches ( http://motif . genome . jp ) . The identified domains of the oxyanion hole and the G-x-S-x-G motif ( including noncanonical ) from each protein were used for the phylogenetic analyses . The phylogenetic analyses included homologous sequences of plants and fungi that were obtained from literature searches , and identified in the tree ( Fig . S11C ) by GenBank accession: F . graminearum , FG06645 . 1; Aspergillus clavatus , XP_001268427; M . grisea , A4QVZ8; M . grisea , ABG79933; S cerevisiae , NP_014044 , and sequences from five plant species ( shown in purple in Supplementary Figure S10 C ) : Gossypium hirsutum , AAX99411; Nicotiana tabacum , AAF98368; Vitis vinifera , CAO61313; Oryza sativa , AAT77905 , Arabidopsis thaliana , NP_180224 . DNA alignments of duplicated lineage-specific ( LS ) sequences ( as shown in Supplementary Fig . S7B ) were performed using DNASIS MAX v2 . 9 ( MiraiBio , Hitachi Software ) . The 354 predicted proteins from the Vd LS regions were downloaded from the Verticillium group database ( Broad Institute ) . The program WoLF PSORT was used to predict subcellular location as described above , and BlastMatrix analyses were performed to identify putative orthologous sequences in other fungal species . BlastMatrix is a modified BLAST program that supports the simultaneous identification in multiple species of genes homologous to a query [73] . The fungal genome dataset archived in the web-based , comparative fungal genomics platform ( CFGP; http://cfgp . snu . ac . kr ) was queried , as were selected stramenopile , plant , protist and animal genomes . The w statistics for odds ratio analyses [57] were calculated to compare the frequency of specific genes of interest within the LS regions , versus the their frequency in the remainder of the genome . Values significantly greater than 1 indicate the preferential ( non-random ) distribution of these genes within the LS region versus the non-LS regions . Transformation was done using the natural log ( ln ) of w , and the error and 95% CI were calculated for lnw . The odds ratios were based on 10 , 535 total predicted proteins encoded in the genome , and 354 total predicted proteins encoded in the LS regions of strain VdLs . 17 . To elucidate potential evolutionary relationships among the bZIP TFs located in the Vd LS regions , a dN/dS analysis was employed . This analysis was done using the Phylogenetic Analysis by Maximum Likelihood ( PAML ) package [95] which estimates synonymous and non-synonymous substitution rates of nucleotide sequences using the pairwise codeml algorithm , assuming realistic evolutionary models . Prior to the dN/dS analysis , codons were reconstructed using Pal2Nal software ( [96]; http://www . bork . embl . de/pal2nal ) . Each of the DNA sequences used as hybridization probes was amplified from genomic DNA of VdLs . 17 by PCR , cloned into vector pCR4-TOPO and sequenced ( MCLAB , San Francisco , CA ) to ensure the correct probe sequence . PCR products were DIG-labeled using the random labeling method ( Roche ) . The DNA probe for sequence VDAG_04871 . 1 was generated with primers 4871F ( 5′-TTTGGCCATCTCAAAAGATGG-3′ ) and 4871R ( 5′-TACTCATCTTGACCTTCTGTCC-3′ ) . The probe for VDAG_05180 . 1 was generated with primers 5180F ( 5′-ACAATGCGGCCCGACGTTTTCG-3′ ) and 5180R ( 5′-AGCTGCACGGCACAACGATGTC-3′ ) . Similarly , primers 9197F ( 5′-AGAGACTGTCCGACACAGGAAG-3′ ) and 9197R ( 5′- CATCAGCTCGCGCAAACAATGG-3′ ) and 09220F ( 5′-GTGCATACTGATACGCAGTTGC-3′ ) and 09220R ( 5′-TGAGTTCCCAGAAAGAGCGGTGC-3′ ) were used for probes VDAG_09197 . 1 and VDAG_09220 . 1 , respectively . For DNA blot hybridizations , the DNA was extracted from the conidia of each of the strains using a bead beater protocol , and RNA was removed with RNAse A ( Promega , Madison , WI ) . Five micrograms of DNA from each strain was digested overnight at 37°C with either HindIII or PstI enzyme ( 10 U/reaction , Promega ) . The entire reaction mix containing digested DNA was loaded onto a 0 . 8% agarose gel for electrophoresis . DNA was transferred to a Zeta Probe Membrane ( Bio-Rad Laboratories , Hercules , CA ) overnight by capillary transfer , using 20× saturated sodium citrate ( SSC ) . Blots were fixed by cross-linking with a UVC500 cross-linker ( Hoefer , San Francisco CA ) , rinsed with sterilized water ( Millipore , Billerica , MA ) , and air-dried at room temperature . Pre-hybridization was done for 4 to 5 hours at 42°C , in 50% formamide , 5× SSC , 49 mM Na2H2PO4 , 2 . 94% SDS and 0 . 2× blocking buffer ( Roche , Mannheim , Germany ) . Hybridization was done overnight ( 16–18 hours ) , using 25 ng/ml DIG-labeled probe in 5 ml pre-hybridization buffer . Membranes were washed twice at 42°C for 30 minutes , with 125 ml 1 mM EDTA , 40 mM Na2H2PO4 , 5% sodium dodecyl sulftate ( SDS ) , and twice for 30 minutes at 55°C with 125 ml 1 mM EDTA , 40 mM Na2H2PO4 , 1% SDS . Membranes were rinsed in 1× washing buffer ( Roche ) two times and once in 1× detection buffer ( Roche ) , and hybrids detected with anti-digoxigenin-AP conjugate Fab fragments ( Roche ) , according to the manufacturer's instructions , and exposure to BioMax film ( Eastman Kodak Company , Rochester , NY ) . To identify potential wilt pathogen-specific proteins , the total protein set from Vd was used in comparative BlastMatrix [73] searches against sequences from the vascular wilt fungi Vaa and Fo , and predicted protein sets from other fungi , including Fusarium solani , F . graminearum , and F . verticilliodes . The same comparison was conducted with the other five genomes ( 15 genome comparisons in total were made via BLAST searches ) , resulting in the classification of the Vd genes into 32 classes ( Web File 6 at http://www . broadinstitute . org/annotation/genome/verticillium_dahliae/SupplementaryPage . html ) . For example , class 1 included proteins that did not display any significant matches of Vd proteins to any from the other five genomes , and can potentially be considered as Vd-specific . The class containing those predicted proteins present in Vd , Vaa and Fo but not present among the other protein sets included 28 candidate proteins ( BLASTp E value cutoff <1e-6 ) . Additional manual screening of these candidate genes was performed by BLASTp analysis ( Verticillium and Fusarium group databases , Broad Institute ) , and by comparison of the protein alignment lengths ( alignment lengths <50% were excluded ) . The manual screening limited the number of potential wilt-specific proteins to 14 . Motif searches ( http://motif . genome . jp/ ) were performed for each of the 14 proteins against the Pfam library , using the program Hmmpfam [90] . tBLASTn and BLASTp searches of the NCBI nr database were performed to identify similar proteins from other organisms ( BLASTp E value cutoff <1e-6 ) , and WoLF PSORT [21] was used to infer subcellular localization of the predicted proteins , as described above . Additional tBLASTn and BLASTp searches of the NCBI ( nr and WGS ) databases were performed using the glucosyltransferase gene ORFs of VDAG_02071 , VDBG_03162 and FOXG_02706 . Twenty kb windows flanking either side of the open reading frames for VDAG_02071 , VDBG_03162 and FOXG_02706 were examined by BLASTn analyses to determine if these sequences were of fungal origin . The maximum-likelihood tree including four glucosyltransferase proteins was constructed employing a maximum likelihood-based package , PhyML [97] . Branch lengths in substitutions per site were calculated using the WAG evolutionary model [98] . The deletion construct for the knockout of gene VDAG_02071 , the glusosyltransferase in Vd , was prepared by Paz et al [99] and used in Agrobacterium tumefaciens-mediated transformation of VdLs . 17 to obtain independent mutant strains , ΔGT-A and ΔGT-B . The knockout of the glucosyltransferase gene in each was confirmed by nucleic acid hybridization ( Fig . S14 ) . The 1643 bp probe , DIG-labeled ( Roche ) as described above ( See Nucleic acid hybridizations ) , was amplified using primers OSC-F 5′-CGCCAATATATCCTGTCAAACACT-3′ and Hyg-F , 5′-AGAGCTTGGTTGACGGCAATTTCG-3′ . Five micrograms of DNA was obtained from VdLs . 17 and the respective mutant strains , and digested with BamH1 enzyme ( 10 U/reaction , Promega ) overnight at 37°C . DNA transfer for the blot , probe hybridization , and DIG detection was carried out as described above ( See Nucleic acid hybridizations ) . Light microscopy analyses of the ΔGT-A and ΔGT-B strains , in comparison to strain VdLs . 17 , was performed to assess microsclerotia and conidia production , and morphology of conidiophores . For reverse transcription-PCR detection of the VDAG_02071 transcripts , RNA was extracted from VdLs . 17 and the mutant strains using the RNeasy Kit ( Qiagen , La Jolla , CA ) with an on-column DNAse digestion . Reverse transcription reactions included 100 ng RNA template and 0 . 5 ug oligo-dT15 , were incubated at 70°C , cooled on ice and then added to 1× GoScript Buffer ( Promega ) , 0 . 5 mM dNTPs , 7 . 5 mM MgCl2 , 40 U RNAsin ( Promega ) , and 1 ul GoScript Reverse Transcriptase ( Promega ) to a final volume of 20 ul . The reaction was incubated at 25°C for 5 min , 55°C for 45 min , and 70°C for 15 min . The amplification reactions included 5 µl of cDNA template , 1× Promega GoTaq Flexi Buffer , 3 mM MgCl2 , 0 . 2 mM dNTP's , 0 . 15 pmol ORF1F primer ( 5′-ATGAGCAACAACATCCTTACACC-3′ ) , 0 . 15 pmol ORFR primer ( 5′-CTCTAGGGTTGAGCCGATG-3′ ) , and 1 . 25 U GoTaq Flexi Polymerase . The thermocycling program included denaturation at 94°C for 3 min , followed by 40 cycles of 94°C for 30 sec , 55°C for 30 sec and 72°C 1 min 30 sec with a final extension of 72°C 5 min . For the β-tubulin control reaction , 2 µl of cDNA was used as template in a total reaction volume of 20 µl , containing 1× GoTaq Green Master Mix ( Promega ) , 200 nM each of primers VertBtF and VertBtR ( Atallah et al . 2007 ) , and 1 . 25 U GoTaq Flexi polymerase . The thermocycling program consisted of a denaturation at 95°C for 3 min followed by 35 cycles of 95°C 15 sec , 63°C 35 sec , and 72°C 30 sec with a final extension of 72°C for 3 min . Two independent glucosyltransferase ( VDAG_02071 ) knockout strains and the wild type VdLs . 17 strain were subcultured on PDA for 10 days at 22°C . Inoculum was prepared by harvesting conidiospores and adjusting the concentration to 106 spores/ml in water . For each experiment six two-week-old Nicotiana benthamiana plants were inoculated with each of the Vd genotypes by dipping the roots for 5 min in inoculum , and transferring the plants into soil . Plants were scored at two weeks post-inoculation for the display of symptoms . The experiment was performed three times with similar results . ImageJ ( http://rsb . info . nih . gov/ij/ ) was used to measure plant height , and data were analyzed using a T-test . Pathogenicity tests on lettuce PI 251246 were conducted using a soilless assay as previously described [100] , except that the inoculum was adjusted to 2×107 spores/ml and each treatment consisted of 15 plants inoculated with water or one of knockout or wild-type genotypes as described above . Dead plants were discarded from the data set for analyses . Data were analyzed using analysis of variance ( ANOVA ) statistics of ranked data using the PROC Mixed procedure of SAS ( Version 9 . 1 , SAS Institute , Cary , NC ) , with the LD_CI macro to generate relative effects ( RME ) for each treatment , and confidence intervals for detection of statistical differences between treatments [101] , [102] . Leaf symptom data were expressed as the proportion of symptomatic leaves per treatment , and the root vascular discoloration data was expressed as the proportion of discolored roots per treatment . The data from the water control was not included in the analysis , but is summarized in Table S5 . For the one-way ANOVA , “isolate” was treated as a fixed effect , and three independent experiments were combined into a single analysis with “experiment” treated as a random variable . The median and maximum percentage of symptomatic leaves and the percentage of plants with root vascular discoloration were calculated for each strain .
|
Vascular wilts are chronic and very often severe plant diseases that cause billions of dollars in annual crop losses . The characteristic wilt symptom is a result of water blockage caused by the colonization and proliferation of pathogenic microbes in the plant xylem , a water-conducting system . We sequenced genomes of two Verticillium wilt pathogens and compared them to the genome of another wilt fungus , Fusarium oxysporum . The shared genomic features among these three wilt fungi suggest the acquisition of homologs of a bacterial glucosyltransferase , involved in adaptation to osmotic stress , through horizontal transfer . Analyses of glucosyltransferase gene deletion mutants in Verticillium dahliae revealed decreased virulence in the host plant Nicotiana benthamiana . Compared to other fungi , both Verticillium genomes encode more plant cell wall degrading enzymes , including those that are able to degrade cell walls of live plants . Between the two closely related Verticillium genomes , we discovered flexible genomic islands in the primary causal agent of Verticillium wilts , Verticillium dahliae . Coupled with the impressive arsenal of plant cell wall-degrading enzymes , these flexible genomic islands may have contributed to expanding genetic diversity for this organism to invade more plant hosts . In summary , our study reveals insights into the evolution and niche adaptation of fungal wilt pathogens and sheds light on the development of novel disease management strategies for combating the destructive wilt diseases .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] |
[
"genome",
"evolution",
"plant",
"biology",
"microbiology",
"genome",
"sequencing",
"plant",
"science",
"fungal",
"evolution",
"microbial",
"evolution",
"plant",
"pathology",
"genome",
"complexity",
"mycology",
"microbial",
"pathogens",
"comparative",
"genomics",
"biology",
"plant",
"pathogens",
"genomics",
"computational",
"biology",
"genetics",
"and",
"genomics"
] |
2011
|
Comparative Genomics Yields Insights into Niche Adaptation of Plant Vascular Wilt Pathogens
|
In developing brain neuronal migration , dendrite outgrowth and dendritic spine outgrowth are controlled by Cdc42 , a small GTPase of the Rho family , and its activators . Cdc42 function in promoting actin polymerization is crucial for glutamatergic synapse regulation . Here , we focus on GABAergic synapse-specific activator of Cdc42 , collybistin ( CB ) and examine functional differences between its splice isoforms CB1 and CB2 . We report that CB1 and CB2 differentially regulate GABAergic synapse formation in vitro along proximal-distal axis and adult-born neuron maturation in vivo . The functional specialization between CB1 and CB2 isoforms arises from their differential protein half-life , in turn regulated by ubiquitin conjugation of the unique CB1 C-terminus . We report that CB1 and CB2 negatively regulate Cdc42; however , Cdc42 activation is dependent on CB interaction with gephyrin . During hippocampal adult neurogenesis CB1 regulates neuronal migration , while CB2 is essential for dendrite outgrowth . Finally , using mice lacking Gabra2 subunit , we show that CB1 function is downstream of GABAARs , and we can rescue adult neurogenesis deficit observed in Gabra2 KO . Overall , our results uncover previously unexpected role for CB isoforms downstream of α2-containing GABAARs during neuron maturation in a Cdc42 dependent mechanism .
During CNS development , GABAergic transmission regulates key steps of neurogenesis and neuronal circuit formation [1 , 2] . A similar role has been observed also for the regulation of adult neurogenesis [3–5] . The diverse effects of GABA can be observed best upon inactivation of specific GABAAR subtypes by targeted deletion of the α subunit variants , impairing cell migration , dendrite formation and synaptic integration [6 , 7] . Furthermore , functional inactivation of the scaffolding protein gephyrin , present at GABAergic synapses , strongly impairs formation and growth of dendrites , presumably by reducing GABAergic transmission [8] . Hence , gephyrin and other postsynaptic proteins represent essential components of downstream signaling and cytoskeletal function associated with GABAergic synapses [7] . In addition to GABAAR and gephyrin , several proteins have been found to be essential for GABAergic synapse structure and function [9] . Among these , collybistin ( CB ) is of special interest , as it has emerged as a key organizer of GABAergic synapses . CB is a member of the Dbl family of guanine nucleotide exchange factors ( GEF ) that , through interaction with gephyrin and the postsynaptic cell adhesion protein neuroligin 2 , contributes to the formation and stabilization of GABAergic synapses [10–12] . CB selectively activates Cdc42 , albeit weakly [13] and TC10 , but their roles down-stream of CB remains enigmatic . Targeted deletion of Cdc42 has been reported not to affect gephyrin clustering [14] , whereas , in vitro , Cdc42 overexpression can rescue gephyrin clustering impaired by the presence of a CB mutant lacking the pleckstrin homology ( PH ) domain [15] . In addition , Cdc42 , either alone or in combination with CB , modulates the size and shape of postsynaptic gephyrin clusters [15] . A major unresolved issue about CB is the functional role of alternative splicing of its mRNA , giving rise to three main isoforms differing exclusively in the C-terminal domain ( CB1-3 ) and having or lacking an SH3 domain close to the N-terminus . While the latter is thought to regulate the catalytic activity of CB , the specific functions of the CB1-3 isoforms are currently not understood [15] . CB3 is the ortholog of human hPEM2 , mutations of which are associated to two cases of X-linked mental retardation without hyperekplexia , associated with epilepsy , anxiety , and sensory hyperarousal [16 , 17] . CB1 , CB2 and CB3 have previously been shown to exert comparable effects on postsynaptic gephyrin clustering when overexpressed in primary neurons [18] . In this study , we investigated possible functional differences between CB isoforms on neuronal development and GABAergic synapse formation in vitro and in vivo , using transfection of primary neurons as well as neuronal precursor cells of the dentate gyrus; and we characterized possible isoform-specific properties of CB1 and CB2 and their interactions with Cdc42 using biochemical assays . Based on previous observations that , upon overexpression , CB isoforms are not restricted to GABAergic postsynaptic sites but distributed in the soma and entire dendritic tree [15] , we reasoned that CB might have distinct functions in synapses and on the cytoskeleton , which might depend also on interaction with gephyrin and/or GABAAR . Our results demonstrate that CB1 and CB2 differ in their regulation of GABAergic synapse formation and dendritic growth in developing neurons . Unique lysine residues in the C-terminus of CB1 determine its short half-life compared to CB2 and explain major functional differences between the two isoforms . Further , we show that CB interaction with Cdc42 regulates its action on the cytoskeleton and is modulated by gephyrin . Paradoxically , silencing CB expression exacerbates Cdc42 function , providing a basis for functional differences between CB1 and CB2 that depend on their differential half-life . Finally , by using targeted deletion of the GABAAR α2 subunit , we demonstrate that CB isoforms act downstream of α2 GABAAR in adult-born dentate gyrus granule cells .
We aim here to understand the functional significance of CB1 and CB2 isoforms for GABAergic synapse formation . CB1 and CB2 isoforms differ primarily in their C-terminus ( 93% overall sequence identity; unique 30 amino acid and 8 amino acid C-terminus , respectively ) , which suggests a specific role for this region of the protein in setting isoform-specific properties . Given the complexity in specifically targeting each of the CB isoforms ( CB1SH3+ , CB1SH3- , CB2SH3+ and CB2SH3- ) individually in vivo , we resorted to an overexpression system to individually express these four cDNAs for dissecting out their functional differences . We cloned CB1 and CB2 isoforms from a rat cDNA library and added either eGFP , mCherry or V5 N-terminal tags ( Fig 1A ) . eGFP-gephyrin was co-transfected with either of the mCherry-CB isoforms into primary hippocampal neurons at 8 days in vitro ( DIV ) to facilitate visualization of GABAergic synapses , and analyzed for changes in density of postsynaptic gephyrin clusters 7 days post-transfection ( DIV 8+7 ) . Representative images of neurons expressing either eGFP-gephyrin alone , or together with mCherry-CB ( blue ) isoforms opposed to vesicular GABA transporter ( vGAT ) -positive terminals ( red ) are shown ( Fig 1B–1F ) . Independent studies have demonstrated that alterations in postsynaptic gephyrin scaffolds are reflected by corresponding changes in GABAergic transmission [19–21] . Here , we confirmed that eGFP-gephyrin clusters along the proximal-distal axis of dendrites are associated with GABAARs by co-labeling for the γ2 GABAAR subunit ( Fig 1B–1F , lower panel ) . Examination of dendritic proximal segments ( up to 40 μm from soma; upper panels ) and distal segments ( 80–120 μm from soma; lower panels ) revealed a differential distribution of eGFP-gephyrin clusters selectively in the presence of CB1SH3+ or CB1SH3- ( Fig 1G–1I ) . Quantification showed that CB1SH3- and CB1SH3+ specifically enhance eGFP-gephyrin cluster density in distal dendritic segments , while CB2SH3- and CB2SH3+ enhance eGFP-gephyrin clustering uniformly throughout the entire dendrites ( Fig 1G–1I; Table 1 ) . These results demonstrate that CB splice isoforms differentially influence gephyrin clustering ( i . e . , GABAergic synapse formation ) in primary neurons in a compartment-specific manner , pointing towards a functional segregation of CB isoforms . In vivo , proximal and distal dendrites of pyramidal cells are targeted by different interneurons; as this compartment-specific organization is not preserved in vitro , we presume that CB1 mainly facilitates synapse formation in immature segments of growing dendrites . As a corollary , we reasoned that for CB1 to act in immature neurons , it should be expressed early during brain development . To test this possibility , we collected total mRNA from rat neocortex , hippocampus and cerebellum at different time-points after birth ( P5 , P9 , P15 , P30 ) and performed quantitative real-time PCR ( qRT-PCR ) analysis of CB1 , CB2 and PanCB transcript levels . Normalizing the samples to PanCB levels allowed us to evaluate the relative ratios of CB1 and CB2 mRNA ( Fig 2A–2C ) . In all samples , CB2 levels were similar to panCB , confirming that it represents the main splice isoform in the CNS across development ( One-way ANOVA F ( 3 , 16 ) = 3 . 9; P = 0 . 02 ) . Nevertheless , the level of CB1 mRNA was consistently higher at the early time-points in the three brain regions tested with a gradual reduction over time . Interestingly , CB1 mRNA in the neocortex even represented almost 60% of the total CB at P5 , and decreased to 18% at P30 ( One-way ANOVA F ( 3 , 16 ) = 299; P<0 . 0001 ) . In contrast , hippocampus and cerebellum samples showed 20% total CB1 mRNA at P5 , reducing to below 10% of PanCB by P30 ( One-way ANOVA F ( 3 , 16 ) = 4 . 2; P = 0 . 0219; One-way ANOVA F ( 3 , 16 ) = 10 . 5; P = 0 . 0005 ) . Thus , CB1 might play a predominant role during the early phase of postnatal neuronal maturation , when rapid changes in dendritic growth and synapse formation are taking place . To explore the functional specificity of CB isoforms in vivo , we moved to adult neurogenesis in the subgranular zone ( SGZ ) , which gives rise to new dentate gyrus granule cells ( GCs ) [22] . This model provides an amenable in vivo system to test in adult mice the influence of signaling downstream of GABAAR on neuronal maturation in a cell-autonomous manner . We used retroviruses encoding eGFP and eGFP-tagged CB1 and CB2 isoforms ( see Materials and methods ) to infect dividing neural progenitor cells in 8–10 week-old mice . Successful labeling of adult-born neurons allowed us to follow their position in the granule cell layer and morphology over a long time span . We first compared the effects of eGFP-CB1SH3- or eGFP expression in adult-born neurons , focusing on neuronal migration and dendrite maturation at 14 , 28 or 42 days post-virus injection ( dpi ) . These time-points represent three distinct phases of maturation of adult-born GCs and allow direct comparison with earlier studies [6] . GCs overexpressing CB1SH3- penetrated less deeply into the GCL , with 80–90% of cells remaining within <20 μm from the SGZ border at each of the three time-points examined , whereas in the control group , >25% of GCs migrated more than 20 μm . The difference was significant at 28 and 42 dpi ( Kolmogorov-Smirnov test; P <0 . 05; Fig 3A–3A” ) , suggesting that CB1SH3- negatively regulates cytoskeletal reorganization required for cell motility . However , all transduced GCs had moved away from the SGZ , indicating that migration per se was not completely impaired . We then tested the influence of eGFP-CB2SH3- overexpression during neurogenesis in an independent batch of mice . Analysis of eGFP-CB2SH3- infected neurons at 14 , 28 and 42 dpi showed no significant migration differences compared to control ( Kolmogorov-Smirnov test; P = 0 . 438 , 0 . 566 and 0 . 89 respectively; Fig 3B–3B” ) . Next , we quantified dendritic complexity by Sholl analysis , which likely reflects on neuronal maturation and cytoskeleton regulation . Overexpression of eGFP-CB1SH3- did not significantly influence dendritic arborization compared to eGFP-control neurons ( Mann Whitney test for AUC; P = 0 . 7712 , 0 . 0979 , 0 . 0852 respectively; Fig 3C–3C” ) . However , eGFP-CB2SH3- overexpression impaired arborization , as seen by the significant reduction in complexity of dendritic tree compared to eGFP-cells ( Mann Whitney test for AUC; P = 0 . 1714 for D; P<0 . 0001 for D’ , P<0 . 0001 for D”; Fig 3D–3D” ) . To confirm this observation , we also determined the total dendritic length of eGFP , eGFP-CB1SH3- and eGFP-CB2SH3- infected adult newborn neurons ( Fig 3E ) . eGFP-CB2SH3- overexpression reduced total dendritic length and terminal dendrite length ( Fig 3F–3G ) . Furthermore , eGFP-CB2SH3- overexpression caused a significant effect of time ( F ( 2 , 66 ) = 52 . 70 , P = 0 . 0001 ) compared to eGFP control ( F ( 1 , 66 ) = 10 . 24 , P = 0 . 021 ) . These observations indicate that CB1SH3- and CB2SH3- isoforms regulate distinct processes during adult newborn neuron maturation , despite overexpression and the presence of endogenous CB isoforms . In order to elucidate the mechanistic basis of these functional differences , we turned to protein biochemistry . We analyzed the protein half-life of CB1 and CB2 isoforms in HEK-293T cells , based on the fact that the C-terminus of CB1 and CB2 contain different lysine residues , which might be targeted by ubiquitination [23] . To achieve this goal , we blocked protein synthesis using cyclohexamide ( 100 μM ) 12 h post-V5-CB isoform transfection and performed Western blotting ( WB ) for V5 to measure the relative abundance of CB isoforms at different time points ( Fig 4A–4D ) . This analysis showed that both CB1SH3+ and CB1SH3- have significantly shorter half-life than CB2SH3+ or CB2SH3- . Interestingly , CB1SH3- exhibited the shortest half-life ( 1 . 8 hr ) and CB2SH3- was the most stable of the four isoforms tested ( 7 . 3 hr ) . We further determined the CB1SH3- protein half-life in primary hippocampal neurons infected with lentiviruses encoding mCherry-CB1SH3- and we obtained a very similar result ( Fig 4E ) , suggesting that CB protein regulation might be conserved between cell types . We then checked whether CB1SH3- and CB2SH3- are substrates for ubiquitination . We treated one set of HEK-293T cells with the proteasomal inhibitor MG132 or vehicle ( DMSO ) and examined for HA-Ubiquitin ( Ub ) conjugation of V5-CB isoforms ( S1A Fig ) . Immunoprecipitation for HA , followed by WB for V5 showed higher Ub conjugation in MG132-treated samples in comparison to control . We confirmed this observation by performing the experiment in reverse . We immunoprecipitated V5-CB and probed for HA-Ub conjugation by performing WB against HA . We could see a distinct increase in HA-Ub conjugation of all V5-CB isoforms in the presence of MG132 ( S1B Fig ) . CB1 harbors two lysine residues in the penultimate position , suggesting that they might regulate its protein stability . We mutated these two lysine residues to arginine in both CB1SH3+ ( 491/492 ) and CB1SH3- ( 431/432 ) isoforms ( Fig 4F ) and determined the half-life of these CB1 lysine mutants in HEK-293T cells as before . We found a significantly increased protein stability for both isoforms ( Fig 4G and 4H ) ; for instance , mutation of lysine ( 431/432 ) in CB1SH3- increased the protein half-life to similar levels as seen for CB2SH3- ( 7 . 6 hr ) , suggesting that these lysine residues on CB1 largely determine protein turnover . The global contribution of C-terminus sequence in regulating protein stability of CB isoforms is still not fully elucidated . Hence , we deleted the unique C-terminus of CB1 to remove the linker and coiled-coil domain , having only the SH3 domain determining two CB isoforms ( Fig 4I and 4J ) . Analysis of CBΔCSH3+ and CBΔCSH3- protein half-life revealed intermediate values between WT CB1SH3- ( 1 . 8 hr ) and WT CB2SH3- ( 7 . 3 hr ) , around 4 hr . This result suggests that the C-terminus sequence and possibly the tertiary protein structure to be major determinants of CB protein stability . Although CB has been described as a GABAergic synapse-specific RhoGEF with Cdc42 as preferred substrate , their functional relationship within neurons remains unclear . As Cdc42 influence on cytoskeleton organization is well documented [24] , we wondered whether CB modulation of Cdc42 could somehow impinge on actin reorganization . To test this possibility , we co-transfected eGFP-CB2ΔPH , eGFP-CB1SH3- or eGFP-CB2SH3- along with VSVG-Cdc42 in HEK-293T cells and examined actin reorganization 12 hr post-transfection using morphological analysis . We stained actin using phalloidin and counted filopodia in co-transfected HEK-293T cells ( Fig 5A–5C ) . Comparison of cells transfected with eGFP , eGFP-CB1SH3+ , eGFP-CB1SH3- , eGFP-CB2SH3+ or eGFP-CB2SH3- revealed an increase in filopodia formation when eGFP-CB isoforms expressed ( Fig 5D; One-way ANOVA , P< 0 . 0001 , Tukey multiple comparison test ) . This observation was unexpected , given that Cdc42CA causes membrane ruffling and Cdc42DN leads to filopodia formation ( Fig 5E–5G ) . In order to clarify how CB might induce filopodia in HEK-293T cells , we co-transfected either VSVG-Cdc42CA together with eGFP-CB1SH3- or eGFP-CB2SH3- . Co-transfection of eGFP-CB isoforms suppressed VSVG-Cdc42-CA phenotype and induced filopodia formation ( Fig 5H and 5I ) , suggesting that CB might sequester Cdc42 irrespective of its activation status and prevent its interaction with actin . We confirmed this hypothesis by co-transfecting a dominant negative ( DN ) VSVG-Cdc42 mutant along with eGFP-CB1SH3- or eGFP-CB2SH3- isoforms ( Fig 5H’–5I’ ) . While VSVG-Cdc42-DN promoted filopodia formation upon transfection , co-expression of eGFP-CB isoforms amplified this phenotype , suggesting a causal link between CB-mediated inactivation of Cdc42 , likely owing to Cdc42 sequestration . We have reported earlier that CB , Cdc42 , and gephyrin can form a ternary protein complex [15] . Hence , we next tested whether gephyrin co-transfection might influence eGFP-CB1SH3- or eGFP-CB2SH3—mediated Cdc42 sequestration . Interestingly , and in line with a role for CB regulating Cdc42 at postsynaptic sites , mCherry-gephyrin co-expression reduced filopodia formation when the V5-CB1SH3- or V5-CB2SH3- isoforms were expressed . Instead , the HEK-293T cells showed ruffled membrane , indicative of Cdc42 activity ( Fig 5J and 5J’ ) . Quantification confirmed that gephyrin co-expression along with VSVG-CB isoforms significantly reduces filopodia formation ( Fig 5K , One-way ANOVA , P<0 . 001; Tukey multiple comparison test ) . To explore further the relationship between CB and Cdc42 we turned to biochemistry . In this assay , we purified bacterially overexpressed GST-Cdc42 and incubated it with γsGTP or GDP to mimic active and inactive forms , respectively . Next , we incubated lysate of HEK-293T cells overexpressing V5-CB isoforms with immobilized GST-Cdc42 γsGTP or GDP . Pull down of GST-Cdc42 , followed by WB for V5-CB , showed CB interaction with both active and inactive forms of Cdc42 ( S2A and S2B Fig ) . This result confirms that CB can interact with both active and inactive forms of Cdc42 . If gephyrin interaction determines CB activation of Cdc42 , then we should find a reduced Cdc42 interaction with CB in gephyrin co-expressing cells . To test this possibility , we co-transfected VSVG-Cdc42 , V5-CB1SH3- or V5-CB2SH3- with or without FLAG-gephyrin ( S2C Fig ) . We bacterially expressed and purified GST-Pak1 binding domain ( GST-PKB ) . GST-PKB interacts with active Cdc42 with high specificity and affinity [25] . Hence , we depleted free activated Cdc42 from HEK-293T lysate by incubating with GST-PKB immobilized on glutathione agarose beads ( S2D Fig , bottom panel ) . We collected active-VSVG-Cdc42 depleted HEK-293T cell supernatant and immunoprecipitated VSVG-Cdc42 using an antibody against VSVG . WB for V5 allowed us to determine the relative levels of Cdc42 bound to V5-CB1SH3- or V5-CB2SH3- in the presence or absence of gephyrin ( S2C Fig ) . These experiments confirmed the reduced Cdc42 interaction with V5-CB1SH3- and V5-CB2SH3- when FLAG-gephyrin was co-expressed . Taken together , these results indicate that , rather than activating Cdc42 , CB isoforms prevent its activation ( by other GEFs ) by sequestrating it . In turn , gephyrin binding the CB reduces the amount of sequestered Cdc42 and thereby the availability of Cdc42 for remodeling of the actin cytoskeleton . Biochemical experiments have shown consistently a stronger interaction between Cdc42 and CB2 than CB1 , both in vitro and in transfected HEK-293T cells . Therefore , the impairment of dendrite growth observed in adult newborn neurons upon eGFP-CB2SH3- overexpression could be due to Cdc42 sequestration . To confirm this idea , we infected newborn neurons with retroviruses expressing eGFP-Cdc42DN and quantified migration distance of these neurons and dendrite formation at 14 , 28 and 42 dpi ( Fig 6A–6A” ) . We did not observe any significant differences in the migration of cells expressing eGFP or eGFP-Cdc42DN , in line with the results presented in Fig 3 . However , we observed a significant reduction in dendrite complexity at 28 dpi in newborn neurons expressing eGFP-Cdc42DN ( Fig 6B–6B”; Mann Whitney test for the AUC P = 0 . 0226 for B’ ) . This phenotype of eGFP-Cdc42DN overexpression was similar to that produced by eGFP-CB2SH3- overexpression , albeit less severe . In view of this finding , we wondered whether the shorter protein half-life for eGFP-CB1SH3- restricted its function when overexpressed in newborn neurons . If this were true , then expressing the eGFP-CB1SH3- K431R/432R mutant , which exhibits a half-life similar to eGFP-CB2SH3- , should also impair dendritic aborization . Hence , we infected newborn neurons with eGFP or eGFP-CB1SH3- K431R/432R mutant retrovirus , and looked for defects in dendritic arborization . Sholl analyses showed reduced dendritic complexity in cells overexpressing eGFP-CB1SH3- K431R/432R at 28 dpi ( Fig 6C and 6C’ , Mann Whitney test for AUC , P = 0 . 0004 for B’ ) . Furthermore , we also observed a significant reduction in the total dendritic length and a reduction in the number of terminal branches in these neurons ( Fig 6D–6F ) . These observations confirm that protein half-life is a major determinant of CB1 and CB2 isoform functional differences . Given that CB1 differentially influences gephyrin clustering along the proximal-distal axis of dendrites in primary neurons ( Fig 1 ) , we wondered whether this effect also was influenced by CB1 stability . We co-transfected DIV 8 neurons with eGFP-gephyrin and V5-CB1SH3+ , or V5-CB1SH3- , or V5-CB1SH3+2R mutant , or V5-CB1SH3-2R mutant , and analyzed the density of postsynaptic eGFP-gephyrin clusters along the proximal-distal axis at DIV8+7 ( S3A–S3E’ Fig , see Table 1 ) . Neurons co-expressing either V5-CB1SH3+ or V5-CB1SH3- showed significantly increased eGFP-gephyrin clustering at distal dendritic segments . However , in neurons co-expressing V5-CB1SH3+ K491R/492R or eGFP-CB1SH3- K431R/432R we saw an overall increase in gephyrin clustering without any specific gradient along the proximal-distal axis ( S3F–S3H Fig , see Table 1 ) . Therefore , these data identify a further role for protein stability in the functional differentiation between CB1 and CB2 isoforms . Upon overexpression in adult-born neurons , our data identifies a role for CB1 in neuronal cell migration and for CB2 in dendrite growth . In order to rule out overexpression artifacts , we reduced endogenous CB expression by PanCB shRNA-mediated silencing to unmask its contribution for regulating neuronal migration and dendrite maturation . We tested several PanCB shRNA sequences in HEK-293T cells and incorporated the most effective sequence in a retrovirus IRES GFP to label infected newborn neurons . We analyzed for migration differences at 21 dpi and 42 dpi as our earlier experiments showed strongest effects during these time points . As control , eGFP alone or scrambled shRNA sequence was used . Expression of PanCB shRNA reproduced the CB1SH3- overexpression phenotype on neuronal migration . Fewer than 70% PanCB shRNA-expressing cells had migrated up to 15 μm at 21 dpi whereas cells expressing either eGFP or scrambled shRNA control had migrated more than 25 μm from the border of the SGZ ( Fig 7A and 7A’ ) . This finding confirms the importance of a delicate regulatory balance to maintain correct levels of activated Cdc42 . Interestingly , PanCB shRNA-expressing neurons showed increased dendritic complexity at 21 dpi , which is opposite of the eGFP-CB2SH3- , eGFP-Cdc42DN or eGFP-CB1SH3- K431R/432R mutant overexpression phenotypes ( Fig 7B and 7B’; One-way ANOVA for AUC , P = 0 . 0154 for B ) . We confirmed this finding by analyzing the total dendritic length in GCs expressing PanCB shRNA ( Fig 7C ) . We even observed distal branches in PanCB shRNA positive cells to revert their radial course when reaching the outer border of the ML , in both the upper and lower blades of the DG ( Fig 7D ) . These results demonstrate that overexpression of CB1 or CB2 isoform into newborn neurons limit dendrite growth presumably via sequestration of Cdc42 . We have reported that CA1 pyramidal cells in Gabra2 KO mice exhibit a reduced frequency of miniature IPSCs , but no change in GABAergic current amplitude [26] . Hence , we infected adult-born neurons in Gabra2 KO mice using a eGFP retrovirus and analyzed for gephyrin clustering defects along the proximal-distal axis at 14 , 28 and 42 dpi ( Fig 8A–8B’ ) . We examined α1 subunit expression to determine whether this GABAAR subtype compensates in Gabra2 KO cells . Staining for the α1 subunit in eGFP-positive newborn neurons in Gabra2 KO at 14 , 28 and 42 dpi showed elevated α1 GABAAR levels at 28 dpi in both proximal and distal dendritic segments ( Fig 8C and 8C’ ) . This increase of α1 subunit clusters was not seen at 42 dpi , suggesting a role for extrinsic factors in shaping inhibition during the maturation process . α2 GABAAR can directly interact with CB and facilitate gephyrin scaffold recruitment at GABAergic postsynaptic sites [12 , 27] . Hence , we examined in Gabra2 KO mice , where we have reported deficit in gephyrin scaffolding both in hippocampal formation and neocortex [19 , 26] , whether CB overexpression can compensate for the absence of the GABAAR in facilitating gephyrin clustering . We quantified gephyrin cluster density in the first 40 μm segment of dendrite from the soma and found a significant reduction in GCS from Gabra2 KO compared to the WT mice . Similar observations were made in distal dendritic segments , as well ( 80–120 μm ) ( Fig 8B and 8B’ ) . To test whether CB1SH3- overexpression could rescue endogenous gephyrin scaffolding in Gabra2 KO cells , we overexpressed eGFP-CB1SH3- using a retrovirus in Gabra2 KO newborn neurons and analyzed for gephyrin clustering in the proximal and distal dendritic segments at 42 dpi . eGFP-CB1SH3- overexpression consistently rescued endogenous gephyrin clustering in GCs from Gabra2 KO mice ( Fig 8D and 8D’ ) . We have reported a direct link between the Gabra2 gene and adult neurogenesis [6] . Hence , we wanted to examine whether CB1SH3- function occurs downstream of α2 GABAAR and infected WT and Gabra2 KO mice with eGFP retrovirus to analyze cell migration differences at 14 , 28 and 42 dpi . α2 GABAARs contribute towards cell migration , as in the absence of the Gabra2 gene product , newborn neurons migrate 60% longer distance away from the SGZ . This difference was highly significant at 14 dpi ( S4A–S4B” Fig; Kolmogorov- smirnov test , p<0 . 0001 ) , but diminished over time , as the WT cells caught up at 28 and 42 dpi . If CB1SH3- functions downstream of α2 GABAARs , we reasoned that it should be able to reverse the migration phenotype observed in adult-born neurons of Gabra2 KO mice . We infected Gabra2 KO mice with either eGFP or eGFP-CB1SH3- retrovirus and quantified migration of labeled GCs at 14 , 28 and 42 dpi . The expression of eGFP-CB1SH3- reduced the migration of the newborn neurons by 60% shorter distance from the SGZ . The migration difference did not normalize at later time points ( S4C–S4C” Fig; Kolmogorov-Smirnov test , P = 0 . 0044 , 0 . 0274 , 0 . 0096 respectively ) . This confirms a role for CB1SH3- downstream of α2 GABAARs and highlights the relevance of CB-mediated regulation of Cdc42 function for proper cell migration .
Our RTqPCR analysis revealed differential regulation of CB splicing during development , favoring CB1 isoform expression at early stages of development , especially in the neocortex . Assuming equal mRNA translation , these data favor a role of CB1 in immature neurons , possibly at the onset of synaptogenesis . However , the short half-life of CB1 relative to CB2 suggests a transient role for this isoform . The half-life of CB1 determines its specific role even upon overexpression in primary neurons , as seen upon mutation of C-terminal lysine residues that render CB1 CB2-like . Therefore , we conclude that ubiquitination of CB1 is a novel mechanism restricting its functional repertoire . It would be interesting to test whether it regulates the interaction of CB1 with Ccd42 , which is overall weaker than that of CB2 with Cdc42 . A pyramidal cell-specific transcriptional program involving transcription factor NPAS4 has been reported to coordinate the distribution of GABAergic synapses by increasing the number synapses on the cell body , while decreasing the synapse number on the apical dendrite [28] . In addition , neuronal cell-specific expression of splicing factors SLM1 and SLM2 has been shown to contribute towards the postsynaptic biochemical diversity [29] . Hence , identification of CB1-facilitated gephyrin clustering selectively in the distal portion of immature dendrites adds to the repertoire of biochemical diversity at GABAergic synapses , complementing the compartment-specific interneuron connectivity . The role of the ubiquitin proteasome system ( UPS ) in synaptic plasticity is an emerging concept at inhibitory synapses . However , at glutamatergic synapse protein ubiquitination and degradation are important determinants of synapse function [30] . In the current study , we highlight the importance of CB1 protein turnover for compartment-specific eGFP-gephyrin clustering . This finding implicates the local availability of CB1 protein pools as a basis for GABAergic synapse distribution along the elongating neuronal dendrite . We also report the importance of UPS in adult newborn neurons in vivo to explain the GABAergic basis for actin regulation and dendrite outgrowth . However , the distribution and regulation of UPS in developing neurons are currently not understood and require further investigations . Strikingly , in wildtype mice , CB2SH3- overexpression and shRNA-mediated downregulation of all CB isoforms had opposite effects on dendrite arborization , suggesting that CB2 activates signaling factors to regulate dendritic growth . A role for small Rho-GTPase , such as Cdc42 ( or possibly TC10 ) downstream of CB2 might explain the enhanced dendrite outgrowth phenotype upon CB silencing . This hypothesis implies , however , that CB inhibits Cdc42 activation . Our data confirms this hypothesis by demonstrating the failure of dendrites to stop growing when they reach the outer surface of the molecular layer upon CB down regulation ( Fig 7D ) . Evidence supports the notion that specific GABAAR subtypes , determined by their α subunit variants , regulate distinct stages of adult neurogenesis in the dentate gyrus [6 , 31 , 32] . Thus , strong genetic and pharmacological evidence points towards the regulation of cell-fate determination by α5 GABAARs , whereas proliferation on precursor cells is modulated by α4 GABAARs and migration of adult-born GCs away from the SGZ is regulated in opposite directions by α4- and α2 GABAARs [6] . Finally , dendritic growth and arborization , as well as GABAergic synapse formation , are modulated , at least in part by α2 and α5 GABAARs [6 , 32] . Our present results implicate CB1 and CB2 downstream of α2 GABAARs for both GC migration and postsynaptic gephyrin clustering , providing a mechanistic explanation how GABAergic synaptic transmission might act on the actin cytoskeleton regulating neuronal mobility and dendrite formation . Overexpression experiments can be challenging to interpret , because the effects might be minimized , or even reverted , due to exhaustion of an essential factor present at a limited level . In this context , the paradoxical observation that the inhibitory effect of CB1 on migration of adult-born GCs could be mimicked upon overexpression of a DN Cdc42 mutant provided the first clue towards the possibility that CB overexpression might affect the regulation of Cdc42 activation . More direct evidence for an inhibitory effect of CB isoforms on actin remodeling mediated by Cdc42 was obtained , therefore , in HEK-293T cells , where the formation of filopodia and membrane ruffling are robust and well-established readout of Cdc42 inactivation and activation , respectively [33] . Cdc42 gene encodes for two splice isoforms , namely Cdc42E6 and Cdc42E7 [24] . The Cdc42E6 isoform is post-translationally modified by palmitoylation; and has been shown to be enriched in dendritic spines to influence spine density [34] . On the other hand , Cdc42E7 isoform is exclusively prenylated [34] , and whose function occurs within axons [35] . Given that CB expression is localized to neuronal dendrites , it is very unlikely to regulate both Cdc42 isoforms in the brain . The results of our experiments are best explained by assuming that binding of Cdc42 to CB limits ( or prevents ) its activation , presumably by other GEFs like TC-10 present in HEK-293T cells . Importantly , this interaction between CB and Cdc42 is modulated by gephyrin , providing an elegant mechanism to explain how CB action on the cytoskeleton at postsynaptic sites might be locally restricted . Our pull-down experiments confirmed that CB2 interacts stronger than CB1 with Cdc42 ( possibly due to CB1 ubiquitination , see above ) and that CB1 and CB2 bind to both activated ( i . e . , GTP-bound ) and inactive Cdc42 ( GDP-bound ) . These biochemical observations are compatible with the hypothesis that , albeit being a weak activator of Cdc42 , CB mainly acts upon the actin cytoskeleton by limiting its availability for other activators of this ubiquitous small GTPase .
Taken together , our results provide a new framework for the differential functions of CB isoforms in neurons . CB1 is an early expressed , short-lived isoform regulated by ubiquitination . It modulates GABAergic synapse formation and regulates neuronal migration downstream of α2 GABAAR . The latter effect involves downregulation of Cdc42 activity by trapping it away from other Cdc42 activators . CB2 is a constitutively expressed variant , with a longer half-life , which promotes dendrite growth and GABAergic synapse formation . The action of CB2 also involves negative regulation of Cdc42 activation , but , based on its distribution in adult CNS [12] , might be localized mainly at GABAergic postsynaptic sites . The Ubiquitin-Proteasome System ( UPS ) , by limiting the action of CB1 ( and possibly its binding to Cdc42 ) , thus appears to be a major regulator of the function of the GABAergic system .
All experiments were performed in accordance with the European Community Council Directives of November 24 , 1986 ( 86/609/EEC ) and approved by the cantonal veterinary office of Zurich . CB isoforms CB1SH3+ , CB1SH3- , CB2SH3+ and CB2SH3- were cloned from rat whole brain RNA using primers Fwd: 5’-ATGCAGTGGATTAGAGGC-3’; Rev: 5’- CTAATAGTGCCATTTTCTTTGG-3’ for CB1 isoforms , and Fwd: 5’-ATGCAGTGGATTAGAGGC-3’; Rev: 5’-CGCTAAGCTTCATGACTCTGCT GATCA-3’ for CB2 isoforms . mCherryC2-CB1SH3+ , mCherryC2-CB1SH3- , mCherryC2-CB2SH3+ , mCherryC2-CB2SH3- , expression vectors were generated by subcloning CB into eGFPC2 vector backbone using HindIII and BamHI . eGFP sequence was later replaced by mCherry sequence using NheI and XhoI sites . pCR3-V5-CB1SH3+ , pCR3-V5-CB1SH3- , pCR3-V5-CB2SH3+ and pCR3-V5-CB2SH3- were generated by subcloning the V5 tag into the pCR3-CB vectors using HindIII restriction site . The eGFP-gephyrin P1 variant has been described previously ( Lardi-Studler et al 2007 ) . FLAG-gephyrin is described previously ( Tyagarajan et al 2011 ) . pGEX-2T-GST Cdc42WT was obtained from Addgene ( #12969 ) and pGEXTK-Pak1 ( 70–117 ) was obtained from Addgene ( #12117 ) . pRKVSV-Cdc42 WT , pRKVSV-Cdc42 DN ( N17 ) , pRKVSV-Cdc42 CA ( QL ) were gifts from Prof . Kenneth Yamada ( NIH ) . CB mutants were generated using site directed mutagenesis according to vendor manual ( Agilent Technologies , USA ) using pCR3-V5-CB splice isoforms as the template . The truncated V5-CBΔCSH3+ and V5-CBΔCSH3- were generated by inserting a Stop codon after Ile 440 ( Ile380 for CBSH3- , respectively ) into pCR3-V5-CB1 isoforms . HA-ubiquitin was a gift from Dr . Teier , Heidelberg , Germany . Primary hippocampal neuron cultures were prepared as described previously in [36] . Hippocampal cultures were transfected with eGFP-gephyrin and mCherry-CB or V5-CB construct ( 500 ng of each plasmid ) , using a combination of Lipofectamine 2000 ( Life Technologies ) and CombiMag ( OZ Biosciences ) . The neurons were grown in 2 mL of growth media for 8 days prior to transfection . The transfection mix was incubated at room temperature for 15 min before adding to the neurons . The transfection was stopped 25 min later by transferring the coverslips into a fresh 12 well dish containing the conditioned media . HEK-293T cells were cultured at 37°C under a 5% CO2 atmosphere in Dulbecco’s modified Eagle’s medium ( DMEM ) supplemented with 10% fetal bovine serum ( FBS ) . They were transfected with 2–3 μg DNA at 14–16 hr post-plating using polyethylenimine ( PEI ) according to the manufacturer’s recommendation . The whole cell lysate was prepared 24 hr post-transfection using EBC buffer containing Complete-mini ( Roche ) and a phosphatase inhibitor cocktail ( Sigma ) . Wild type C57Bl6/J-Crl1 mice were purchased from Charles River Laboratories ( Germany ) . Gabra2 KO mice were bred at the Institute of Pharmacology and Toxicology , University of Zurich and genotyped as described in [26] . For retrovirus production , a non-replicative vector was adapted from the Moloney murine leukemia virus ( MMLV ) . DNA constructs with the gene of interest ( GFP-CB1SH3- , eGFP-CB2SH3- , eGFP-CB1SH3-2R were cloned into the pCAG-V-PRE-eGFP vector . The eGFP-Cdc42DN and pCMV-gp expressing gag/pol genes for virus packaging and pCMV-vsv-g used as envelope protein plasmids were a kind gift from Prof . Sebastian Jessberger ( University of Zurich ) . ShRNA against ArhGEF9 were ordered from Origene Technologies ( Rockville , USA ) in a HuSH pRFP-C-RS ( TF517554 ) backbone . GFP-panCB shRNA CTGATGAAGGACAGCCGCTATCAACACTT and 29-mer scrambled ShRNA cassette in pRFP-C-RS ( TR30015 ) was used as control . pCAG-V-PRE-eGFP-panCB shRNA was subcloned using PstI restriction sites introduced using PCR primer 5’ to U6 promoter and 3’ to termination codon . HEK-293T cells were transfected with the “CalPhos Kit” viral transfection protocol with three separate plasmids containing the capsid ( CMV-vsvg ) , viral proteins ( CMV-gag⁄pol ) and the gene of interest . Medium containing the virus was collected and the virus was purified with ultracentrifugation , re-suspended in cold phosphate-buffered saline ( PBS ) and stored at -80°C . The virus titer of minimum 106 cfu/ mL was used for sterotactic injections , after determination using serial dilution in HEK-293T cells . The eGFP-CB1SH3-K431/K432R ( 2R ) viral vector and viral vector plasmid were generated at the Viral Vector Facility ( VVF ) of the Neuroscience Center Zurich . Mice 8–12 weeks old were anesthetized by inhalation of isoflurane ( Baxter ) in oxygen , injected intraperitoneally ( i . p . ) with 1 mg/kg buprenorphine ( Temgesic , Essex Chemicals , Lucerne , Switzerland ) , and head-fixed on the stereotaxic frame ( David Kopf Instruments ) . The Nanoject II Auto-Nanoliter Injector ( Drummond Scientific Company ) was used to deliver 69 nL of MMLV bilaterally into the hilus of the dentate gyrus 15x ( antero-posterior = -2 mm; lateral = ±1 . 5 mm; dorso-ventral = -2 . 3 mm , with the bregma as reference ) , under stereotaxic guidance ( “Stereodrive software” ) . During the operation and recovery , the mice were held on a warm pad . After the operation they were received a second injection of buprenorphine . Mice were deeply anesthetized with i . p . injection of 50 mg/kg sodium pentobarbital ( Nembutal ) and perfused intracardially through the left ventricle with approximately 20–25 mL ice-cold , oxygenated Artificial Cerebrospinal Fluid ( ACSF: 125 mM NaCl , 2 . 5 mM KCl , 2 . 5 mM CaCl2 , 2 mM MgCl2 , 26 mM NaHCO3 , 1 . 25 mM NaH2PO4 , 25 mM glucose; pH 7 . 4 ) at a flow rate of 10 mL/min [37] . The brain was immediately dissected and a block containing the hippocampal formation was fixed for 3 hr in ice-cold 4% paraformaldehyde solution ( 4% PFA in 0 . 15 M sodium phosphate buffer; pH 7 . 4 ) . The tissue was rinsed with PBS and stored overnight at 4°C in solution of 30% sucrose in PBS for cryo-protection . The tissue was sectioned coronally after being frozen on tissue mounting fluid ( M-1 Embedding matrix , Shandon , Thermo Scientific , USA ) , on the frozen ( -40°C ) block of a sliding microtome ( MICROM HM 400 , MICROM International GmbH , Walldorf , Germany ) . The tissue was cut into 40 μm thick sections , which were immediately transferred in ice-cold PBS as free-floating sections . For immunoprecipitation 0 . 8 μl anti V5 antibody was added to 500 μl cell lysate and incubated over-night on a rotating wheel at 4°C and the protein complex was precipitated by using 50 μl of Protein A and Protein G Plus-Agarose beads ( Calbiochem ) in EBC buffer . The beads were washed once in EBC-based high-salt buffer ( 50 mM Tris-HCl pH 8 . 0 , 500 mM NaCl and 1% Nonidet P-40 ( Sigma-Aldrich ) and twice in EBC buffer . After boiling the samples in 2xSDS sample buffer containing 15% 2-Mercaptoethanol ( Bio-Rad ) for 10 minutes at 72°C the supernatant was loaded onto SDS-polyacrylamide gels and run at 140V at room temperature . After transferring the protein bands onto PVDF membranes with constant 35 mA in Tris-glycine transfer buffer , WB were performed by blocking the membranes with 5% western blocking reagent ( Roche Diagnostic ) in Tris-buffered saline with Tween 20 ( TBST ) for 1 hr at room temperature and later incubating with primary antibody overnight at 4°C . Secondary antibody coupled to horseradish peroxidase or IRDye ( LI-COR ) was used to visualize the protein bands . The following antibodies were used in this study: Chicken anti GFP ( 1:5000 , Aves Labs Inc . , USA ) , Mouse anti-V5 antibody ( 1:5000 , Invitrogen , Carlsbad , USA ) and ( 1:3000 , Acris , SanDiego , USA ) , mouse anti-FLAG ( 1:5000 , Sigma , Saint Louis , USA ) , mouse anti-actin ( C4 clone ) antibody ( 1:20000 , EMD Millipore Corporation , Billerica , USA ) , rabbit anti-VGAT antibody ( 1:2000 , Synaptic Systems , Göttingen , Germany ) , mouse anti-gephyrin antibody ( mAb7a , 1:1000; or 3B11 , 1:10000; Synaptic Systems , Göttingen , Germany ) , guinea pig anti-α1 subunit ( home-made; [38] , as well as a STrEP-tag Purification ( IBA GmbH , Göttingen , Germany ) . Cells were fixed for 10 minutes in 4% PFA , rinsed in PBS and permeabilized with 0 . 1% Triton X-100 containing 10% normal goat serum . Immunohistochemistry was performed by incubating the cells with the primary antibodies diluted in PBS containing 10% normal goat serum for 60 min . After washing in PBS the cells were incubated with the secondary antibodies coupled to Cy3 or Cy5 ( 1:1000 , Jackson ImmunoResearch ) for 30 min . After drying the cells were mounted with fluorescent mounting medium ( Dako Cytomation , Carpinteria , CA ) . Everything was performed at room temperature . Sections were transferred into primary antibody solution ( 0 . 2% Triton X-100 , 2% normal goat serum ( NGS ) and primary antibodies in PBS; pH 7 . 4 ) and incubated at 4°C , in darkness and under continuous agitation ( 100 rpm/min ) for 72 hr . The sections were then washed 3x10 min in PBS and incubated in the secondary antibody solution ( 2% NGS and secondary antibodies targeting the species of the respective primary antibodies used in PBS; pH 7 , 4 ) at room temperature , in the darkness and under continuous agitation ( 100 rpm/min ) , for 6 h . DAPI 1:3000 was added into the secondary antibody solution to stain cell nuclei . After incubation in the secondary antibody solution the sections were washed again 3x10 min in PBS , mounted onto gelatin-covered glass slides ( Menzel , GmbH & Co KG ) , air-dried and coverslipped with Dako fluorescence mounting medium ( Agilent Technologies , Santa Clara , CA , USA ) . RNA was isolated from WISTAR rat brain sub regions as indicated at different ages ( N = 5 ) using Sigma-Aldrich GenElute Mammalian Total RNA Miniprep Kit; cDNA was prepared using random hexamers and SuperScript II Reverse Transcriptase ( Invitrogen ) . Quantitative PCR was performed on a 7900 HT Fast Real Time PCR system ( Applied Biosystems ) . HOT FIREPol EvaGreen qPCR Mix Plus ( ROX ) ( Solis Biodyne ) was used with the designed primers to amplify cDNA ( 10ng/sample ) . Changes in mRNA levels were calculated using the ΔΔCt Method [39] relative to PanCB mRNA . Primer pairs common to all CB isoforms were designed and tested to normalize specific CB isoforms to PanCB levels . Bar charts and statistics were performed using Graphpad Prism . One-way ANOVA with Bonferroni post-hoc correction with significances: P<0 . 001 ***; P<0 . 01 **; P < 0 . 05 * CB1 fwd 5’-GTAGGGTTGGAGAGGAAGAG-3’ CB1 rev 5’-TTGTGGTGGATAGGAAGGTG-3’ CB2 fwd 5’-GCACAACAAGGAAACCGAAGA-3’ CB2 rev 5’-TGGGTTACTTTCTGTTTAGACGCTTT-3’ panCB fwd 5’- CCCTGCTTCTTGGAGCATCA -3’ panCB rev 5’- TGATAGCGGCTGTCCTTCATC -3’ HEK-293T cells were transfected with 1 μg of V5-CB and HA-ubiquitin using PEI as per vendor suggestions . Sixteen hours post-transfection , 5 μM MG132 ( Tocris ) was added and 5 hr post-treatment , cells were lysed for protein detection . V5-CB isoforms were transfected into HEK-293T cells using the PEI protocol . Sixteen hours after transfection , 100 μM cyclohexamide dissolved in culturing medium was added and the cells were lysed in EBC buffer ( 50 mM Tris- HCl pH 8 . 0 , 120 mM NaCl and 0 . 5% Nonidet P-40 ) containing complete mini-protease inhibitor ( Roche Diagnostics ) and phosphatase inhibitor cocktail 2 and 3 ( Sigma–Aldrich ) after 0 , 1 , 2 , 4 or 8 hr of treatment for 20 min at 4°C . WB quantification of protein intensity was done using LI-COR odyssey scanner and image studio . The area under the curve was normalized to actin and a half-life curve was fitted using Graphpad Prism . For analysis of morphology of HEK-293T cells and hippocampal neurons as well as gephyrin and α1 subunit clustering in neurons , a confocal laser-scanning microscope ( LSM700 or LSM710 , Carl Zeiss AG , Jena , Germany ) with a 25x or 40x oil immersion objective was used for image acquisition . The pinhole was set for all channels at 1 Airy unit; pixel size typically was 90 nm and a z-stack ( 3–4 steps at 0 . 5 μm interval ) was acquired by sequential scanning of each channel . For each condition in in vitro experiments ( primary neurons and HEK-293T cells ) , 15 cells per group from 3–4 independent experiments were imaged . For in vivo experiments , at least 3–5 mice per condition were used for analysis . In each mouse , images from 10–15 cells were acquired . Image analysis was performed using ImageJ ( http://rsb . info . nih . gov/ij/ ) . In order to measure the migration distance , CSLM images were acquired with a 25x oil immersion objective , in a way that the position of GFP-labeled granule cells could be clearly observed into the granular cell layer ( GCL ) , labeled with DAPI . Using ImageJ , the orthogonal distance from the center of the cell to the base of the GCL was measured and compared using one-way ANOVA . In order to image the neuronal cells on their whole , a 25x oil immersion objective was used to acquire z-stack images in a slice interval of 1 . 5μm . The analysis was carried out using ImageJ , version 1 . 49o; Java 1 . 6 . 0_12 ( Wayne Rusband , National Institutes of Health , USA ) . The NeuronJ plugin ( NIH ImageJ , Meijering et al . , 2004 ) was used to trace the dendritic trees and measure the length of the dendrites . In order to evaluate the complexity of the dendritic trees , Sholl analysis was performed [40] . The Sholl Analysis macro was used to count the number of intersections of the dendrites with imaginary concentric cycles , designed to have the middle of the cell soma as their center , and 10 μm subsequently increasing radii . The number of intersections for each given radius ( and hence distance from the soma ) were used to plot the diagrams , and the area under the curve ( AUC ) was used for statistical analysis . For the analysis of synaptic clusters in primary neurons a 40x oil immersion objective ( N . A . 1 . 4 ) was used and the pictures were acquired with a 1 . 8 zoom , a 0 . 45 μm inter-image interval and a 1024x pixels length in segments of proximal and distal dendrites . Images of proximal and distal dendritic segments of GFP-labeled cells were analysed with an ImageJ macro designed to identify eGFP-gephyrin clusters opposed to VGAT-positive terminals using threshold segmentation algorithms . The density of such postsynaptic clusters was normalized to a length of 20 μm . Postsynaptic clusters in eGFP-labeled adult-born GCs were analyzed individually in each image of stacks covering the entire dendritic tree , using the same macro for identification of clusters by threshold segmentation . Cumulative probability analysis was performed to analyze distribution of postsynaptic eGFP-gephyrin cluster size ( Kolmogorov-Smirnov test ) in primary neurons . The effect of CB isoforms on eGFP-gephyrin cluster density , CB isoform mRNA expression levels during brain development , CB effect on filopodia formation in HEK-293T cells and the effect of PanCB shRNA expression were analyzed by one way ANOVA with a Bonferroni post-hoc test . CB isoform effects on CB influence on dendritic arborization , the effect of Cdc42 sequestration by CB were analyzed using Mann Whitney test . All histograms , as well as the statistical analysis of the results of this study were carried out on Prism software ( GraphPad Software Inc . , La Jolla CA , USA ) .
|
GABAergic inhibition regulates distinct stages of brain development; however , cellular mechanisms downstream of GABAA receptors ( GABAARs ) that influence neuronal migration , maturation and synaptogenesis are less clear . ArfGEF9 encodes for RhoGEF with Cdc42 and TC10 GTPase as its substrates . Interestingly , ArhGEF9 is the only known RhoGEF essential for GABAergic synapse formation and maintenance . We report that during brain development ArfGEF9 mRNA splicing regulation generates different ratios of CB1 and CB2 splice isoforms . CB1 mRNA splicing is enhanced during early brain developmental , while CB2 levels remains constant throughout brain development . We also show that CB1 protein has shorter half-life and ubiquitin proteasome system restricts CB1 abundance within developing neuron to modulate neuron migration and distributing GABAergic synapse along the proximal-distal axis . On the other hand , CB2 isoform although expressed abundantly throughout brain development is essential for neuron dendrite maturation . Together , our data identifies specific post-transcriptional and post-translational mechanisms downstream of GABAARs influencing ArhGEF9 function to regulate distinct aspects of neuronal maturation process .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Conclusions",
"Materials",
"and",
"methods"
] |
[
"cell",
"motility",
"medicine",
"and",
"health",
"sciences",
"neurogenesis",
"nervous",
"system",
"electrophysiology",
"neuroscience",
"developmental",
"biology",
"cellular",
"structures",
"and",
"organelles",
"neuronal",
"dendrites",
"cytoskeleton",
"contractile",
"proteins",
"actins",
"developmental",
"neuroscience",
"animal",
"cells",
"proteins",
"neuron",
"migration",
"biochemistry",
"cytoskeletal",
"proteins",
"adult",
"neurogenesis",
"cellular",
"neuroscience",
"neuronal",
"morphology",
"cell",
"biology",
"anatomy",
"synapses",
"physiology",
"neurons",
"cell",
"migration",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"neurophysiology"
] |
2017
|
RhoGEF9 splice isoforms influence neuronal maturation and synapse formation downstream of α2 GABAA receptors
|
Plasmodium yoelii YM asexual blood stage parasites express multiple members of the py235 gene family , part of the super-family of genes including those coding for Plasmodium vivax reticulocyte binding proteins and Plasmodium falciparum RH proteins . We previously identified a Py235 erythrocyte binding protein ( Py235EBP-1 , encoded by the PY01365 gene ) that is recognized by protective mAb 25 . 77 . Proteins recognized by a second protective mAb 25 . 37 have been identified by mass spectrometry and are encoded by two genes , PY01185 and PY05995/PY03534 . We deleted the PY01365 gene and examined the phenotype . The expression of the members of the py235 family in both the WT and gene deletion parasites was measured by quantitative RT-PCR and RNA-Seq . py235ebp-1 expression was undetectable in the knockout parasite , but transcription of other members of the family was essentially unaffected . The knockout parasites continued to react with mAb 25 . 77; and the 25 . 77-binding proteins in these parasites were the PY01185 and PY05995/PY03534 products . The PY01185 product was also identified as erythrocyte binding . There was no clear change in erythrocyte invasion profile suggesting that the PY01185 gene product ( designated PY235EBP-2 ) is able to fulfill the role of EBP-1 by serving as an invasion ligand although the molecular details of its interaction with erythrocytes have not been examined . The PY01365 , PY01185 , and PY05995/PY03534 genes are part of a distinct subset of the py235 family . In P . falciparum , the RH protein genes are under epigenetic control and expression correlates with binding to distinct erythrocyte receptors and specific invasion pathways , whereas in P . yoelii YM all the genes are expressed and deletion of one does not result in upregulation of another . We propose that simultaneous expression of multiple Py235 ligands enables invasion of a wide range of host erythrocytes even in the presence of antibodies to one or more of the proteins and that this functional redundancy at the protein level gives the parasite phenotypic plasticity in the absence of differences in gene expression .
Despite the recent renewed onslaught to tackle a disease that infects 300-660 million people and kills one million each year worldwide [1] , the malaria parasite remains an elusive target . During the asexual blood stage , which is responsible for the disease , the parasite invades and develops within erythrocytes , but the precise molecular mechanisms employed to gain entry into the erythrocyte are still being worked out . A number of parasite adhesion proteins have been identified as important in the selection and invasion of host cells and have been grouped according to structural and sequence homology rather than host molecular specificity or cellular phenotype ( reviewed in [2] , [3] , [4] , [5] ) . The role of the actin-myosin motor complex in the invasion of erythrocytes is also being elucidated [6] , [7] . Together , merozoite surface proteins , the adhesion ligands and the motor complex add up to a multifaceted molecular interaction that results in the successful selection and invasion of host cells [3] , [4] , [5] . Understanding the role played in the invasion cascade by adhesion proteins with homologues in both human and rodent Plasmodium is of paramount importance in the quest to design intervention tools that will inhibit invasion pathways and so kill the parasite and prevent disease . Of the Plasmodium adhesion ligand families identified to date , one of the most studied is the erythrocyte binding ligand family ( EBL ) , which includes P . falciparum erythrocyte binding antigen ( EBA ) -175 and the Duffy binding protein ( DBP ) of P . vivax and P . knowlesi ( reviewed in [3] , [4] ) located in the apical organelles of the merozoite . A second group of high molecular mass adhesion proteins , which was first described in the rodent malaria parasite Plasmodium yoelii as Py235 [8] , [9] , is the reticulocyte binding-like ( RBL ) super family , so named because of sequence homology with the reticulocyte binding protein ( RBP ) -1 and RBP-2 , of Plasmodium vivax . In P . vivax , these proteins are thought to be involved in erythrocyte selection as they bind to reticulocytes but not mature erythrocytes thereby restricting P . vivax to the invasion of reticulocytes [10] . P . falciparum contains a small group of genes coding for proteins with similarities to Py235 and PvRBP , the PfRH family [11] , [12] , [13] . In contrast to the PvRBP and PfRH gene families , which are small , the Py235 multigene family contains at least 11 members [9] , [14] , [15] , [16] , [17] , [18] . Analysis of the sequences on fifteen contigs identified in the P . yoelii genome database [15] , which represent members of the Py235 gene family ( and some of which are incomplete ) , show they have overall conserved structural elements [16] , [19] , [20] . The Py235 proteins have been implicated in the selection , recognition and invasion of erythrocytes . For example , passive immunization of mice with monoclonal antibodies ( mAbs ) 25 . 77 and 25 . 37 specific for Py235 , or immunization of mice with mAb 25 . 77-affinity purified protein restricts the growth of the virulent YM line of P . yoelii [8] , [21] . In these experiments the invasion profile was switched from invasion of erythrocytes of all ages to invasion of only reticulocytes , suggesting that the antibodies prevent parasite recognition and invasion of mature erythrocytes . This restriction in cell specificity resulted in a non-lethal infection similar to that of the avirulent 17X line , in contrast to the normal lethal phenotype of the YM parasite . Of note is the finding that a combination of both protective mAbs together conferred greater protection [22] , suggesting that the epitope recognized by each of the mAbs is not identical and may or may not be on distinct members of the family . Py235 proteins are released in soluble form from parasitized cells maintained in vitro , and two of these are recognized by mAb 25 . 77 [23] , [24] , [25] . However , only one of these forms was detected binding specifically and preferentially to the surface of mature mouse erythrocytes [24] . Binding was to neuraminidase-resistant , chymotrypsin- and trypsin-sensitive erythrocyte receptors , and the binding was abolished by incubation with Py235-specific antibodies [25] . Several invasion pathways coexist in a single parasite as exemplified by P . vivax that requires selection of reticulocytes ( using the RBPs [26] ) that are Duffy blood group antigen positive , ( using the DBP [27] , [28] ) to successfully gain entry into the host cell . In P . falciparum the Dd2 clone can switch from being dependent on sialic acid for entry into the erythrocyte , allowing it to invade neuraminidase-treated erythrocytes [29] . This change of phenotype has been found to be due to the up-regulation of PfRH4 [30] , [31] . Polymorphism due to amino acid substitutions in the binding domain of PfRH5 leads to recognition of different erythrocyte surface receptors , [32] . That several pathways are available to a single parasite is further demonstrated by the observation that invasion into an enzyme-treated cell is not all or nothing even though enzymatic treatment goes to completion , ( in P . falciparum [11] , [33] , and in P . yoelii [25] ) . Therefore , the invasion pathway of a parasite depends not only on the set of ligands expressed or silenced , some of which are coded by genes under epigenetic control [34] , but also on a molecular hierarchy that determines which of the expressed ligands are used [2] , [3] . This variant expression of adhesion-/invasion-related proteins is thought to be primarily driven by immune evasion although it may also help to increase the range of erythrocytes that can be invaded [35] , [36] . Populations of P . yoelii asexual blood stage parasites express multiple members of the py235 gene family [17] , [36] , [37] . Multiple gene products were detected in individual schizonts although only single products were identified in single merozoites , leading to the suggestion that the presence of the family allowed clonal phenotypic variation [38] . On the other hand , all merozoites within schizonts reacted with mAb 25 . 77 [17] , suggesting that they either share the protein recognized by this antibody or the epitope is present on multiple members of the family . We have previously identified a specific erythrocyte binding member of the Py235 family ( Py235EBP-1 ) , which is recognized by mAb 25 . 77 , and its corresponding gene ( py235ebp-1[PY01365] ) [20] . Here , we describe the effect on parasite growth in vivo of deleting the gene that encodes the Py235EBP-1 expressed in asexual blood stages of the virulent P . yoelii YM line in order to better understand the role of the Py235 protein family in erythrocyte recognition , binding , and merozoite invasion . We also examine the expression of other family members in this py235ebp-1 knock out ( KO ) parasite line . Furthermore , we identify the proteins recognised by the other protective mAb 25 . 37 to obtain an understanding of the relationship in the invasion process between the two protective mAbs and the Py235 proteins they recognize . Whilst there is no difference in the level of expression of other genes in the family , other proteins compensate for the loss of the erythrocyte binding protein , highlighting the importance of functional redundancy to provide plasticity in interaction with the host .
Disruption of the py235ebp-1 ( PY01365 ) by insertion of the DHFR cassette by double homologous recombination ( Figure 1A ) was carried out . Southern blot analysis of digested gDNA from transfected parasites selected with pyrimethamine identified a single band of the expected size in this population , when the filter was probed with a fragment of DHFR/TS ( Figure 1B , lane 2 ) . In contrast , hybridization with the probe that binds to the 3′ coding region of py235ebp-1 ( Fragment B ) , detected DNA in both the wild type ( WT ) ( 2 . 3Kb ) and the KO ( 3 . 8Kb ) parasite lines as expected ( Figure 1B , lanes 3 and 4 ) . Four individual clones ( 1 to 3 shown ) derived from the population of parasites gave a similar result with both fragment B ( Figure 1B , lanes 5 to 7 ) and the DHFR/TS probe ( Figure 1B , lanes 8 to 10 ) , clearly showing that the PY01365 gene had been disrupted . Further evidence that the PY01365 gene had been deleted from the P . yoelii genome was obtained by chromosome analysis ( Figure 1C ) . Hybridization with the probe that binds to a 5′ coding sequence of PY01365 ( Fragment C ) , detected a signal only in the WT parasite lanes , ( Figure 1C , lanes 3 and 4 ) . Hybridization with a probe that binds the 3′ UTR of DHFR/TS detected both the modified py235ebp-1 locus ( chromosome 13/14 ) in the KO parasite line ( Figure 1C , lanes 5 and 6 ) and the endogenous dhfr locus ( chromosome 7 ) in both KO and WT parasites ( Figure 1C , lanes 5 to 8 ) . Hybridization of a chromosome blot with the Fragment B probe identified a band in all the lanes as expected ( data not shown ) . Southern blot analysis of ten sets of double restriction enzyme digested gDNA from WT P . yoelii YM-parasitized erythrocytes probed with either Fragment B or C gave single bands under low stringency washes ( Figure 2 ) . These data suggest that PY01365 is a single copy gene in the line of P . yoelii YM parasites used in this study . This conclusion is also supported by the absence of any RNA-Seq reads mapping to any part of PY01365 from the PY01365-KO parasite . The mAb 25 . 77 had previously been used to identify Py235EBP-1 , the product of the PY01365 gene . By IFA , this mAb gives a punctuate pattern of fluorescence in the WT parasite line ( Figure 3A ) . Surprisingly , a similar pattern was also observed for the PY01365-KO parasite line , even though Py235ebp-1 is no longer being expressed . The pattern of reactivity ( Figure 3B ) was similar but not identical to that of antibodies specific for the micronemal protein , Apical Membrane Antigen 1 ( AMA1 ) [39] , the erythrocyte binding ligand protein ( EBL ) , which has a dense granule location in this parasite line [40] , and rhoptry neck protein 4 ( RON 4 ) [41] . Furthermore , when proteins released into in vitro culture supernatant from radiolabeled WT and PY01365-KO parasitized erythrocytes were immunoprecipitated using mAbs 25 . 77 and 25 . 37 ( Figure 3C ) , or bound to erythrocytes , eluted and then immunoprecipitated ( Figure 3D ) , both mAbs recognized proteins of approximately 235 kDa showing that Py235 proteins were expressed by both WT and PY01365-KO parasite lines . Clearly the Py235 proteins now expressed by the KO parasite line , although at least in part different to those being expressed by the WT parasite , share common epitopes bound by the antibodies . Further confirmation that merozoites express epitopes recognized by both mAb 25 . 77 and 25 . 37 was obtained using WT parasitized erythrocytes in a dual labeling fluorescent assay . Alexa Fluor 488-conjugated mAb 25 . 77 and Alexa Fluor 594-conjugated mAb 25 . 37 were used to probe the same thin blood smears of mixed stage WT P . yoelii YM parasites ( Figure 3E ) . Overlay of the individual images showed clearly that both antibodies recognized the same parasites . The gene encoding the protein recognized by mAb 25 . 77 and expressed in WT parasites has previously been identified . A second protective mAb 25 . 37 also recognizes Py235 proteins and we wished to identify the protein ( s ) to which it binds . To identify the corresponding genes , peptide mass fingerprinting was carried out on the Py235 proteins affinity purified using mAb 25 . 37 from both schizonts ( Figure 4A ) and from culture supernatant of WT parasites maintained in vitro ( Figure 4B ) and fractionated by SDS-PAGE on a 5% gel . The peptides detected were derived from three contigs in the genome database; one contained a full length ( 8172bp ) py235 gene sequence , PY01185 , and the remaining two contained partial gene sequences ( Table 1 ) . Contig PY05995 is 2685bp in length and contains sequence that aligned with the 5′ end of other Py235 gene family members , while PY03534 ( 5478bp ) aligned with the 3′ of other Py235 genes . To establish whether or not these two contigs are part of the same gene , primers designed to the 3′-sequence of PY05995 and to the 5′-sequence of PY03534 were used to amplify sequence from gDNA , and gave a single PCR product of the expected size , 392bp ( Figure 5A ) . Sequence analysis and alignment with the gene sequences from the database showed perfect alignment ( Figure 5B ) and confirmed that the contigs were part of the same single full length Py235 gene , PY05995/PY03534 . This conclusion was further confirmed by read pairs of the RNA-Seq data . 153 mates mapped to the end of PY03534 and the beginning of PY05995 , and the entire gene could be assembled from the mapping reads ( Figure 5C ) . Sequence data for the Py235 family was aligned and examined for structural features . Partial or full length sequences were compiled from the literature [15] , [16] , [17] , [19] , [42] , [43] resulting in eleven essentially full length protein sequences , one almost full length but lacking the C-terminus and two others , one representing the N- and the other the C-terminal sequence of one or two further genes . The PY01365 , PY01185 , and PY05995/PY03534 sequences form a discrete subset of the family with a degree of similarity in pairwise alignment of greater than 80% at the amino acid sequence level ( Table 2 ) . None has the Asp-Ile-Asn ( DIN ) repeats close to the C-terminus of the protein found in some members of the family . A further subgroup contains genes 11 , 10 , PY03184_E3 , PY02104_E5 , and PY04438_PY0618_E8 . We were interested to identify the Py235 proteins expressed by the PY01365-KO parasite line that were recognised by mAb 25 . 77 in the absence of Py235EBP-1 . Py235 proteins affinity purified using mAb 25 . 77 from both detergent-solubilised parasite preparations ( Figure 6A ) and culture supernatant ( Figure 6B ) from the PY01365-KO parasite line were fractionated on a 5% SDS-PAGE gel and processed for mass spectrometry analysis . MASCOT searches using the peptides and the NCBI database gave significant matches to three Py235 contigs , PY01185 , PY05995 and PY03534 ( Table 2 ) , mirroring the results obtained with mAb 25 . 37 and WT parasites . The results suggest that mAb 25 . 77 has a higher affinity for PY01365 , than PY01185 and PY05995/PY03534 protein products , since in the absence of PY01365 mAb 25 . 77 was able now to detect the other proteins . While mass spectrometry analysis of proteins affinity purified from WT parasites using mAb 25 . 77 routinely clearly identified PY01365 ( Py235EBP-1 ) ( Table 2 ) , peptides derived from PY01185 , PY05995 and PY03534 were also present in small amounts ( data not shown ) . Due to the high number of unique peptides required for positive identification of these large proteins , the few unique peptides obtained for PY01185 , PY05995 and PY03534 was insufficient . Radiolabeled proteins from WT and PY01365-KO parasites that had been released into the supernatant of in vitro cultures were used in erythrocyte binding assays . Proteins bound to and eluted from the erythrocyte surface were immunoprecipitated using mAbs 25 . 77 and 25 . 37 . These mAbs recognized single protein bands of approximately 235 kDa in this fraction ( Figure 3C and D ) . We have shown previously that of the several biosynthetically-labeled Py235 proteins released into the supernatant of parasites incubated in vitro , only one binds to the surface of erythrocytes and is recognized by mAb 25 . 77 ( Py235EBP-1 ) . Similarly , we now show that mAb 25 . 37 also recognizes two Py235 proteins released into culture supernatants ( Figure 3C ) and that only one of them , the upper of the two bands ( Figure 3D ) , binds to erythrocytes . This upper band has been identified as the protein encoded by the gene , PY01185 ( Figure 4A and B ) , identifying another Py235 protein that binds to the surface of erythrocytes and is recognized by the protective mAb 25 . 37 . This protein has been designated Py235 erythrocyte binding protein-2 ( Py235EBP-2 ) , as the second known erythrocyte binding protein from this family and encoded by the PY01185 gene . Our result suggests that in WT parasites there are at least two erythrocyte binding proteins , Py235EBP-1 , encoded by PY01365 and Py235EBP-2 encoded by PY01185 . In the PY01365-KO parasite line , the erythrocyte binding protein is Py235EBP-2 . By western blotting , similar amounts of Py235 protein were detected by mAbs 25 . 37 and 25 . 77 in extracts of both WT and PY01365-KO parasites ( Figure S1 ) indicating that there has been no compensatory change in protein levels such as upregulation of Py235EBP-2 . To test the hypothesis that there had been a switch , for example the up-regulation of other members of the Py235 family expressed in the PY01365-KO parasite line , two methods were used: quantitative RT-PCR ( qPCR ) for some specific members and RNA-Seq for all known members of the family . qPCR was carried out using primers specific to the genes of interest and to reference genes coding for PyEBL ( PY04764 ) , which is expressed at the same developmental stage as Py235 proteins , and the gene for the constitutively expressed protein Pyβ-tubulin ( PY05711 ) ( Table 2 ) . Of the 3 genes in the Py235 family that were examined , PY01365 had the lowest transcription level followed by PY05995/PY03534 , with PY01185 having the highest transcription level in the WT parasite line ( Figure 7a ) . Two reference genes , PyEBL and Pyβ-tubulin had similar amounts relative to each other in both parasite lines . The quantification cycle ( Cq ) value obtained for PY01365 was similar to those for the negative controls ( -RT or no template control ) , confirming that PY01365 had been deleted from the genome of P . yoelii YM in the PY01365-KO line . Fold change transcriptional calculations between the KO and WT lines were made . There was a fold change increase of 1 . 5 in the transcription level of PY01185 in the KO line , and for PY05995/PY03534 the fold change increase was 1 . 9 . The measured fold change decrease of PY01365 in the KO line was 350 , showing its absence in the PY01365-KO line . In the RNA-Seq data , the Pearson correlation of all expressed genes between both parasite lines is nearly 0 . 99 ( 0 . 9899136 ) , taking the log of the geometric mean [44] ( Figure S2 ) . Also the ratio of expression between the two reference genes of the qPCR ( PY04764 and PY05711 ) was between 0 . 9 and 1 . 1 ( Table 3 ) . The data show that within the family , in the WT parasite , two genes are very poorly expressed ( PY02104 and PY06381 ) . Preliminary analysis of P . yoelii YM genomic DNA suggests that PY06381 is absent from this genome ( data not shown ) . Of the remaining genes all are expressed ( geometric means 159 . 56 to 725 . 4 ) with the PY05054 transcript being most abundant . Comparing the WT and PY01365-KO lines , as expected PY01365 is not expressed in the PY01365-KO parasite line but is clearly present and expressed in the WT , confirming the deletion of the PY01365 gene ( Figure 7b ) . For the other members of the family the average ratio of expression in KO versus WT lines was 1 . 022 and for PY05995 and PY01185 it was 1 . 406 and 0 . 999 , respectively . This indicates that there was no compensatory significant upregulation of expression of any of the other Py235 genes in the KO parasite . To evaluate the phenotypic effect of deleting py235ebp-1 from the genome of the virulent P . yoelii YM line , for example on the age of the host cell invaded or on the course of infection , groups of 5 mice were injected with parasitized erythrocytes . P . yoelii parasites of the KO line showed no changed preference for a particular host cell type relative to WT parasites . All host cell types , both mature erythrocytes and reticulocytes , were invaded , indicating that deletion of the PY01365 gene ( py235ebp-1 ) did not restrict the parasites to invasion of reticulocytes . In mice made reticulocytemic there was no difference in cell preference or growth rate between the two parasite lines ( data not shown ) . Parasite growth kinetics for all groups of mice were very similar and there was no clear difference in the parasite multiplication rate ( Figure 8 ) . Only in the group injected with a thousand parasitized erythrocytes was there a significant reduction in parasite growth , when comparing the KO and WT parasite lines ( P<0 . 05 ) . Disruption of the PY01365 gene was not lethal to the parasite and no phenotype was detectable with respect to the age of host cell invaded .
We were interested to examine the effect on growth in vivo of deleting the gene that encodes the Py235EBP-1 protein expressed in asexual blood stages of the virulent P . yoelii YM line . Additionally , we wished to examine the expression of other family members in this py235ebp-1-KO parasite line and identify the proteins recognised by the other protective mAb 25 . 37 in WT parasites . We have previously shown that although there are several biosynthetically-labeled soluble Py235 proteins released into the supernatant of parasites incubated in vitro , only one of these proteins binds to the surface of erythrocytes and is recognized by the protective mAb 25 . 77 [24] . We show here that the second protective mAb 25 . 37 also recognizes two Py235 proteins in the in vitro culture supernatant , namely the products of PY01185 and PY05995/PY03534 . We have obtained peptide mass and sequence information from the Py235 proteins either purified from parasitized erythrocytes or from the in vitro culture supernatant , which identifies the corresponding genes as members of the Py235 family . As in the case of the proteins recognized by mAb 25 . 77 , only one of the two Py235 proteins , the upper of the two protein bands , PY01185 , binds to erythrocytes . Therefore PY01185 is the gene identified as coding for the erythrocyte binding protein recognized by 25 . 37 , which has been designated Py235 erythrocyte binding protein-2 ( Py235EBP-2 ) . We sought to identify the proteins being expressed by the py235ebp1-KO parasite line that could still be recognized by mAb 25 . 77 . Interestingly , proteins affinity purified from both detergent solubilized parasites and culture supernatant using mAb 25 . 77 were shown to be Py235EBP-2 ( PY01185 ) and PY05995/PY03534 - the same gene products recognized by mAb 25 . 37 in WT parasites . Our results clearly show that in WT parasites there are two erythrocyte binding proteins , namely , Py235EBP-1 , recognized by the protective mAb 25 . 77 and encoded by PY01365 and Py235EBP-2 recognized by the protective mAb 25 . 37 and encoded by PY01185 . In the PY01365-KO parasite line , the erythrocyte binding protein recognized by mAb 25 . 77 is Py235EBP-2 . Immunofluorescence studies indicate that each merozoite within a schizont expresses proteins recognized by both protective mAb 25 . 77 and 25 . 37 and that proteins recognized by the protective mAbs are expressed by all merozoites; this confirms and extends the conclusions of Narum et al [17] . The location of the proteins still needs to be resolved: in the immunofluorescence studies there was only partial overlap of mAb 25 . 77 staining and that of other antibodies specific for microneme , dense granule and rhoptry neck proteins . It has been reported that some P . yoelii lines contain two copies of the PY01365 gene [45] . For the YM line we have analysed , the data indicate that only one copy of the gene is present . This conclusion is based on the Southern blot analysis of separated chromosomes , and of digested gDNA , and is supported by the qPCR and RNA-Seq analyses . However , other members of the gene family with a significant homology to PY01365 may be detected with certain probes at low stringency . This is in agreement with work carried out by Iyer et al [36] . The absence of the PY06381 gene in the YM genome is consistent with an extra gene being detected on chromosome blots of 17X parasites [18] . We examined gene transcription in both the WT and PY01365 parasites by qPCR and as expected , there was no difference in the mRNA levels of PyEBP and Pyβ-tubulin between the WT and KO parasite lines . Of the 3 genes in the py235 family whose transcriptional level was examined , PY01365 had the lowest transcription level followed by PY05995/PY03534 and then PY01185 . This result is in contrast to that of Iyer et al [36] who reported that PY01365 was the most highly expressed Py235 family member . This difference may be due to the use of different P . yoelii YM parasite lines with different gene copy numbers . There was a small difference in the level of transcription of PY01185 and PY05995/PY03534 between the WT and KO parasite lines . A more detailed analysis of the Py235 family using RNA-Seq indicated that most of the Py235 gene family is transcribed in these asexual blood stage parasites and there is no change following the deletion of PY01365 other than in the absence of products from this gene . Thus the redundancy in the function of this family must occur at the protein level rather than being reliant on genes being up-regulated; PY01185 protein probably takes over the function of PY01365 , although the level and role of other Py235 proteins cannot be addressed . This result is in contrast to the picture in P . falciparum where individual RH protein genes are under epigenetic control and change in expression can lead to change in receptor recognition and host cell invasion pathway . It is possible that the expression of most of the Py235 genes at the same time could contribute to the noted virulence of the YM parasite [46] . Targeted disruption of py235ebp-1 ( PY01365 ) did not lead to a change in the invasion phenotype . Although there was a significant difference in the course of infection in groups of mice injected with 1000 parasitized erythrocytes ( P<0 . 05 ) , there was no significant difference in the groups of mice that received either 200 or 5000 WT and KO parasites . It will be of interest to delete both PY01365 and PY01185 , since this double deletion might be expected to result in a much more severe phenotype . The most puzzling result was that we were unable to detect at a significant level by affinity purification the PY01185 and PY05995/PY03534 proteins in extracts of WT parasites using mAb 25 . 77 even though the transcripts were present in the parasite at similar or higher levels than that of PY01365 . In contrast these proteins were clearly detectable with mAb 25 . 77 in extracts from the KO parasite line using exactly the same methodology and could be purified from the extract of WT parasites using the second mAb , 25 . 37 . One limitation of MALDI-TOF fingerprinting is that large proteins require a relatively large number of matched peptides to generate a significant MASCOT score . The few peptides unique to PY05995/PY03534 identified on detailed analysis of the peptide mass finger print data were insufficient to establish the presence of PY01185 and PY05995/PY03534 in the proteins extracted from WT parasite using mAb 25 . 77 and so these proteins were below the level of detection by mass spectrometry . Even if the level of transcription as judged by qPCR and RNA-Seq did not correlate with the level of protein expression , this does not explain the discordant results obtained with the two antibodies . It is possible that the two mAbs may recognize common binding domains in the Py235 proteins but with different affinities , dependent upon the precise amino acid sequence of the antigens . It is conceivable that in the absence of Py235EBP-1 ( which may have a higher affinity for mAb 25 . 77 ) other members of the family , such as Py235EBP-2 , and PY05995/PY03534 , can now bind to mAb 25 . 77 . In the absence of the gene coding for the Py235EBP-1 , other family member proteins carry out the same function; this redundancy facilitates binding and erythrocyte invasion . We show that the removal of the erythrocyte ligand expressed by PY01365 , in the py235ebp1-KO line allows other expressed members of the py235 gene family to be used . In this instance , there was no change in phenotype with respect to the age of host cell invaded and mediated by the new set of parasite ligands . The invasion phenotype/pathway of a parasite depends not only on the set of ligands expressed or silenced , but also on a molecular/functional hierarchy that determines which of the expressed ligands are used [3] , reviewed by Cortes [2] . Several pathways probably coexist in a single parasite so that invasion into different cells such as those in different mammalian hosts or experimentally generated , such as enzyme-treated cells is not all or nothing even though enzymatic treatment goes to completion , ( in P . falciparum , [11] , [33] and in P . yoelii [25] ) . Our data would also fit in with the ‘limited space hypothesis’ proposed for the P . falciparum ( PfRH family ) whereby the position of a particular PfRH ligand at the apex of the merozoite determines which ligand is used for invasion [47] , [48] . In this current study , perhaps the absence of Py235EBP-1 allows space for the binding of another member of the Py235 family member , such as Py235EBP-2 ( PY01185 ) to initiate erythrocyte invasion . Different levels of protein expression in Plasmodium that are not matched by the level of transcription as seen by qPCR or RNA-Seq may arise through post-transcriptional controls of these proteins [34] . The sub-telomeric location of invasion-associated multigene families , including the Py235 family [18] , may be important for variant expression , alternatively , new forms of protein regulation at the level of translation [49] may occur . However , none of these mechanisms appears to contribute to our findings because the level of proteins seems to be unchanged . Analysis of the protein sequences we have identified showed that they are coded by a subset of the Py235 family . For example , they all lack the short repetitive sequence based on the tripeptide , Asp-Ile-Asn ( DIN ) , which is located just N-terminal to the transmembrane domain . The significance of this is obscure but the observation does cast doubt on the validity of using this repeat sequence as a diagnostic marker for the expression of all genes in the Py235 family [37] , [38] , [50] . In this study we have shown by targeted disruption that the py235ebp-1 ( PY01365 ) gene is not essential to the parasite and the KO did not result in a change in the invasion phenotype with respect to the age of mouse cell invaded or the parasite growth rate . However , deletion of py235ebp-1 did seem to result in an alteration in the level or accessibility of other Py235 protein family members such that they became able to bind to mAb 25 . 77; the proteins coded by PY01185 and PY05995/PY03534 appeared to compensate for the absence of py235ebp-1 ( PY01365 ) . The basis for this new accessibility is obscure , but it is possible that these proteins form complexes either with each other or other proteins , which could make the antibody binding site cryptic . Our result suggests that in WT parasites there are at least two Py235 erythrocyte binding proteins , Py235EBP-1 , ( recognized by mAb 25 . 77 and encoded by PY01365 ) and Py235EBP-2 ( recognized by mAb 25 . 37 and encoded by PY01185 ) . In the PY01365-KO parasite line , the erythrocyte binding protein is changed to Py235EBP-2 recognized by both mAb 25 . 77 and 25 . 37 in the absence of Py235EBP-1 . In conclusion , in the absence of Py235EBP-1 , invasion of erythrocytes by P . yoelii takes place using Py235EBP-2 , an alternative Py235 erythrocyte binding protein; modulation of erythrocyte binding appears to occur at the level of the proteins without significant changes in gene expression .
All animal work protocols were reviewed and approved by the Ethical Review Panel of the MRC National Institute for Medical Research and approved and licensed by the UK Home Office as governed by law under the Animals ( Scientific Procedures ) Act 1986 ( Project license 80/1832 , Malaria parasite-host interactions ) . Animals were handled in strict accordance with the “Code of Practice Part 1 for the housing and care of animals ( 21/03/05 ) ” available at http://www . homeoffice . gov . uk/science-research/animal-research/ . The numbers of animals used was the minimum consistent with obtaining scientifically valid data . The experimental procedures were designed to minimize the extent and duration of any harm and included predefined clinical and parasitological endpoints to avoid unnecessary suffering . Female BALB/c mice , with an average weight of 18 to 22 g and 6 to 8 weeks old were obtained from the specific pathogen-free unit at the MRC National Institute for Medical Research . The cloned virulent YM line of P . yoelii [46] , [51] , was obtained from Dr . David Walliker , University of Edinburgh . Parasites were passaged no more than five times in the same mouse strain , before returning to a fresh stabilate . All full-length and partial py235 gene sequences identified in the database ( www . PlasmoDb . org ) were retrieved and Clustal X 1 . 81 [52] was used to align them . Areas of sequence similarity and difference were identified at both the amino acid and nucleotide levels and analysed using Bioedit [53] . The sequence information was then used to design gene-specific reagents , and compare features of the individual sequences . PY01365 ( py235ebp-1 ) gene sequences were amplified from P . yoelii YM line genomic DNA using gene specific primers . A 500bp fragment from the 5′UTR of PY01365 ( Fragment A ) was amplified with forward primer , ( restriction sites are underlined ) , 5′-gccgggggcccACTATAACACTAATTATTTATTATAAAACG-3′ and reverse primer , 5′-gccggaagcttATGTATGTATCTATGTATGCATGCATG-3′ . A region from the 3′ coding sequence of PY01365 ( Fragment B ) was amplified with forward primer , 5′-gccgggaattcACGAACTCACTCGAATACAAAGTCGTTTAG-3′ and reverse primer , 5′-ggcggtctagaATAATTTTTATATTTTGCATCATCATTATTATTATGG-3′ . Fragment A PCR product was digested with ApaI and Hind III and Fragment B PCR product was digested with EcoRI and XbaI . The targeting construct was made by the cloning of fragments A and B sequentially into the plasmid vector pBSDHFR , in which the Toxoplasma gondii dihydrofolate reductase/thymidylate synthase gene ( DHFR/TS ) is flanked by the upstream and downstream control elements from P . berghei DHFR/TS . First , Fragment A was cloned into pBSDHFR that had been digested with Apa1 and Hind III and the inserted DNA sequence was verified by sequencing . This construct was digested with EcoRI and XbaI , and then Fragment B cloned into it , and its sequence verified . The final targeting construct was digested with the enzymes ApaI and XbaI and inserted by double homologous recombination into the PY01365 gene following transfection of the virulent YM line of P . yoelii . Transfection of parasites was carried out essentially as described previously [54] , [55] . Briefly , erythrocytes containing late stage parasites , were harvested at 20 to 25% parasitaemia and schizonts were purified by centrifugation for 25 min at 600 g at room temperature ( RT ) on a 55% Nycodenz ( Nycoprep ) cushion ( NYCOMED Pharma AS ) . 5×107 schizonts were mixed with 90 µl AMAXA nucleofactor T-cell solution ( plus supplements ) and 5 µg of targeting construct DNA was added . These parasites were transfected using AMAXA Nucleofector programme U33 . Immediately , 100 µl of RPMI 1640 medium containing 20% foetal calf serum ( FCS ) was added to the transfected parasites and the suspension injected intravenously ( i . v . ) into the tail vein of a single mouse . Electroporation of parasites with targeting construct and injection into individual mice was carried out twice independently using the above conditions . Twenty seven hours post injection , day ( D ) 1 , and on D3 and D4 mice were treated with 1 mg kg −1 pyrimethamine , intraperitoneally ( i . p . ) . From day 2 post injection , pyrimethamine was administered continuously in the drinking water at a final concentration of 70 µg/ml . Three sets of control mice were set up . One set was injected with transfected schizonts as above but did not receive any drug treatment , a second set was injected with schizonts in T-cell solution ( without DNA or electroporation ) and was drug treated as above , and the third set of controls was as the second set but without subsequent drug treatment . The parasitaemia of each set of mice was monitored daily . Stabilates of drug resistant parasites were made and stored in liquid nitrogen , additionally the parasites were used to infect naïve mice ( 2×107–5×107 parasitized erythrocytes , administered i . v . ) for a second round of drug pressure . These mice were given pyrimethamine continuously in their drinking water as above . Parasites were allowed to grow sufficiently for samples to be taken for analysis and for stabilates to be made . After verification of the pyrimethamine-resistant parasites by PCR and Southern blot analysis , the transgenic parasite line was cloned by limiting dilution using 10 mice injected i . v so that each inoculum contained a maximum of one parasite . Four clones were obtained and genotyped . Genomic DNA ( gDNA ) was isolated from leukocyte-depleted , Percoll-purified late stage WT and PY01365-KO parasitized erythrocytes lysed in a buffer containing SDS . The DNA was phenol chloroform extracted and precipitated with ethanol . Various restriction enzymes were used to digest the gDNA and samples were resolved by electrophoresis on a 0 . 8% agarose gel and transferred onto Hybond N+ in 7 . 5 mM NaOH overnight . The filter was neutralized in 2 x SSC and UV cross-linked prior to hybridisation [13] . DNA probes used were: Fragment B ( see above ) , and a 739 bp TgDHFR DNA sequence PCR amplified from plasmid DNA using the forward primer: 5′-GCCGGGATCCCATCATTCGACCCTGATATATATAACGA-3′ and the reverse primer: 5′-GCCGGGAATTCATTCTAAAAATTCATAGTAATAAGGTG-3′ . For an estimation of the number of copies of PY01365 in the P . yoelii genome , WT gDNA was digested with various restriction enzymes in double digest reactions and the samples transferred onto nylon filters as above . DNA probes used were Fragment B , ( as above ) and Fragment C derived from the 5′ coding sequence of PY01365 , amplified using the forward primer: 5′- ATCATCTGCACCATCATTCGAC-3′ and the reverse primer: 5′- CAATATGGAATCTAATAGACG-3′ . DNA probes were labelled using DECAprime II labelling system ( Ambion ) and hybridized to the filters . Chromosomes from leukocyte–depleted , Percoll-purified late stage WT and PY01365-KO parasites were fractionated by contour-clamped homogeneous electric field ( CHEF ) electrophoresis as described [18] . The gel was blotted and hybridized sequentially with three different probes . First a probe that binds to the 5′ coding region of py235ebp-1 ( Fragment C ) ; second , a probe that binds to the 3′ coding region of py235ebp-1 ( Fragment B ) ; and finally , a probe that binds to the 3′ UTR of DHFR/TS . Erythrocytes were harvested from BALB/c mice infected with P . yoelii YM WT or PY01365-KO lines and depleted of leukocytes [23] . Py235 protein from both supernatant and detergent solubilized parasite preparations was purified on separate columns by affinity chromatography using mAb 25 . 77 as described previously [20] . Affinity chromatography using mAb 25 . 37 was also carried out as above but only using WT parasites . Proteins eluted from the affinity columns were subjected to SDS-PAGE under reducing conditions on a 5% polyacrylamide gel and visualised using colloidal blue stain ( Novex ) . Bands were excised , reduced , alkylated and digested with trypsin [56] . Peptide mass fingerprinting was carried out using a Reflex III MALDI-ToF mass spectrometer ( Bruker Daltonik , Germany ) . The peptide mass fingerprints were used to query sequences in both the rodent malaria database [15] and the general non-redundant database at the National Centre for Biotechnology Information , ( NCBI; http://www . ncbi . nlm . nih . gov ) . The gene accession numbers identified were used to carry out further searches of the NCBI database to obtain full gene sequence information . Two of the three py235 sequences identified by mass spectrometry analysis using peptides from 25 . 37-affinity purified protein , did not correspond to full-length genes , instead they corresponded to two contigs , PY05995 and PY03534 . A gene specific forward primer was designed based on DNA sequences from the last 200 nucleotides of the 3′ coding sequence of PY05995 ( Forward primer 5′-GAAATGAAACGTACAAAAGATGACATC-3′ ) , and a reverse primer was designed based on the first 200 nucleotides of the 5′ coding sequence of PY03534 ( Reverse primer 5′-CTGTATATGATTGTTCTATTAAATTAC-3′ ) . Using WT gDNA , PCR amplification was carried out using Pfu ultra DNA polymerase ( Strategen ) . The PCR product was directly sequenced ( Cogenics ) , analysed , aligned and assembled with the PY05995 and PY03534 contigs to create a single py235 gene , using Bioedit [53] . Total RNA was prepared [57] from leukocyte-depleted , Percoll-purified late stage WT and PY01365-KO parasitized erythrocytes using Trizol ( Invitrogen Life Technologies ) . RNA samples were first treated with RNase-Free DNase I ( Quiagen ) and cleaned up using RNeasy MiniElute to remove contaminating gDNA . First strand cDNA was synthesized using 1 µg RNA , AMV reverse transcriptase ( RT ) and random primers according to the manufacturer's instructions ( Promega ) . RT-PCR amplification using the synthesized cDNA was carried out and samples amplified without the addition of RT were included as controls . Gene specific primers were designed for the Py235 genes of interest PY01365 , PY01885 , and PY05995/PY03534 and the reference genes , P . yoelii erythrocyte binding protein ( PyEBL ) and Pyβ-tubulin ( Table S1 ) . Short regions of the genes ( 150bp–193bp ) were amplified using gDNA extracted from purified late stage WT P . yoelii YM parasites . cDNA was used as a template to PCR amplify Pyβ-tubulin . PCR products were cloned into TA vector , and clones containing inserts were identified by PCR and the insert DNA verified by sequencing . qPCR was carried according to the MIQE guidelines [58] . qPCR reactions ( 25 µl ) were set up in triplicates in Absolute SYBR Green mix ( containing Thermo-Start , DNA polymerase and ROX Dye ) ( Abgene ) , 0 . 2 µM each primer and 1 µl cDNA and amplified in an ABI Prism 7000 Sequence Detection System ( Applied Biosystems ) . Cycle conditions were 50°C , 2 min; 95°C , 15 min; 40 cycles of 95°C , 15 s; 60°C , 1 min . gDNA was used to check that amplification efficiencies of primers were comparable and plasmids used to generate standard curves were included in each assay . Transcript levels for each gene in the WT and PY01365-KO parasite lines were quantified and normalized with Pyβ-tubulin and PyEBL . For analysis , cDNA prepared from two independent RNA samples was used . Parasites were purified using a MACS type-D depletion column with a SuperMACS II magnetic separator ( Miltenyi Biotec GmbH ) [59] . RNA from the purified WT and PY01365-KO parasite lines , prepared as described above , was sequenced and analyzed as described [44] . Briefly , the RNA of both samples was depleted of ribosomal RNAs with exonuclease and sequenced on an Illumina GA II platform using the Illumina RNA-seq protocol . Of the approximately 61 million 76-base pair paired-end reads per run , around 95% percent mapped with SSAHA2 [60] against the P . yoelii 17XNL genome sequence ( GeneDB: ftp: ftp://ftp . sanger . ac . uk/pub/pathogens/P_yoelii/June_2010/ ) . From the coverage of the uniquely mapped Illumina reads , a perl script was used to calculate the geometric mean for expression of each predicted gene , representing the level of messenger RNA . We compared the expression in both samples by obtaining the ratio of expression values for each gene . IFA assay of Py235 protein expression in WT and PY01365-KO parasites was carried out using mAb 25 . 77 on formaldehyde fixed parasitized erythrocytes , followed by Alexa Fluor 488-conjugated affinity purified goat anti-mouse IgG ( Molecular Probes ) . In colocalization studies EBL was detected using rabbit antibodies provided by Dr Osamu Kaneko . Alternatively the slide was first probed with mAbs specific for either RON4 ( 48F8 ) or for AMA1 ( 45B1 ) [39] followed by Alexa Fluor594 congugated secondary antibody . After washing , this was followed by incubation with mAb 25 . 77 directly congugated to Alexa Fluor488 . In a separate assay , Alexa Fluor 488-conjugated 25 . 77 mAb was used to probe thin blood smears of mixed stage WT P . yoelii YM parasites , followed by Alexa Fluor 594-conjugated 25 . 37 mAb in a dual labeling experiment [61] . mAbs were labeled with Alexa Fluor ( Molecular Probes ) succinimidyl esters according to the manufacturer's instructions . The labeled antibodies were separated from excess labelling reagent by gel filtration on PD-10 columns ( Amersham Pharmacia ) , eluted using PBS/1% BSA . All slides were examined and images captured on an Axioplan 2 imaging system ( Zeiss ) . Percoll-purified late stage parasites were solubilized under reducing conditions in a buffer containing DTT , resolved by SDS-PAGE on a 5% Bis-Tris polyacrylamide gel and transferred onto nitrocellulose membrane . Primary antibodies ( at 10 µg/ml ) were used to immunostain the membrane and were detected by incubation with HRP-congugated goat anti-mouse IgG ( H+ L ) antibody ( Bio-Rad ) and the ECL Western Blotting detection reagent ( GE Healthcare/Amersham ) . Protein bands were visualized on a Kodak BioMax MR film . The blots were stripped with Restore PLUS according to the manufacturer's instructions and then probed with mAb 48F8 to detect PyRON4 [41] as a control for protein loading . [35S]methionine/cysteine ( Promix , GE Healthcare , Little Chalfont , UK ) ) radiolabeled proteins from P . yoelii YM ( WT and PY01365-KO ) either released into culture supernatant , extracted in a buffer containing 0 . 5% ( w/v ) sodium deoxycholate , or eluted from erythrocytes were immunoprecipitated using mAbs 25 . 77 and 25 . 37 , hyperimmune serum ( HIS ) and normal mouse serum ( NMS ) . The erythrocyte binding assay and immunoprecipitations were carried out as described previously [24] . Groups of 5 Balb/c mice were infected i . v . on day 0 with either 200 , 1000 , or 5000 WT or PY01365-KO parasitized erythrocytes [62] . Blood smears from each mouse were made daily from D3 , stained with Giemsa's reagent and infected cells counted to monitor the course of infection .
|
Malaria parasites invade erythrocytes where they develop and multiply before bursting out and invading fresh cells . There are sequential steps to invasion; early in the process , specific parasite proteins bind to molecules on the surface of the erythrocyte . Tight binding forms a junction between parasite and host cell leading to the next steps in the invasion process . Several of these parasite proteins , which establish contact with the host cell surface , are coded by gene families . One family , first described in the rodent parasite Plasmodium yoelii and found in all Plasmodium spp , is often referred to as the reticulocyte binding ligand family . In P . yoelii the proteins are called Py235 and are coded by at least eleven genes . Previously , we identified one family member which is the target of protective antibodies that prevent parasite invasion . Here we have deleted the gene for this protein and examined the consequences . Other members of the family take the place of the missing protein but their genes are not up-regulated . The family provides the parasite with the potential to recognize erythrocytes with different surface receptors and evade the binding of protective antibodies through plasticity at the level of its adhesion molecules .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/gene",
"expression",
"microbiology/parasitology",
"infectious",
"diseases/protozoal",
"infections",
"cell",
"biology/cell",
"adhesion"
] |
2011
|
Targeted Disruption of py235ebp-1: Invasion of
Erythrocytes by Plasmodium yoelii Using an Alternative Py235
Erythrocyte Binding Protein
|
A crucial step in the development of muscle cells in all metazoan animals is the assembly and anchorage of the sarcomere , the essential repeat unit responsible for muscle contraction . In Caenorhabditis elegans , many of the critical proteins involved in this process have been uncovered through mutational screens focusing on uncoordinated movement and embryonic arrest phenotypes . We propose that additional sarcomeric proteins exist for which there is a less severe , or entirely different , mutant phenotype produced in their absence . We have used Serial Analysis of Gene Expression ( SAGE ) to generate a comprehensive profile of late embryonic muscle gene expression . We generated two replicate long SAGE libraries for sorted embryonic muscle cells , identifying 7 , 974 protein-coding genes . A refined list of 3 , 577 genes expressed in muscle cells was compiled from the overlap between our SAGE data and available microarray data . Using the genes in our refined list , we have performed two separate RNA interference ( RNAi ) screens to identify novel genes that play a role in sarcomere assembly and/or maintenance in either embryonic or adult muscle . To identify muscle defects in embryos , we screened specifically for the Pat embryonic arrest phenotype . To visualize muscle defects in adult animals , we fed dsRNA to worms producing a GFP-tagged myosin protein , thus allowing us to analyze their myofilament organization under gene knockdown conditions using fluorescence microscopy . By eliminating or severely reducing the expression of 3 , 300 genes using RNAi , we identified 122 genes necessary for proper myofilament organization , 108 of which are genes without a previously characterized role in muscle . Many of the genes affecting sarcomere integrity have human homologs for which little or nothing is known .
Muscle tissue is important for humans in a myriad of processes including movement , digestion , and the pumping of blood through the cardiovascular system . Afflictions that affect muscle can be debilitating in any of these processes . The underlying cause of many myopathies lies within the cells that make up muscle tissue . Specifically , defects in components of the functional repeat unit of muscle , the sarcomere , are implicated in over 20 diseases [1] . The nematode Caenorhabditis elegans is a valuable model organism for the study of muscle due to the similarity of worm body wall muscle to vertebrate muscle , along with its semi-transparent cuticle that allows for visualization of muscle structures in vivo . The basic protein components within a C . elegans sarcomere have vertebrate counterparts , with only minor differences in protein composition and organization [2]–[4] . Much of the work focused on sarcomere assembly is restricted to muscle cells that contain multiple sarcomeres and make up body wall muscle in C . elegans . Building a functional sarcomere in C . elegans is a complex process that requires the assembly of two main attachment complexes , the M-line and dense body . Both of these structures are anchored in the sarcolemma , projecting inwards to allow for anchoring of actin filaments in the case of dense bodies , or myosin filaments in the case of the M-line . This anchorage is necessary to transmit the force created from the contraction of myofibrils to the muscle cell basement membrane [2] , [5] . Many of the proteins needed to form a functional sarcomere in the worm are homologs of proteins required in vertebrate focal adhesion complexes [3] . Adhesion complexes are found in migrating cells involved in a number of processes including tissue repair , immune responses , and tumor formation [6] . Much of what we know about the protein composition of adhesion complexes in C . elegans muscle was uncovered via mutational screens . Uncoordinated movement ( Unc ) and Paralyzed and Arrested at Two-fold stage ( embryonic arrest , Pat ) are the two primary phenotypes used to identify muscle mutants in the worm . An Unc phenotype can be caused by a defect in muscle cells , the neural architecture or neuronal function which triggers their contraction . The Pat phenotype is more severe and appears to be primarily muscle specific in nature , as almost all Pat mutants identified by Williams and Waterston ( 1994 ) [7] have since been shown to be components of the attachment complex or associated with the basement membrane . All of the proteins linked to this phenotype are essential for the initial assembly of the sarcomere and/or involved in the attachment of muscle cells to the basement membrane [7] , [8] , including UNC-52/perlecan [9] , PAT-3/β-integrin [10] , PAT-4/ILK [11] and DEB-1/vinculin [12] . Proper muscle function , attachment to the basement membrane and communication with the underlying hypodermis are critical for elongation of embryos beyond the two-fold stage [7] . Interestingly , less severe mutations in some Pat genes lead to an Unc phenotype due to partial function of the protein . The unc-89 , unc-95 and unc-98 genes also encode proteins found in the attachment structures that anchor the myofilament lattice [13]–[15]; however , loss of their gene products is associated with an Unc , rather than a Pat , phenotype . Animals homozygous for null mutations of another gene class , the Dim ( disorganized muscle ) mutants , exhibit a far less acute phenotype . These animals are wild type in appearance and movement , but display minor muscle defects upon close observation [16] . The work done in C . elegans first demonstrated the important role of the basement membrane protein UNC-52/perlecan in the initial assembly of adhesion structure in the muscle cell membrane [8] , [9] , [17] as well as the key role of the UNC-112 protein and , by inference , its human orthologs Kindlin-1 , 2 , 3 in the assembly of integrin-containing attachment complexes [18] , [19] . Furthermore , studies on flies and worms demonstrate that ILK is an adaptor molecule and all of its behavior in regards to muscle attachment can be explained as an adaptor molecule , not as a kinase [11] , [20] , [21] . The many studies of adhesion receptors that have been done over the years have led to the idea that adhesion complexes are platforms where many molecules meet to organize the cytoskeleton and trigger signalling . Recent studies highlight the striking similarity between vertebrate focal adhesion plaques and C . elegans muscle adhesion structures and position integrin linked kinase , as well as FERM and LIM domain proteins as central players at focal adhesions [3] , [22] . In humans , mutations in several sarcomeric and sarcolemmal proteins have been shown to cause muscular dystrophy and cardiomyopathies ( reviewed in [23] ) . For example , loss of Kindlin-1 , an actin-ECM linker protein first described in C . elegans as UNC-112 [18] , causes Kindler syndrome , characterized by neonatal blistering , sun sensitivity , muscle atrophy , abnormal pigmentation , and fragility of the skin [24] . Clearly , an understanding of muscle development and function is an important step towards the development of treatments for muscle-related disease . To date , mutational screens for genes affecting C . elegans body wall muscle have identified about 70 genes . Recent studies using DNA microarray technology have identified thousands of transcripts that are made in muscle cells . Using this method of transcriptional profiling Roy et al . [25] were able to identify 1 , 364 genes enriched in the body wall muscle of L1 larvae , and Fox et al . [26] were able to identify 5 , 710 genes expressed in embryonic body wall muscle . Another method of identifying large numbers of transcripts from isolated tissue specific RNA is Serial Analysis of Gene Expression or SAGE [27] . This technology differs from gene chip technology because it allows the identification of all transcripts including those not previously identified by analysis of the sequenced genome . Here , we use this technology to identify over 7 , 900 genes that are potentially expressed in embryonic muscle cells . Several large-scale transcriptional profiling screens using DNA microarray technology and more recently RNA-Seq have been done on various mouse and human tissues and cell lines including muscle [28]–[30] . In this study we describe the first SAGE gene expression profile done on muscle tissue in C . elegans . In addition to generating a list of genes expressed in embryonic muscle cells , we have carried out two RNAi screens designed to identify muscle-affecting genes in C . elegans . Screening was performed using a subset of about 3 , 600 genes that were found to be expressed in muscle using both SAGE and microarray analysis [26] and thus have a high probability of being expressed in body wall muscle . In order to find as many muscle affecting genes as possible , our screens were designed to identify myofilament-affected animals , regardless of whether the worms exhibited a visible phenotype or appeared wild type . We have identified 122 genes involved in proper myofilament organization , 108 of which had no previously characterized role in muscle structure or function .
The C . elegans embryo has a total of 123 muscle cells , the majority of which ( 81 ) are body wall muscle cells [31] . The myo-3 gene encodes a myosin heavy chain protein which is the central component of body wall muscle thick filaments and is present in at least 85 embryonic muscle cells [32] . Myosin protein expression is first detected during mid embryogenesis . Tagging MYO-3 with GFP and expressing it under the control of its own promoter is therefore an excellent way of identifying muscle cells prior to and during early myofilament formation using GFP fluorescence [33] ( see Figure 1A and 1B ) . We have used a strain expressing a GFP-tagged MYO-3 protein [34] to obtain a relatively purified population of embryonic muscle cells after Fluorescence Activated Cell Sorting ( FACS ) . Whole embryos were manually disrupted to obtain a suspension of mostly single cells that was then passed through a FACS machine to collect viable GFP expressing cells ( see Figure 1C–1F ) . Typically , we obtained a range of 90% to 95% GFP expressing cells at the final step . Only samples that consisted of at least 90% fluorescing MYO-3::GFP live cells were used for RNA extraction . We have constructed two Long SAGE biological replicate libraries using RNA isolated from our FAC-sorted embryonic muscle cells ( data available at http://elegans . bcgsc . bc . ca ) . A total of 49 , 655 and 120 , 825 sequenced tags were obtained from the SWEM1 and SW031 libraries , respectively . We used a sequence quality filter of 0 . 99 to reduce the number of tags incorrectly assigned due to sequencing errors . This gave us adjusted tag numbers for the two libraries of 33 , 827 and 89 , 561 , respectively and these numbers were used for all subsequent analyses . By combining the data from both libraries we identified 10 , 724 unique sequence tags representing 7 , 974 genes expressed in embryonic muscle cells . Despite the difference in sampling size between these two libraries almost 88% of the genes identified are present in both SWEM1 and SW031 . In addition , the correlation coefficient of tag counts observed for genes found in the two libraries is 0 . 94 , indicating that variability between these samples is low ( Figure 2 ) . This result is comparable to the correlation coefficient observed for Affymetrix GeneChip biological replicates ( see for example [26] ) . Not all of the 7 , 974 genes we identified in the embryonic muscle libraries are necessarily specific to muscle , or even enriched in muscle . To gain some insight into these facets of expression we compared the results from the embryonic muscle libraries to those obtained from libraries using RNA isolated from purified embryonic intestinal cells ( SWEG1 ) or pan neuronal cells ( SW028 ) ( Figure 3A , [35]; data available at http://elegans . bcgsc . bc . ca ) . We found that more than half of the muscle-expressed genes ( 4 , 315 ) are also present in both the intestinal and pan neuronal libraries , while about 15% of the genes ( 1 , 294 ) we detected are candidates for unique expression within muscle ( Dataset S1 ) . The remaining 2 , 365 genes are expressed in either neurons or the intestine in addition to muscle . Similar percentages of overlapping gene expression were observed for the other two tissues as well . To identify genes with enriched expression in muscle we compared the relative abundance of sequence tags for a particular gene in the muscle libraries to the number of tags for the same gene in the whole-embryo SAGE library SWN22 . A total of 192 , 661 sequence tags were obtained for the SWN22 library . Within this data set , there are 133 , 862 tags that pass the 0 . 99 sequence quality threshold representing 12 , 165 unique tag species and 9 , 064 genes ( data available at http://elegans . bcgsc . bc . ca ) . Sequence tags for 6 , 984 of the muscle-expressed genes are also present in the whole embryo library . We used fold-change and a minimal tag count filter to determine enrichment . Depending on the fold-change cut-off that was used we observed between 1 , 459 ( 1 . 7 fold-change ) and 228 ( 5 . 0 fold-change ) genes enriched for expression in muscle ( Dataset S2 ) . Several different approaches have been taken to identify genes that are expressed in muscle cells . These include a genetic approach to isolate mutants with disorganized or non-functioning muscle , determining the cell expression patterns of proteins using antibodies or promoter GFP fusions , and identifying the transcripts present in isolated muscle cells by SAGE , GeneChip or rtPCR or cDNA analysis . Here we determine that the majority of the muscle genes identified by these other methods is present in our SAGE transcriptome . Previous genetic analyses have identified about 70 genes that , when mutated , affect muscle structure or function in some way . Sixty-four of these known muscle-affecting genes are included in the 7 , 974 genes that define our muscle transcriptome ( Dataset S3 ) . Some muscle affecting genes that were not found in the muscle transcriptome encode proteins that are expressed in the hypodermis ( vab-19 and mua-1 ) or in neurons ( ace-2 ) and not muscle [36]–[38] , while others like unc-22 are expressed in muscle cells but just not during embryogenesis [39] . A large-scale project involving the mapping of cell expression patterns of promoter GFP fusions has identified 587 genes that are expressed in muscle; data available at http://www . wormatlas . org/ ) . We have determined that 452 of these genes ( ∼77% ) are present in at least one of the two SAGE embryonic muscle libraries ( Dataset S4 ) . Interestingly , GFP expression in body wall muscle cells was observed for 39 genes that were represented by only a single SAGE tag and 167 genes that were represented by fewer than 5 SAGE tags in either SAGE library . We also compared the muscle transcriptome generated by SAGE to one generated using an alternative gene expression analysis platform , the Affymetrix GeneChip [26] . The GeneChip study is comparable to our study as the analysis was also done using RNA isolated from FAC sorted embryonic muscle cells . The Fox et al ( 2007 ) study [26] identified 5 , 170 different transcripts in freshly sorted embryonic muscle cells after direct isolation at ‘0’ hours . A comprehensive list of 9 , 567 genes can be obtained by combining the SAGE and Affymetrix data; however , only 3 , 577 ( 37% ) of these genes overlap the two data sets ( Figure 3B; Dataset S5 ) . As might be expected more highly expressed genes are more likely to be detected by both platforms . The mean sequence tag count for the genes detected by both SAGE and Affymetrix GeneChips is 19 tags per 100 , 000 , but only 6 tags per 100 , 000 for the genes whose expression is detected by SAGE only . Similarly , the median sequence tag counts are 6 tags per 100 , 000 and 3 tags per 100 , 000 , respectively . The moderate correlation ( 0 . 77 ) of SAGE sequence tag counts per gene and Affymetrix signal intensities for the 3 , 577 genes determined to be ‘present’ in both the SAGE and Affymetrix GeneChip data is consistent with previously published comparisons between SAGE and Affymetrix GeneChips [40] . Partially because of the large dynamic range of individual gene expression , ( i . e . three orders of magnitude in our study ) low abundance messages are a problem for all expression platforms to detect . They simply get swamped out by the more abundantly transcribed messages and are often lost during the stochastic process of sampling the RNA populations . For our SAGE studies we define a transcript as ‘present’ if it is represented by at least one sequence tag , even though there is some controversy as to whether singletons ( i . e . single sequence tags representing a gene ) are a reliable measure of gene activity [40] . The majority of the 4 , 397 genes that were identified in our SAGE study , but that were not identified in the Afffymetrix GeneChip study are low abundance transcripts with an average sequence tag count of 6 per 100 , 000 . Before initiating a functional study of the muscle transcriptome we felt it prudent to refine the transcriptome and only study genes with a high probability of actually being expressed in muscle cells . We have generated a refined embryonic-muscle transcriptome consisting of the 3 , 577 genes that are present in at least one of the two SAGE libraries and also in the GeneChip library generated by Fox et al , 2007 [26] ( Figure 3B , Dataset S5 ) . Our refined embryonic-muscle transcriptome identified 3 , 577 genes that are expressed in muscle cells . A small number of these genes have been studied in great detail and much is known about their role in a functioning muscle cell ( reviewed in [3] ) . However , little or nothing is known about the function and/or relevance to muscle development of the majority of the genes identified by transcriptional profiling . Gene inactivation or knockdown by RNAi has been used successfully to screen much of the C . elegans genome and possible functions have been suggested for some genes [41]–[43] . Rather than simply repeat previously performed RNAi screens , we have designed two muscle specific RNAi screens , one to identify genes essential for embryonic muscle formation , and another to identify novel genes involved in myofilament assembly . Both screens utilized the RNAi feeding library constructed by the Ahringer lab [42] . For various reasons we were not able to obtain double stranded RNA for all 3 , 577 genes in our muscle transcriptome , screening 85 and 92% of the genes respectively . The purpose of our first RNAi screen was to identify genes essential for embryonic muscle formation . Previous work by Williams and Waterston ( 1994 ) [7] has shown that animals lacking functional muscle exhibit a similar embryonic arrest phenotype designated the Pat phenotype . To date , 19 genes that affect muscle have been shown to produce this phenotype when mutated [7] , [9]–[12] , [17] , [44]–[51] . Here we utilized the approach outlined in Figure 4 pathway A to screen for Pat mutants after RNAi treatment of wild type animals . Our initial screen of 3 , 031 genes was done in liquid culture and identified 371 RNAi treatments resulting in embryonic lethality , sterility , or movement abnormalities . After eliminating well characterized and published genes , the remaining 170 positive genes were tested again using RNAi feeding on solid media . Further analyses revealed that four genes T27B1 . 2 , F31D5 . 3 , T28B4 . 3 , and F25B3 . 6 consistently produced Pat animals when inactivated . We were also able to identify effects on post-embryonic muscle development by feeding double stranded RNA to wild type animals after hatching . Adult hermaphrodites were observed using polarized light microscopy and significantly disorganized body wall muscle was seen after RNAi treatment for the genes T27B1 . 2 , F31D5 . 3 , and T28B4 . 3 ( Figure 5 ) , but not F25B3 . 6 ( data not shown ) . The second RNAi screen utilized the approach outlined in Figure 4 pathway B , and involved screening for defects in myosin localization , using the MYO-3::GFP expressing strain RW1596 [34] . The GFP-tagged MYO-3 protein in this strain is transcribed from a modified gene contained on a transgenic array and is the only myo-3 gene product produced . Loss of the array results in animals that arrest at the two-fold stage of embryogenesis ( the Pat phenotype ) . It has been shown previously that RW1596 animals experience an acute loss of myofilament stability with age [52] . To avoid false positive results in the screen , we first analyzed synchronized RW1596 worms from young adults until 3-days of adulthood for abnormalities in MYO-3::GFP expression . We found that the percentage of animals displaying myosin defects increased from 22 . 8% in the L4/young adult stage to 31 . 9% in 1-day old adults , 42 . 5% in 2-days old adults , and 75% in 3-days old adult animals . Figure 6 shows the observed myofilament abnormalities in aging RW1596 animals as visualized by GFP fluorescence microscopy ( Figure 6A–6D ) as well as polarized light microscopy ( Figure 6F ) . This age related myofilament instability does not become apparent in N2 wild-type animals until much later ( Figure 6G; our unpublished data ) . Thus , it would appear that the presence of the GFP-tagged MYO-3 protein renders the myofilaments more susceptible to perturbation than normal , providing a sensitized background for the detection of sarcomere affecting genes . As a result of this analysis , we screened only L4/young adult animals in the following RNAi experiment . In a preliminary screen , we examined a total of 3 , 297 genes for RNAi-induced myofilament defects ( Figure 7A ) . Briefly , adult hermaphrodites that had been grown on plates containing RNAi feeding bacteria were transferred to fresh plates containing the same bacteria for 24 hrs and then removed . Their progeny were scored as L4 larvae or young adults . After a brief inspection for overt phenotypic anomalies , animals were scored for myofilament defects and then assigned to one of four classes . In order for RNAi-treated hermaphrodites to be placed in the High class ( HC ) , 75 to 100% of the animals screened had to exhibit some sort of MYO-3 abnormality . The proportions of affected animals for the Intermediate ( IC ) , Low ( LC ) and Wild type ( WT ) classes were 50 to 75% , 25 to 50% and 0 to 25% , respectively . The majority of the genes that were tested caused either a Low ( 1 , 710 ) or a Wild type ( 1 , 252 ) class phenotype when inactivated by RNAi treatment . We only considered the RNAi treatments resulting in High or Intermediate phenotypes to be muscle-affecting . To confirm this data , we repeated the RNAi treatments for the 290 genes that fell in these two categories . In addition , a slightly modified screen ( see Materials and Methods ) was used to test the 45 essential genes identified in the primary screen that resulted in few , if any , adult progeny . In the re-screen , both the number of analyzed animals and the scoring stringency were increased to reduce false-positive results . We required an animal to have myofilament defects in multiple muscle cells to be deemed affected . After this second round of screening , we identified a total of 118 genes affecting MYO-3::GFP localization and/or stability in the RW1596 strain ( Figure 7B ) . The remaining genes failed to meet the imposed stringency of the second screen . The clones from the RNAi feeding library used for each of the 118 positive treatments were verified by PCR using gene specific primers , or in some cases , sequencing . Fifty-eight of the 118 genes that affect myosin localization and/or stability did not display any obvious anatomical or behavioral phenotypes when knocked down by RNAi and , therefore , would not have been identified in a conventional screen . The MYO-3::GFP abnormalities observed have been arranged into three categories that are described below and in Dataset S6 . In some cases gene inactivation by RNAi resulted in more than one of these myofilament abnormalities , albeit in different animals . This could be due to variability in gene knockdown from animal to animal . Figure 8A and 8B show the phenotypes observed for the control experiments that were included in this screen . The animal in Figure 8A was fed bacteria containing an empty L4440 feeding vector and represents the unaffected state . The MYO-3::GFP protein is well ordered in the myofilament lattice and there are no unlocalized clumps of GFP . The animal in Figure 8B was fed dsRNA for GFP and has a marked reduction in the amount of MYO-3::GFP protein produced . The most common phenotype observed after RNAi knockdown is a general disorganization of the myofilaments , including minor GFP aggregations , that somewhat resembles the age related sarcopenia observed in older RW1596 hermaphrodites . Examples of this phenotype are shown in Figure 8C and 8D . Eighty-two of the genes we tested have been assigned exclusively to this category ( i . e . this was the only phenotype observed after RNAi knockdown ) , although most animals assigned to the two other categories also exhibited this phenotype in some muscle cells . The two other distinct phenotypes that we observed are less common . A total of 22 genes have been assigned to the second category . In muscle cells exhibiting this phenotype , the MYO-3::GFP containing filaments are well ordered but small aggregations of GFP appear along the filaments ( shown in Figure 8E and 8F ) . In many cases large or small gaps in the lattice are also present indicating that some myofilaments have detached from the muscle cell membrane . Only 13 genes have been assigned to the third category . The MYO-3::GFP abnormalities in muscle cells exhibiting this phenotype are characterized by large GFP deposits , and are accompanied by disorganization of the myofilaments ( shown in Figure 8G and 8H ) . Again large or small gaps in the lattice may also be present . Only one gene did not fall exclusively into any one category . Animals fed dsRNA for the known muscle gene unc-112 displayed myofilament abnormalities consistent with both categories 2 and 3 . The other 13 known muscle genes identified in this screen fall into all three phenotypic categories; unc-45 , uig-1 , unc-23 , and mup-4 were assigned to category 1 , dim-1 , unc-95 , and unc-15 were assigned to category 2 and pat-4 , pat-6 , vab-10 , unc-89 , act-3 , and tnt-3 were assigned to category 3 ( see Dataset S6 ) . All available information about our screen , including diagnostic images , may be accessed at http://www . zoology . ubc . ca/~alorch/rnai A subset of 22 genes that scored positive in the GFP reporter myofilament screen was also analyzed by RNAi in N2 wild-type animals using polarized light . The observed phenotypes in body wall muscle range from minor structural irregularities making it very difficult to visualize filaments under polarized light to severe disorganization ( Figure S1 ) . Six out of 22 genes did not result in a phenotype distinguishable from wild type filament structure using this approach and would therefore have been missed without the sensitized background . Together our two RNAi screens have identified 108 genes not previously known to be involved in C . elegans muscle development and/or function ( Dataset S6 ) . Only 15 of the 122 genes we identified in our screens were previously characterized and shown to be required for proper muscle function . Working from GO annotation terms we have determined that several of these new muscle-affecting genes fall into the categories of cell structure/cellular processes , metabolism , gene expression , signal transduction , protein processing and protein-protein interactions . Several of these newly identified muscle-affecting genes have no assigned GO category and are , therefore , of completely unknown function . Based on Ensembl annotations ( http://ensembl . org ) , about 58% of these new muscle-affecting genes encode proteins with sequence similarity to human proteins ( Dataset S6 ) . In some cases protein function can be inferred from the information known about the human ortholog . However , that still leaves many genes without a described function .
In this study we have combined two large scale technologies , transcriptional profiling using SAGE and gene inactivation using RNAi , to identify novel genes involved in myofilament assembly and/or stability . Our SAGE data identified 7 , 974 protein-coding genes that are expressed in embryonic muscle cells , whereas a very similar study using Affymetrix GeneChip technology identified 5 , 170 genes expressed in embryonic muscle [26] . In an attempt to avoid sampling bias and low-level contamination by non-muscle transcripts we chose to focus on genes identified by both expression platforms; these would , presumably , have a higher probability of being expressed in muscle cells . Our more limited embryonic muscle transcriptome consists of ∼3 , 500 genes , and we used gene inactivation or knockdown by RNAi to determine the function of some of these genes . We generated two replicate long SAGE libraries for sorted embryonic muscle cells , identifying 7 , 974 protein-coding genes , including 64 previously described muscle affecting genes . More than half of the genes identified by SAGE were not detected using GeneChip technology , although the majority of these are represented among the 22 , 150 transcripts represented on the Affymetrix C . elegans genome array [26] . Conversely , about one third of the genes detected by the GeneChip microarray are missing form the SAGE libraries . Presumably a large portion of this discrepancy is due to sampling and distinguishing signal from noise . Consistent with previously published comparisons most of the genes detected by SAGE but not detected by GeneChip technology ( ∼68% ) have fewer than 5 sequence tags/library [40] . The total number of genes with only one sequence tag ( singletons ) in either SW031 , SWEM1 or both libraries accounts for 31 . 4% of the total number of genes identified by SAGE analysis . In contrast , less than 19% of the total number of genes identified by both SAGE and GeneChips are singletons , 675 of the 2504 singletons ( 27% ) identified by SAGE are present in the Affymetrix GeneChip study . In addition , we confirmed a number of singletons and genes with low tag counts ( <5 ) to be present in muscle cells by GFP reporter studies . These data demonstrate that a not insignificant number of genes are transcribed at very low levels in muscle and one cannot simply ignore these low level transcripts if one wishes to determine the full transcriptome of any specific tissue or cell type . In future deep sequencing SAGE libraries may be the best way to surmount the difficulty of gene expression across four orders of magnitude . Low-level contamination will still be a problem but at least detection will not be an issue . As expected , many of the genes identified in embryonic muscle also function in several types of cells . Comparing the data obtained from the muscle , intestinal and pan neuronal libraries revealed that between 11–16% of the genes identified in any one tissue are specific to that tissue , although these estimates are possibly high . Some of the genes in the muscle embryonic library are enriched in muscle when compared to the other tissue specific libraries or the whole embryo library . Generating SAGE library expression profiles is useful for identifying candidate muscle genes for further functional studies . We have identified 228 genes enriched 5-fold or more in embryonic muscle cells compared to the whole embryo , including known genes like mup-2 ( troponin [49] , 21 . 8-fold ) , unc-95 ( LIM-domain containing protein [14] , 7 . 7-fold ) , and mlc-2 ( myosin light chain [53] , 14 . 3-fold ) as well as many unknown or uncharacterized genes ( Dataset S2 ) . Lowering the fold-change cut-off to 1 . 7 reveals 1 , 459 muscle-enriched genes . Our RNAi screens identified 108 new muscle-affecting genes . Only four of these were identified in the first screen that specifically looked for genes required for embryonic muscle function . Many embryonic lethal mutants have been found in various RNAi screens [43] , [54] , but until now there has not been a screen focused solely on identifying animals exhibiting the Pat phenotype . In fact , two of the four genes found in our study ( F31D5 . 3 and T28B4 . 3 ) have not been reported to cause embryonic lethality in any previous RNAi screens . Several genetic studies have identified at least 19 loci that encode proteins essential for muscle formation during early development ( reviewed in [3] ) . Most of these genetic loci , with the exception of pat-9 and pat-11 , have been identified at the sequence level and thus are correlated with the physical map . Seventeen of the 19 known Pat genes were included in our muscle transcriptome , and our screen identified only eight of these . Given these results it would seem that RNAi knockdown by feeding is not a very efficient method for identifying Pat mutants . If this is the case , there are probably more “pat” genes to be discovered . Injection of double stranded RNA instead of feeding has been used quite successfully by others [54] , [55] and our preliminary RNAi knockdown treatments using this method suggests that it is a much more effective , but also more time consuming and expensive way to identify Pat mutants . The majority of the new muscle-affecting genes that were discovered were identified by specifically screening our refined embryonic muscle transcriptome for disorganized myosin filaments using a MYO-3::GFP marker strain , RW1596 [34] . The only functional myosin heavy chain A ( MYO-3 ) protein produced in these animals is fused to GFP , a bulky , tubular protein composed of 238 amino acids [33] . We have discovered that the myofilaments in the body wall muscle of RW1596 hermaphrodites are more sensitive to perturbation and the effects of aging than those in the body wall muscle of wild type hermaphrodites . Presumably , the barrel-shaped GFP reporter fused to MYO-3 prevents the tight packing of these molecules in the central bipolar H zone of thick filaments . This would negatively affect the levels of myosin heavy chain B ( UNC-54 ) and paramyosin ( UNC-15 ) , as well as prevent the proper assembly of these molecules into a functional thick filament [56] . Thus , it is very likely that GFP-tagging of the MYO-3 protein results in looser packing of the myosin rods , thereby making the strain more sensitive to the absence of auxiliary proteins resulting from RNAi treatment . It should be noted that several genes previously shown to be required for proper muscle assembly and maintenance were not identified in this muscle RNAi screen . For example , treatment of animals with pat-3/ß-integrin dsRNA , despite having displayed both sterility and uncoordinated movement , did not exhibit a highly penetrant disorganized myofilament phenotype , and therefore it was assigned to the LC class . There are several possible explanations for not detecting previously described muscle affecting genes , including the inherent variability in RNAi screening , loss of the genomic insert in the bacterial clone , or the means by which we screened animals for myofilament defects . Validation of our approach was provided when we identified several known muscle-affecting genes in the RNAi screens ( Dataset S6 ) . For example , the UIG-1 protein was identified as a novel dense body protein in a yeast two-hybrid screen for UNC-112 interacting proteins [57] . Although animals homozygous for the uig-1 ( ok884 ) null mutation appear healthy and fertile , disorganization of the body wall muscle filaments was observed in a polarized light assay [57] . In our RNAi screen , as well as in previous RNAi screens [43] , [54] , [58] , [59] , knockdown of the uig-1 gene by RNAi treatment did not result in any obvious morphological or growth phenotypes . However , in the myofilament screen we were able to detect MYO-3::GFP disorganization in 55% of scored animals ( n = 63 ) , resulting in an intermediate class phenotype , and this phenotype correlates nicely with the earlier results [57] . The uig-1 gene is an example of a gene that affects muscle organization but does not display a Pat or Unc phenotype . Such genes would be difficult to identify in a high throughput screen without a reporter gene for sarcomere integrity . In total , we identified 58 genes without an obvious growth or anatomical phenotype when inactivated by RNAi treatment , but which displayed an intermediate or high class phenotype when investigated for MYO-3::GFP localization . While some of these genes may encode proteins involved in muscle sarcomere assembly , we suspect most are involved in accessory functions involved with sarcomere maintenance and stability . Screening specifically for myofilament abnormalities , therefore , has proven to be an effective way to identify new genes affecting body wall muscle structure and function . The myofilament abnormalities we have observed in our screen range from small aggregates to large deposits of GFP , which may be within the filaments or adjacent to them . We also see gradients of disorganization ranging from mild discontinuities in the filaments to large interruptions often accompanied by the accumulation of the MYO-3::GFP reporter protein . On occasion , we also observed a marked decrease of MYO-3::GFP signal in the myofilaments . These defects may be the result of several factors , including abnormal or aberrant localization of MYO-3::GFP , early onset sarcopenia , loss of myofilament stability and/or defective myofilament assembly . With one exception ( unc-112 ) , we assigned the 118 genes identified in our screen to different categories based on the myofilament phenotype observed after RNAi knockdown ( Dataset S6 ) . The first category represents the least severe phenotype observed and includes the majority of the genes identified , although only 4 of the 14 known muscle-affecting genes were assigned to this category . This phenotype is consistent with an early onset of the age related myofilament disorganization that is seen in older RW1596 and wild type animals . However , our current data do not allow to us to determine whether the disorganization observed is due to defective myofilament assembly , loss of myofilament stability or perhaps both . Two of the 4 known genes in this category encode molecular chaperones and another encodes a signaling molecule . The third category represents the most severe phenotype observed and has the fewest assigned total genes ( 14 ) but half of the known muscle affecting genes ( 7/14 ) . The majority of the known genes assigned to category three encode structural components of dense bodies and/or M-lines or the myofilaments , and most have been shown to be essential for the assembly of the sarcomere during embryogenesis . Annotation of our new candidate muscle-affecting genes using ENSEMBL identified many different functions . Obviously , genes encoding structural proteins or those with protein binding domains are good candidates to be involved in sarcomere assembly and/or stability , but predicted metabolic or signaling molecules can have a structural function as well , as observed for pat-4 , dim-1 and uig-1 [11] , [16] , [57] . We also identified seven genes involved in protein turnover that appear to cause defects in sarcomere structure . It has been shown previously that the degradation of UNC-45 , a molecular chaperone for myosin , by the E3/E4 complex formed by CHN-1 and UFD-2 is indispensable for proper myosin folding and assembly into thick filaments [60] , [61] . These newly identified genes may be involved in comparable pathways . Using integrated strategies like SAGE and RNAi provides an initial step towards a comprehensive analysis of all the genes required to form and maintain a muscle sarcomere in C . elegans . A recent RNAi study from the laboratory of Norbert Perrimon using Drosophila primary culture cells is , in principal , similar to our study [62] . They identified 49 genes involved in late muscle differentiation after examining 1 , 140 randomly chosen genes , and 22 of these genes were not previously known to be involved in muscle function . Thirty-five of the 49 muscle-affecting genes identified in the Drosophila RNAi study have a C . elegans ortholog . However , only 16 of these met the SAGE and microarray criteria of our study , and only 2 , unc-15 and unc-112 , were identified as affecting myofilament organization . The remaining genes were assigned to either the low class ( 9 ) or the wild type class ( 5 ) . The data obtained in these two studies reveals a significant number of new players involved in muscle development . With the identification of so many new protein involved in sarcomere formation and maintenance perhaps we are nearing the time where some of the more outstanding problems concerning sarcomere assembly can be solved .
All strains were maintained using standard culture methods [63] . The following strains were used: N2: wild-type Bristol isolate , RW1596: myo-3 ( st386 ) V; stEx30 [34] . GFP expressing muscle cells were isolated from freshly dissociated RW1596 embryos as described previously [64] with minor modifications . Embryos were treated with Chitinase ( 0 . 5 units ) to dissolve the eggshell , concentrated by high-speed centrifugation and then repeatedly passed through a 21-gauge needle to dissociate the cells . The resulting cell suspension was filtered through a 5 µm Millex-SV syringe filter , and pelleted by gentle centrifugation . Isolated cells were resuspended in ice-cold egg buffer and maintained on ice in preparation for the following FACS procedure . GFP expressing embryonic muscle cells were isolated using a FACSVantage SE Diva cell sorter , equipped with a 488 nm Argon laser and a 530±15 nm emission filter for GFP , and a 585±22 nm emission filter for propidium iodide ( used to discriminate dead cells ) . Non-GFP embryonic cells from wild type N2 worms were used as controls to set instrument settings , and to establish GFP sorting gates . Initially cells were analyzed for light scatter properties such that a population of cells in a forward scatter ( cell size ) vs . side scatter ( granularity ) plot was selected that eliminated clumps and debris . This population was then analyzed in a GFP vs . PI plot to visualize GFP positive cells . The GFP sorting gates were defined by comparing the profile of wild-type embryonic cells with that of the MYO-3::GFP expressing cells . Sorted cells were collected in 0 . 5 ml of egg buffer supplemented with 5% FBS and immediately frozen in liquid nitrogen in preparation for RNA isolation . An aliquot of cells from each sort was cultured overnight on peanut lectin coated cover slips with L-15 and 10% FBS as a culture medium . This overnight culture was examined using fluorescent-light microscopy to gauge sort purity . Library construction involved pooling RNA recovered from sorts that were deemed to have a purity of ≥90% . SAGE data for SWEM1 , SW031 , SWEG1 , SW028 , SWN22 were prepared by standard methods and analyzed as described elsewhere [65] , [66] . All information is available at http://elegans . bcgsc . ca/home/sage . html; Wormbase version WS180 ( September 2007 ) was used for gene identification . The N2 strain was used for these experiments following the protocol described previously [67] . RNAi clones were grown in 96-well plates overnight at 37°C in 150 µl of L-broth containing 50 µg/ml carbenicillin and 8% glycerol . The following morning , 10 µl of each freshly grown bacterial stock was transferred to 96-well plates of Liquid NGM media ( 50 µl per well ) containing 4 mM IPTG and 50 µg/ml carbenicillin . Liquid NGM plates were incubated for 16 hours at 20°C and then 37°C for 2 hours . Four synchronized L3 worms were then transferred into each well using a Copas Biosorter-250 ( Union Biometrica ) and incubated at 20°C for 72 hours . Each well was scored for animals that exhibited embryonic lethality , paralysis , or sterility . All RNAi treatments that resulted in animals with the aforementioned phenotypes were re-screened using solid NGM media . RNAi bacterial cultures were grown overnight in 96-well plates containing 200 µl of L-broth with 50 µg/ml carbenicillin . NGM plates containing 1 mM IPTG and 50 µg/ml carbenicillin and were then streaked with the freshly grown RNAi bacterial cultures and incubated at 20°C for 24 hours . Two L4 animals were then transferred to each plate , and incubated for 72 hours at 20°C . Two F1 progeny at the L4 stage were then transferred to each of three replicate plates for each RNAi treatment , and incubated at 20°C for 24 hours to lay eggs . All of the plates were examined for animals exhibiting an early embryonic arrest phenotype , a mixed stage embryonic arrest phenotype , or a two-fold stage arrest phenotype . The strains RW1596 and N2 wild-type was used for these experiments following the protocol described previously [41] . RNAi colonies were grown overnight in L-broth with 50 µg ml−1 ampicillin and then streaked onto NGM plates containing 1 mM IPTG and 50 µg ml−1 carbenicillin . The plates were incubated overnight at room temperature to induce dsRNA expression . Approximately 30 synchronized L1s ( P0 ) animals were spotted onto each RNAi plate . These plates were incubated at 20°C until the worms reached the young adult stage ( ∼60–68 hours ) , then 2 P0 were transferred onto fresh plates , in duplicate for each RNAi clone . To synchronize the F1 generation , the adult animals were removed after an 18-hour egg-laying period . Embryos were allowed to hatch and develop to the L4/young adult stage ( ∼36 hours ) . Prior to the myofilament screen , animals were scored for a variety of gross anatomical and growth defects such as sterility , embryonic or larval lethality , uncoordinated movement , and slow growth . Then , a random subset of worms ( 10–40 ) from each duplicate plate were picked into 15 µl M9 buffer containing 10% sodium azide and their body wall muscle was screened for MYO-3::GFP defects using a compound fluorescence microscope ( Zeiss Axiophot ) . An animal was scored positive when the majority of muscle cells showed abnormal GFP fluorescence . Animals were examined blind to the identity of the dsRNA . The empty vector ( L4440 ) was used as a negative control , and worms feeding on bacteria expressing dsRNA for unc-97 or GFP were used as positive controls on each day of screening . Digital photographs were taken using a Qimaging QICAM digital camera running Qcapture version 1 . 68 . 4 . For clones resulting in an embryonic lethal or sterile phenotype in the first screen , the RNAi screening protocol described above was modified as followed . The synchronized L1 worms were spotted onto empty vector ( L4440 ) RNAi control plates and then transferred to gene specific RNAi plates after reaching the adult stage . The F1 generation was then examined for myofilament defects as L4/young adults . All essays were carried out at 20°C . Images and observations for each gene screened were archived in an online database , created using MySQL 3 . 23 . 49 with a web interface written in PHP 4 . 3 . 4 . Hosting was set up on a server maintained by the Department of Zoology , University of British Columbia ( Canada ) .
|
Muscular diseases affect many people worldwide . While we have learned much about the sarcomere , the basic building block of muscle cells , there are still numerous questions that remain to be answered . We must learn more about proteins expressed in muscle and how they interact so that better treatments for myopathies can be developed . The nematode Caenorhabditis elegans is a valuable model organism for the study of muscle due to similarities between worm body wall muscle and vertebrate muscle , along with its semi-transparent cuticle that allows for visualization of muscle structures in live animals . We have used transcriptional profiling methods to identify the majority of genes that are expressed in the embryonic body wall muscle cells of C . elegans . To gain insight into possible functions performed by these genes and their corresponding proteins , we examined animals and muscle cells for abnormalities after the targeted inactivation of about 3 , 300 genes . We identified 122 genes necessary for proper myofilament organization , 108 of which had no previously characterized role in muscle . This approach proved to be a rapid and sensitive means to identify genes that affect muscle differentiation and sarcomere assembly .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/gene",
"discovery",
"cell",
"biology/gene",
"expression",
"cell",
"biology",
"cell",
"biology/cell",
"adhesion",
"genetics",
"and",
"genomics",
"genetics",
"and",
"genomics/bioinformatics"
] |
2009
|
An Integrated Strategy to Study Muscle Development and Myofilament Structure in Caenorhabditis elegans
|
Schistosome parasites cause schistosomiasis , one of the most important infectious diseases worldwide . For decades Praziquantel ( PZQ ) is the only drug widely used for controlling schistosomiasis . The absence of a vaccine and fear of PZQ resistance have motivated the search for alternatives . Studies on protein kinases ( PKs ) demonstrated their importance for diverse physiological processes in schistosomes . Among others two Abl tyrosine kinases , SmAbl1 and SmAbl2 , were identified in Schistosoma mansoni and shown to be transcribed in the gonads and the gastrodermis . SmAbl1 activity was blocked by Imatinib , a known Abl-TK inhibitor used in human cancer therapy ( Gleevec/Glivec ) . Imatinib exhibited dramatic effects on the morphology and physiology of adult schistosomes in vitro causing the death of the parasites . Here we show modeling data supporting the targeting of SmAbl1/2 by Imatinib . A biochemical assay confirmed that SmAbl2 activity is also inhibited by Imatinib . Microarray analyses and qRT-PCR experiments were done to unravel transcriptional processes influenced by Imatinib in adult schistosomes in vitro demonstrating a wide influence on worm physiology . Surface- , muscle- , gut and gonad-associated processes were affected as evidenced by the differential transcription of e . g . the gynecophoral canal protein gene GCP , paramyosin , titin , hemoglobinase , and cathepsins . Furthermore , transcript levels of VAL-7 and egg formation-associated genes such as tyrosinase 1 , p14 , and fs800-like were affected as well as those of signaling genes including a ribosomal protein S6 kinase and a glutamate receptor . Finally , a comparative in silico analysis of the obtained microarray data sets and previous data analyzing the effect of a TGFβR1 inhibitor on transcription provided first evidence for an association of TGFβ and Abl kinase signaling . Among others GCP and egg formation-associated genes were identified as common targets . The data affirm broad negative effects of Imatinib on worm physiology substantiating the role of PKs as interesting targets .
Schistosomiasis is an infectious disease of worldwide importance caused by parasitic platyhelminthes of the class trematoda such as Schistosoma haematobium , S . intercalatum , S . japonicum , S . mansoni , or S . mekongi . About 780 million people are at risk of schistosomiasis , and more than 240 million infections emerge annually requiring treatment [1] , [2] . Adult schistosomes live in the abdominal veins of their vertebrate hosts . Only if paired , females produce eggs , half of which reach the gut lumen ( e . g . S . mansoni ) or the bladder ( S . haematobium ) , to be transported to the environment for continuing the life-cycle . Gut invasion is accompanied by inflammatory processes . The remaining eggs migrate through the blood stream and become trapped in spleen and liver tissue , where granulomas are formed and fibrosis occurs , leading to hepatosplenomegaly and liver cirrhosis [3] , [4] . The disease has a high socioeconomic impact causing annual losses of 1 . 7 to 4 . 5 million disability adjusted life years ( DALYs ) of humans living in endemic areas [5] , [6] , but tourists and travelers can also be affected [7] . Besides humans , animals including cattle can get infected , too , which causes economic losses [8]–[11] . Since there is no vaccine available yet , the main strategy to control schistosomiasis is the regular use of drugs , of which three have the potency to kill schistosomes . Of the drugs available today Metrifonate is active against S . haematobium only and Oxamniquine is active against S . mansoni only . In contrast to these limitations , although its effectiveness against immature stages is limited Praziquantel ( PZQ ) is effective against all important schistosome species mainly affecting adults [12] . This and its low price have promoted PZQ as the drug of choice , which is also used in large-scale treatment programs today [13] , [14] . However , drug resistance has been recognized as a potential problem since several studies demonstrated PZQ resistance to be inducible in laboratory settings , and field studies provided first indications for the possibility of reduced PZQ efficacy [15]–[18] . Furthermore , multidrug transporters were discovered in schistosomes , of which one was shown to respond to a PZQ challenge [19] . With respect to these facts it is commonly accepted that new drugs are required urgently . To this end research on signal transduction processes in S . mansoni has opened new perspectives . Protein kinases ( PKs ) are highly conserved signal transduction molecules in the animal kingdom and known to be involved in diverse biological processes such as cell growth and differentiation [20] . Thus PK deregulation can lead to cancer development [21]–[22] . This prompted the search for inhibitors , and meanwhile a number of anticancer drugs targeting PKs are approved for use in humans [21]–[24] . Different studies elucidating principles of schistosome development have shown that PKs play important roles during parasite development [25]–[30] . Due to this , and to the fact that schistosomes can be kept in culture , providing access to adults ex vivo , several studies were conducted to investigate whether anticancer drugs would negatively affect PK-controlled processes in schistosomes and cause phenotypic consequences in adults . Indeed , targeting a variety of different kinases using different drugs with PK-inhibiting activities not only showed a negative influence on the reproductive biology of parasites but also remarkable effects on other physiological processes and/or survival in vitro [29]–[37] . Among the PKs studied in more detail were protein tyrosine kinases ( PTKs ) such as Abl kinases , three of which exist in S . mansoni . SmAbl1 and SmAbl2 exhibit high sequence similarities to conserved Abl-kinases [38] , whereas SmTK6 revealed a Src/Abl hybrid character that was confirmed by structural and biochemical studies [39] . Imatinib was used as an inhibitor to analyze Abl-PK activities in adult S . mansoni [38] , [39] . Also known as Glivec ( Gleevec; Novartis ) Imatinib is a small-molecule inhibitor acting as a competitive antagonist of adenosine triphosphate ( ATP ) binding to Abl-PK , which is applied successfully in human cancer therapy [40] . Structural analyses revealed that the S . mansoni Abl-PKs possess the majority of amino acid residues known from studies with the human Abl-kinase to interact with Imatinib [38] , [41] . Furthermore , the Xenopus oocyte system was shown to be suitable to test the catalytic activity of schistosome tyrosine kinases ( TKs ) [31] , [39] . Thus it was demonstrated that SmAbl1-TK , SmTK6-TK , and SmTK3-TK were able to induce 100% germinal vesicle breakdown ( GVBD ) [39] . Using competitive GVBD assays it was further demonstrated that Imatinib negatively influenced the kinase activities of SmAbl1-TK ( 0% GVBD at 1 µM ) and SmTK6-TK ( 0% GVBD at 100 µM ) . Although the latter required a 100-fold higher concentration compared to SmAbl1-TK , this was explained by the unusual Src/Abl hybrid character of SmTK6 . Herbimycin A ( Herb A ) , a Src-TK inhibitor , was not able to fully reduce the GVBD-inducing activity of SmAbl1 ( 60% GVDB at 10 µM ) in contrast to SmTK6 , whose activity was fully suppressed at this concentration . The enzymatic activity of SmTK3 , a Src kinase used as control , was fully suppressed by Herb A ( 0% GVBD at 0 . 01 µM ) but revealed nearly no decrease under the influence of Imatinib ( still 90% GVBD at 100 µM ) , confirming the specificity of these inhibitors [39] . Treating adult schistosomes with Imatinib in vitro led to dose- and time-dependent effects such as reduced pairing stability , the occurrence of bulges and swellings along the body and , finally , the death of the worms . Microscopic analyses showed not only morphological changes within the gonads of both genders , which appeared disordered , defective in differentiation , and in part apoptotic , but also a detachment and degradation of the gastrodermis . Its complete collapse explained the observed death of the parasites [38] . In a follow-up study it was shown that Dasatinib and Nilotinib , second generation Abl-PK inhibitors of high selectivity , were less effective compared to Imatinib in causing severe or even lethal effects on adult schistosomes in vitro . Since Dasatinib and Nilotinib were designed to be even more specific for mutated forms of the human Abl-PK , it was concluded that the more specialized the inhibitor for the human kinase is , the more efficacy for the schistosome kinase gets lost [42] . Because in situ hybridizations detected various regions within adult S . mansoni where these Abl- and Abl-like PKs of S . mansoni were transcribed [36] , [38] , pointing to different physiological functions these kinases may be involved in , we were interested in investigating the effects of Imatinib at the gene transcription level in treated schistosomes . Besides confirming that Imatinib also affected SmAbl2 kinase activity in a heterologous test system , a transcriptome study was performed by microarray analyses and qRT-PCR verification experiments . Strong evidence was obtained that gene transcription is widely influenced supporting the view that a variety of physiological processes have been affected by this Abl-PK inhibitor . This is in line with previous hypotheses suggesting that PKs , due to their pleiotropic and fundamental roles for schistosome biology , are substantiated targets for novel strategies to treat schistosomiasis [42] , [43] . Furthermore , first evidence was obtained that Abl-kinase activities could be part of/or associated with transforming growth factor β ( TGFβ ) signaling in schistosomes .
Experiments with hamsters were performed in accordance with the European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purposes ( ETS No 123; revised Appendix A ) and were approved by the Regional Council ( Regierungspraesidium ) Giessen ( V54-19 c 20/15 c GI 18/10 ) . Biomphalaria glabrata as intermediate snail-host , and hamsters ( Mesocricetus auratus ) as final host were used to maintain the parasite cycle of S . mansoni [44] . Adult worms were isolated from hamsters by hepatoportal perfusion 42 days post infection . Chemical Computing Group's MOE 2011 ( http://www . chemcomp . com/ ) molecular modeling suite was used for the homology modeling . The highly conserved catalytic tyrosine domains [38] of human Abl1 ( 2HYY ) and Abl2 ( 3GVU ) were used as templates for the model building of SmAbl1 and SmAbl2 , respectively . The two homologous sequences were aligned using MOE's kinase constraints , and the models were built using the Amber99 force field with R-Field solvation . Crystallographic water molecules in an H-bond network with the ligand Imatinib and the ligand itself were used as “Environment for Induced Fit” during model building . Intermediates were refined to medium using the GB/VI scoring , while the final model was refined to Fine with an RMS Gradient of 0 . 5 . Molecular Docking was carried out using Cambridge Crystallographic Data Centre's GOLD suite 2 . 5 ( http://www . ccdc . cam . ac . uk/ ) . Imatinib was sketched in 2D and converted to 3D using Molecular Networks' CORINA ( http://www . molecular-networks . com/ ) . Docking was performed using the Chemscore scoring function with kinase parameters , and the binding site was defined using the position of the Imatinib ligand as modeled in the SmAbl1 and SmAbl2 homology models . Parameters for the genetic algorithm were set to auto . Water molecules used for the model building were allowed to participate in the docking . Sequences of the tyrosine kinase ( TK ) domains of SmAbl1 , SmAbl2 , SmTK6 , and SmTK3 were obtained by PCR amplification ( primers used are given in Supplementary data S1 ) and cloned into the plasmid pcDNA3 . 1B ( Invitrogen ) , which contained a T7 promoter for in vitro transcription . The resulting constructs SmAbl1-TK , SmAbl2-TK , SmTK6-TK , and SmTK3-TK were sequenced confirming intact open reading frames ( ORFs ) . The plasmid constructs were linearized by PmeI and cRNAs were generated using the T7 mMessage mMachine Kit ( Ambion , USA ) . This way capped messenger RNAs ( cRNAs ) were synthesized in vitro and analyzed as previously described [31] , [38] , before they were injected into stage VI oocytes of Xenopus laevis . To this end each oocyte was injected with 60 ng cRNA in the equatorial plane , followed by incubation in ND96 medium at 19°C . GVBD ( germinal vesicle breakdown ) activity was determined according to the appearance of a white spot at the animal pole 18 h following injections . As shown before , this system was used successfully to monitor schistosome kinase activities [31] , [38] , [39] . Here it was used to investigate kinase activities under the influence of Imatinib ( Enzo Life Sciences; 170 mM stock solution in water ) , an Abl-kinase inhibitor , or the Src kinase inhibitor Herbimycin A ( Herb A ) as control ( Tocris Bioscience; 10 mM stock solution in DMSO ) . Pools of 10 oocytes each were injected with SmAbl1-TK , SmAbl2-TK , SmTK3-TK , or SmTK6-TK cRNA and placed in ND96 containing different concentrations of Imatinib ( 0 . 01 µM to 100 µM final ) or Herb A ( 0 . 0001 µM to 10 µM final ) . As negative control , non-injected oocytes were used . As positive controls , oocytes were incubated with the natural hormonal stimulus progesterone leading to 100% GVBD without further manipulation [45] . Perfusion was done with M199 medium ( Gibco ) . Paired adult worms were collected using fine tweezers and washed with M199 medium ( 2x ) . Subsequently , they were maintained in culture in M199 supplemented with FCS ( Gibco; 10% ) , HEPES ( Sigma; 1 M , 1% ) , and antibiotic/antimycotic mixture ( Sigma; 1% ) at 37°C and 5% CO2 [36] , [38] . Inhibitor treatment was performed for 24 h or 48 h with 50 µM Imatinib ( Imatinib mesylate , C29H31N7O·CH3SO3H , dissolved in water; Enzo Life Sciences ) as previously described [38] . Control couples were kept in culture for 24 h or 48 h without inhibitor addition but otherwise treated using the same conditions . During the treatment periods , pairing stability and vitality were checked regularly . We defined pairing stability of couples when males kept their female partners within the gynecophoral canal while being sucked with their ventral suckers to the Petri dish . When couples separated , or when males stopped sucking to the Petri dish and/or lay on the side ( a sign of decreasing vitality ) , the appropriate worms were not used for experiments and removed . After completion of treatment , the couples ( inhibitor-treated and control ) were carefully separated using featherweight tweezers , immediately shock-frozen in liquid nitrogen , and stored at −80°C . Trizol ( Invitrogen ) was used to extract RNA from treated or control worms ( combined sexes in both cases ) followed by a DNAse digestion ( RNAeasy kit; Qiagen ) . The quality of RNA was checked by microfluidic electrophoresis ( Bioanalyzer; Agilent Technologies ) . For microarray experiments a S . mansoni custom-designed oligonucleotide platform ( 60-mers ) was used containing 44 , 000 probes representing nearly the complete S . mansoni and S . japonicum transcriptome ( Agilent Technologies; [46] , [47] ) . All associated information ( probes , annotation ) is available at Gene Expression Omnibus ( GEO ) under the accession number GPL8606 . For the microarray experiments RNA from treated and control males and females ( 300 ng each for biological replicates ) was used for cDNA amplification followed by Cy3 and Cy5 labelling during in vitro transcription ( Quick Amp Labelling Kit , two colors; Agilent Technologies ) . Dye-swap approaches were done as internal technical replications for each sample . Three microarray hybridizations were performed for each sample ( inhibitor treatment for 24 and 48 h as well as control ) , which included two technical replicates for each of the three biological replicates . As probe for hybridization , 825 ng cRNA of each labelled inhibitor sample was used and combined with a control sample labelled with the opposite dye . Hybridization was done at 65°C for 17 h with rotation followed by slide washing ( according to the Agilent manual ) and scanning ( Gene Pix 4000B Scanner; Molecular Devices ) . Raw data were acquired using the Feature Extraction software ( Agilent Technologies ) . They are available under GEO study number GSE53154 . For subsequent data analyses , genes were considered as transcribed only if the corresponding probe had a signal significantly higher than background ( using default parameters from the Feature Extraction software and considering the “IsPosAndSig” result from the output ) . In addition , signals of a probe had to occur in at least 75% of all replicates and in at least one of the two conditions ( inhibitor-treated or control ) independent of the length of cultivation ( 24/48 h treatment versus 24/48 h untreated control ) . The quality of the microarray expression data was assessed by the overall Pearson correlation among technical replicates , which was found to be in the range of 0 . 93 to 0 . 99 ( average 0 . 98 ) . LOWESS algorithm was used for normalisation of intensities [48] , and the log2ratios were calculated between inhibitor-treated and control groups . Finally , the filtered data were analysed on the basis of the updated genome annotation to eliminate redundancy of the probes per gene [49] , [50] . Inspection of box plots revealed that intensities from both dye channels of all technical and biological replicates were in a similar range , showing that no additional normalization steps were necessary . SAM ( Significance Analysis of Microarrays ) was used [51] to detect genes with a significant change in transcript level . Data sets for the two treatment periods ( 24 h and 48 h ) were analysed by one-class analysis , in which transcripts were evaluated that showed the same direction of regulation for both time points ( sustained regulation direction ) . Here , genes with a q-value ≤0 . 01 were considered as significantly differentially transcribed between the inhibitor-treated and the control worm populations comprising protein-coding genes and putative antisense-oriented oligonucleotide probes ( labelled as “to be used in analysis = YES” in the updated annotation of the array [49] , [50] ) . Spotfire was used for hierarchical clustering [52] . For first functional analyses of differentially transcribed protein-coding genes , Gene Ontology ( GO ) enrichment analysis was performed [53] using the software tool Ontologizer [54] . Parent-child union [55] was used to detect categories containing enriched genes , and the p-value was adjusted according to Benjamini-Hochberg ( BH ) correction [56] . For identifying potential candidates for further analyses and their putative involvement in hypothesized pathways , Ingenuity Pathway Analyses ( IPA; http://www . ingenuity . com; [57] ) were performed in addition , as described before [47] . IPA provides curated information from the literature for human , mouse and rat models about canonical pathways , regulated molecular networks , including signal transduction cascades ( of which some are involved in human cancer or other diseases ) , and regulated transcription factors and their putative targets . To use this tool all S . mansoni genes were annotated with the corresponding human homolog and uploaded to IPA along with their corresponding microarray transcription measurements . The validity of the obtained results of qRT-PCRs ( see below ) and microarrays was determined by Spearman's rank correlation coefficient ( rs ) [58] . The reliability of the transcriptional changes detected by microarray analyses was tested by quantitative RT-PCR ( qRT-PCR ) analyses of a number of genes . RNAs of inhibitor-treated or control couples were isolated by TriFast ( PeqLab ) , and 1 µg each was reverse transcribed ( QuantiTect Reverse Transcription kit; Qiagen ) . Following cDNA dilution 1∶20 , qRT-PCRs were done using Rotor Gene Q ( Qiagen ) . Amplification rates were determined by SYBRGreen incorporation ( PerfeCTa SYBR Green Super Mix; Quanta ) . Melting point analyses were done to distinguish between the specific amplification product and unspecific primer-dimer formation following each qRT-PCR analysis . For primer design , the software Primer 3 Plus was used ( http://www . bioinformatics . nl/cgi-bin/primer3plus/primer3plus . cgi ) . The expected amplification products were between 140–160 bp in size . Primers were designed flanking predicted introns to be able to differentiate between cDNA and genomic DNA , and melting points were between 59°C–62°C depending on sequence composition . A list of all primers used , which were commercially synthesised by Biolegio ( Netherlands ) , is shown in Supplementary data S1 . Standard reference genes normally used for relative quantification analyses such as α-tubulin , actin , Cu/Zn SOD ( superoxide dismutase ) , or histone showed regulation following inhibitor treatment . Therefore , absolute quantification was performed on the basis of standard curves generated by purified PCR products ( used in dilution series ) [59] . Fold changes are given where appropriate . As a basis for comparing microarray and qRT-PCR results , log2ratios ( treated/control ) were calculated as described before [47] , [60] . The efficiency of each qRT-PCR was evaluated to be between 90–100% . Spearman's rank correlation coefficient ( rs ) was assessed to validate the ratios obtained from qRT-PCRs and microarrays [58] , [61] , [62] . In addition to those already mentioned , the following public domain tools were used: SchistoDB ( http://www . schistodb . net; [63] ) , BLASTx ( http://www . ncbi . nlm . nih . gov/BLAST ) , the Welcome Trust Sanger Institute S . mansoni OmniBlast ( http://www . sanger . ac . uk/cgi-bin/blast/submitblast/s_mansoni/omni ) , BLAST ( http://blast . ncbi . nlm . nih . gov/ ) , and Gene Cards , which is a database of human genes providing concise genome-related information on all known and predicted human genes , to authenticate IPA-identified gene acronyms ( http://www . genecards . org ) .
On the basis of their human counterparts , homology models of the S . mansoni Abl kinase 1 and 2 were created which corresponded well with the protein template structures 2HYY ( human Abl 1 ) and 3GVU ( human Abl 2 ) . The ten highest scoring docking poses of Imatinib in the homology model of SmAbl2 were found in good structural agreement with the crystal structure pose of Imatinib in the human Abl2 crystal structure ( Figure 1 A; SmAbl1 data not shown ) . The highest scoring docking pose is virtually identical to the crystal structure pose ( Figure 1 B ) . While human Abl2 forms seven directed interactions with Imatinib ( Figure 2 ) , for the SmAbl2 homology model four directed interactions were detected ( Figure 1 C , Figure 2 ) . Two out of the four SmAbl2 interactions are shared with the human Abl2 interactions ( Figure 2 ) . For SmAbl1 the situation was similar . Key residues involved in direct interactions with Imatinib in the human Abl proteins were found to be conserved for all four protein sequences . Four out of the six directed interactions were also detected for the SmAbl1 homology model . In contrast to human Abl 1 , one of these residues ( D568 ) did not interact directly with Imatinib; however , it did via an H-bond network involving a water molecule . Since we docked Imatinib to homology models , the recovery of individual directed interactions should not be overstated . Although we docked Imatinib to homology models , the data clearly indicated that Imatinib is able to bind both SmAbl1 and SmAbl2 . An inhibitor swap-like approach [38] was used to test the enzymatic activity of SmAbl2 , for its susceptibility towards Imatinib and Herb A . To this end , cRNA encoding the TK domain of SmAbl2 ( SmAbl2-TK ) was injected into Xenopus oocytes under selection conditions using different inhibitor concentrations . GVDB was monitored as read out , and the results compared to SmAbl1 , SmTK6 and SmTK3 [38] . Under Herb A selection , SmAbl2-TK induced 100% GVBD at 1 µM and still 90% GVBD at 10 µM ( Figure 3 ) . The Src-TK SmTK3 and the Src/Abl-hybrid TK SmTK6 were completely inhibited by Herb A inducing GVBD at concentrations of 0 . 01 µM ( SmTK3 ) , or 10 µM ( SmTK6 ) [39] . Using Imatinib , however , SmAbl2-TK enzymatic activity was reduced to 70% GVBD-inducing capacity at 0 . 01 µM and completely suppressed GVBD at 0 . 1 µM . At the latter concentration still 90% GVBD was observed for SmAbl1-TK , whose activity was completely suppressed using 1 µM Imatinib ( Figure 3 ) . These results showed that Imatinib effectively inhibits both Abl kinases of S . mansoni , which is supported by the modeling data presented above . Based on our previous findings of remarkable effects of Imatinib on morphology , physiology , and survival of adult S . mansoni in vitro [38] , [42] , we focused on the elucidation of molecular effects induced by this inhibitor . To this end , a large-scale transcriptional analysis was performed using a microarray platform representing nearly the complete S . mansoni and S . japonicum transcriptomes [46] , [47] . Since in previous experiments treatment with 50 µM Imatinib showed slight effects after 24 h and strong effects after 48 h treatment [38] , we anticipated that these treatment profiles represented starting ( 24 h ) and peak points ( 48 h ) of the effects induced , thus being interesting for analysis . Therefore , the 50 µM concentration and both time points were chosen for comparative transcriptomics . Based on the results of the GVBD assays it was anticipated that 50 µM Imatinib would not induce potential off-target effects through co-influencing Src kinases such as SmTK3 since its inhibition will not occur using this inhibitor concentration [39; this study] . With this set-up , the expectation was to find genes differing in their transcript levels under inhibitor influence . It was hypothesized that transcript levels of genes strongly regulated by signaling pathways including SmAbl1/2-kinases would show a continuous tendency of regulation during the treatment period of 48 h representing sustained transcriptional changes . Following microarray hybridization and data evaluation a one-class statistical analysis was performed that revealed sustained transcriptional changes of 1429 significantly differentially transcribed genes which were up-regulated following Imatinib-treatment . Of these , 1094 were protein-coding genes ( Supplementary data S2 , S3 ) . The remaining transcripts represented antisense RNAs , intronic and UTR sequences . Among the protein-coding genes were candidates coding for serine/threonine PKs , PTKs , proteins with female-preferential or -specific functions such as egg synthesis , transcription factors , muscle-associated proteins , small GTPases , heat-shock proteins , and signal transduction/associated proteins ( Supplementary data S4 ) . Furthermore , 939 protein-coding genes were found to be significantly down-regulated . These genes potentially code for cathepsins , lipoproteins , VAL ( venome allergen-like ) proteins , glutamate receptors and further transporters , the gynecophoral canal protein ( GCP ) , motor and/or muscle proteins , drug efflux proteins , transmembrane receptors , calmodulin and other calcium binding proteins , histones , spermatogenesis-/testis-associated proteins , a morphogen-binding protein , signal transduction ( -associated ) proteins , and a cell adhesion protein ( Supplementary data S5 ) . GO analyses of differentially transcribed genes revealed ontology categories enriched with genes being up- or down-regulated ( BH adjusted p-value ≤0 . 05; threshold = 0 . 1 ) . Examples of GO categories represented in the up-regulated genes were: gene expression and transcription ( Biological process ) , myosin complex ( Cellular compenent ) , kinase activity and transcription factor activity ( Molecular Function ) ( Supplementary data S4 , S6 ) . Within the data set of down-regulated genes , GO categories were found for enriched genes coding for functions such as e . g . transmembrane transport or cell surface receptor-linked signalling pathways ( Biological process ) , membrane and microtubule-cytoskeleton/associated complex ( Cellular component ) , and signal transducer activity , transporter activity or cysteine-type endopeptidase activity ( Molecular function ) ( Supplementary data S5 , S7 ) . Using IPA , the following five networks were identified , which were enriched with proteins coded by differentially transcribed genes involved in the following functions: ( 1 ) Protein Degradation , Protein Synthesis , Tumor Morphology ( adjusted p-value 10−80 ) , ( 2 ) RNA Post-Transcriptional Modification , DNA Replication , Recombination , and Repair , Cellular Assembly and Organization ( adjusted p-value 10−69 ) , ( 3 ) Post-Translational Modification , Protein Folding , Carbohydrate Metabolism ( adjusted p-value 10−60 ) , ( 4 ) Developmental Disorder , Gene Expression , Genetic Disorder ( adjusted p-value 10−40 ) , and ( 5 ) Carbohydrate Metabolism , Drug Metabolism , Lipid Metabolism ( adjusted p-value 10−39 ) ( Supplementary data S8 ) . Among the molecules with the largest fold-changes of transcription was a potential pseudo-glutamine synthetase ( LGSN ) strongly ( about 9-fold ) up-regulated , which in the human system is reported to have a chaperone function for the reorganization of intermediate filaments acting as a component of the cytoskeleton [64] . Amongst the strongly down-regulated ( about 8 to 11-fold ) transcripts were those encoding peptidases such as cathepsins ( CatK , CatS , CatL ) , which are members of the peptidase C1 protein family and known in humans to participate in protein processing during immunological processes and several disease-associated pathologies [65] , [66] . In summary , a number of processes were highlighted by both GO analyses and IPA that were influenced by inhibitor treatment pointing to candidate genes such as cathepsins for further analyses . The selection of candidates for qRT-PCR experiments to verify differential transcription was based on GO and IPA results , but also on literature studies including the Imatinib-induced phenotypes in adults obtained previously ( negative effects on pairing-stability , oogenesis and spermatogenesis , integrity of the gastrodermis , and locomotion [38] , [42] ) . Since GO and IPA analyses indicated influences of Imatinib treatment on endopeptidase activity and cathepsins , respectively , we chose cathepsin K ( Smp_139240 ) and cathepsin B ( Smp_085180 ) . The latter was already shown to be active in the gut [67] . This applies also to the selected hemoglobinase ( Smp_075800 ) , which was localized to the gut [68] . Venom allergen-like proteins ( VALs ) of platyhelminths are members of the SCP/TAPS ( Sperm-Coating Protein/Tpx-1/Ag5/PR-1/Sc7 ) protein superfamily and hypothesized to play not only roles in spermatogenesis but also beyond [69] , which led to the choice of VAL7 ( Smp_070240 ) . The metabotropic glutamate receptor ( Smp_128940; [70] ) was selected as a representative for cell surface receptors and its potential role in the nervous system of adult male and female schistosomes [71] . Finally , GCP was included due to its hypothesized role in male-female interaction [72] . The results of the qRT-PCR experiments , which were performed with the RNA of schistosome couples following 48 h Imatinib treatment , confirmed in each case the down-regulation of these transcripts ( Figure 4 ) . Since the GO analysis of up-regulated genes pointed to muscular activities ( myosin complex within the ontology cellular component ) but also to signal transduction ( kinase activity within the ontology molecular function ) , further candidates were selected for qRT-PCR . Among these were paramyosin ( Smp_129550 ) and titin ( Smp_105020 ) , both proteins involved in muscle activity [73] , [74] , and a ribosomal protein S6 kinase ( Smp_017900 ) due to its potential role as a MAPK-activated PK in signalling processes controlling diverse processes including survival [75] , [76] . Furthermore , HSP70 ( Smp_106930 ) was chosen due to its known roles in stress response and signal transduction processes , but also as egg-shell component in schistosomes [77]–[79] . The results of the qRT-PCR analyses confirmed up-regulation in each case ( Figure 5 ) . Unexpectedly , a manual data screen indicated that a number of egg production-related genes were significantly up-regulated following Imatinib treatment such as the egg-shell precursor proteins p14 ( Smp_131110; [80] ) , fs800-like ( Smp_000270; [79] ) , a predicted egg-shell precursor protein ( Smp-000430; [47] ) , and the eggshell protein cross-linking tyrosinase SmTYR1 ( Smp_050270; [81] ) . Also for these genes , qRT-PCR confirmed up-regulation following Imatinib treatment ( Figure 5 ) . The results obtained for the qRT-PCR analyses of all studied genes significantly correlated to the microarray data according to Spearman's Correlation Coefficient ( rs = 0 . 784 , p<0 . 001; [58] ) . Extended analyses indicated that Imatinib treatment can lead to a sustained effect on specific genes . By qRT-PCR analyses of RNA of couples treated with Imatinib for 24 h or 48 h , the amounts of gene transcripts increased or decreased over time as demonstrated exemplary by the analyses of three genes . Compared to 24 h treatment higher transcript levels were determined for the ribosomal S6 kinase after 48 h ( Supplementary data S9 ) , whereas transcript levels declined for hemoglobinase and GCP from 24 h to 48 h ( Supplementary data S10 , S11 ) . Since the couples used for these analyses were separated before freezing , we checked whether pairing had an influence on the transcription of the GCP gene , which was hypothesized before to be a target of a TGFβ-pathway but also a male factor contributing to pairing-dependent female maturation [72] , [82] , [83] . A qRT-PCR analysis using actin as reference gene showed that the status of pairing had no significant influence on the GCP transcript level , since there was no significant difference in transcript levels comparing males with and without pairing experience or males separated from their female partners ( Supplementary data S12 ) . This finding is also supported by results of a recent study comparing the transcriptomes of pairing-experienced males versus naïve males using microarrays , SuperSAGE and also qRT-PCR , in which no evidence for an influence of pairing on GCP expression was found [84] . Thus the decrease of the GCP transcript level following Imatinib treatment represented an inhibitor-specific effect . Recent years have provided compelling evidence for a prominent role of TGFβ signalling in schistosome biology [27] , [47] , [84] , [85] , [86] . Results of a previous study suggested GCP to be part of TGFβ signalling pathways [82] . Furthermore , the effect of a specific TβRI kinase-inhibitor ( TRIKI ) was investigated in schistosomes . In vitro-culture experiments with couples provided first evidence for a role of the TGFβ pathway during the regulation of mitotic activity and egg production [26] . Subsequently , it was shown by microarray analysis using the same technical platform that genes contributing to these processes , such as egg shell-forming genes , were slightly up-regulated upon TRIKI treatment [47] . Analysing the microarray data following Imatinib treatment in the present study we observed that a number of specific genes were differentially regulated that had shown up in the previous analysis as well . Since there is evidence from the literature that Abl-kinases can be part of TGFβ signalling pathways [87] , [88] we investigated whether this may apply for schistosomes as well and compared both data sets in a merging analysis . To this end the Imatinib data set of this study and the TRIKI data set of the previous microarray study [47] were compared using a significance q-value of ≤0 . 05 to identify a comprehensive set of differentially transcribed genes found in common in the two conditions . This approach identified 6754 differentially transcribed protein-coding genes in total , of which 1800 were common in both data sets . The merging analyses finally indicated that out of 2339 genes found in this study to be down-regulated upon Imatinib treatment , 900 matched those differentially ( up- and down- ) regulated upon TRIKI treatment . Out of these 480 were up-regulated and 420 down-regulated by TRIKI . Furthermore , out of 2616 genes found in this study to be up-regulated upon Imatinib treatment , 900 corresponded to those differentially regulated by TRIKI . Out of these 822 were up-regulated and 78 down-regulated by TRIKI . By definition , no gene was found within the intersection of Imatinib up- and down-regulated genes ( Figure 6; Supplementary data S13 , S14 ) .
Based on our previous results on the physiology and morphology of adult schistosomes treated by Imatinib [38] , our present study aimed at identifying transcriptional processes influenced by this inhibitor . To confirm that Imatinib targets not only the Abl kinase SmAbl1 as shown by biochemical analyses before [39] , we investigated its inhibitor effect on SmAbl2 . SmAbl1 and SmAbl2 are the only true Abl-kinases present within the genome of S . mansoni [50] in contrast to SmTK6 which represents a Src/Abl hybrid kinase being less susceptible to Imatinib [39; this study] . By competitive GVBD assays in Xenopus oocytes expressing these kinases we determined specific effects of Imatinib on both Abl1 and Abl2 kinases . Although their susceptibilities differed by a factor of 10 , the results obtained in this and the previous study [39] clearly confirmed their target roles , but also that Src-like and true Src kinases such as SmTK3 are not affected at the Abl-effective concentrations of Imatinib , which reduced the probability of off-target effects . These experimental data were well supported by the modeling and docking data generated confirming that Imatinib is able to bind to both schistosome Abl kinases . With regard to results of the in vitro study showing increasing physiological and morphological effects between 24 h and 48 h treatment using 50 µM Imatinib [38] , we performed transcriptional profiling for these time-points to get a broader view of molecular processes potentially affected by Imatinib . Microarray and subsequent bioinformatics analyses revealed a broad spectrum of genes being differentially regulated following inhibitor treatment . The transcription of genes involved in male-female interaction , gut physiology , muscle activities , and egg production were among those highlighted by the analyses . For qRT-PCR verification a number of genes were selected with regard to GO and IPA results but also to the phenotypes obtained in the preceding in vitro study , which comprised reduced pairing stability ( i ) , reduced sizes of the gonads of both genders combined with disturbed spermatocyte/oocyte differentiation ( ii ) , a degradation of the gastrodermis ( iii ) , and tremor-like movements pointing to altered locomotion activity ( iv ) [38] . In each case qRT-PCR and microarray results correspondingly showed sustained transcriptional changes and reduced transcript levels in each case for ( i ) GCP , a fascicle I-like cell adhesion molecule hypothesized to be involved in male-female interaction [72] , [83] , ( ii ) VAL-7 , a member of the sperm-coating protein SCP/TAPS superfamily [89] , [90] that was found in S . mansoni to be expressed in the esophageal gland in larvae , and adult males and females [91] , ( iii ) a hemoglobinase and the cathepsins B and S , of which hemoglobinase and cathepsin B were already shown to be active in the gut [67] , [92]–[94] , ( iv ) as well as a metabotropic glutamate receptor . Interestingly , previous studies indicated that Abl kinases regulated lysosome functions , especially autophagy by organizing the localization and activity of lysosomes , glycosidases and cathepsins , suggesting that Abl is involved in processes regulating digestion and removal of self- and foreign material [95] , [96] . Metabotropic glutamate receptors have been discussed in the context of seizure-like behavior , defined as paroxysms resulting in disruption of normal locomotor-system activity in planaria [97] . Whether this tremor-like phenotype in schistosomes ( iv ) is also accompanied by higher transcript levels detected for the muscle protein genes paramyosin and titin , of which the latter determines muscle elasticity , stability , and contraction velocity [98] , remains unclear at this stage . In contrast , the up-regulation of the stress protein gene HSP70 following inhibitor treatment meets the expectations as well as higher transcript levels of the ribosomal S6 kinase , a signaling molecule involved in cell growth , proliferation , but also survival [99] . Thus the differential regulation of these genes corresponded well to the phenotypes observed in vitro . Furthermore , a comparison of both time-points used for the analysis ( 24 h/48 h ) showed strong and sustained transcriptional changes by microarrays but also on the basis of qRT-PCRs of selected candidate genes , which indicated an enduring influence of the Abl kinase pathway on these genes . Compared to transcriptome studies in S . mansoni or S . japonicum after exposure to PZQ in vitro [100] , [101] or in vivo [102] , a number of differences can be noted that indicate dissimilar processes affected by this established drug and Imatinib . Using a recent in vivo model of S . japonicum PZQ led to an up-regulation among others of genes associated with muscle function , lipid and ion regulation , and drug resistance in treated males [102] . In females , fewer genes seemed to be affected ( up-regulated ) , examples are involved in pathogen defense , general detoxification , drug resistance and calcium regulation [102] . Similar findings were made in in vitro studies with adult S . mansoni showing that genes encoding multiple drug transporter as well as calcium regulation , stress and apoptosis-related proteins were up-regulated [101] . In contrast to these findings we observed a down-regulation of genes coding for lipoproteins , motor and/or muscle proteins , drug efflux proteins , calmodulin and other calcium binding proteins in S . mansoni couples following Imatinib treatment . With respect to apoptosis-related genes the picture is puzzling since there much variation within all PZQ and Imatinib data sets and little correspondence among these different data . This justifies no precise conclusion on the participation of defined apoptosis-related signaling processes in the primary effects on schistosomes caused by these drugs . Surprisingly , we also identified genes to be up-regulated that contribute to egg formation such as p14 , fs800-like , a predicted egg-shell precursor protein gene , and tyrosinase 1 , a gene involved in final egg-shell synthesis [79] , [81] . This was unexpected since we observed reduced egg production in Imatinib-treated schistosome couples . However , egg production is a complex process and may be influenced by further genes of which some , yet unknown to be important for this process , may be down-regulated by Imatinib , while the known egg-formation genes might be up-regulated to compensate for the overall reduced egg output in Imatinib-treated schistosomes . Conspicuously , the higher transcript levels of these egg formation-related genes resembled the results obtained in a recent microarray study where the effect of TRIKI , a TGFβRI-kinase inhibitor , was investigated on transcriptional profiles in adult schistosome couples in vitro . TRIKI led to an increase of transcript levels of the same egg formation-related genes in paired females in contrast to Herb A , a Src kinase inhibitor , which reduced the transcript levels of these genes . From this it was concluded that a TGFβ and a Src kinase pathway cooperatively control egg formation processes in a balanced manner in schistosomes assigning repressing ( TGFβ/TGFβRI-pathway ) and inducing ( Src-pathway ) tasks [47] . This and the finding of GCP and egg formation-associated genes as common target molecules of TGFβ [82] as well as SmAbl1/2-influenced molecular processes [this study] , prompted us to evaluate the TRIKI-related against the Imatinib-related microarray data sets . Comparing the total amounts of differentially transcribed genes about 27% ( 1800 out of 6754 ) were present in both data sets , of which about 50% ( 900 or 900 out of 1800 ) were differentially regulated and about 70% ( 420 and 822 out of 1800 ) in the same direction . Thus many genes significantly differentially transcribed upon TRIKI- and Imatinib treatment overlapped . This clearly indicates a potential association of TGFβRI-mediated and Abl kinase-containing pathways in schistosomes . Beyond the fact that egg formation-associated genes such as fs800-like , p14 , egg shell precursor , and tyrosinase 1 [47; this study] as well as GCP became noticeable as common targets , SmAbl1/2 transcripts and TGFβRI-transcripts were found in the same tissues by in situ hybridization , mainly in the gonads [38] , [103] . In conclusion , it appears very likely that the schistosome Abl kinase ( s ) are among other possibilities members of signalling pathways induced by TGFβ . Such a molecular connection has been shown before , demonstrating c-Abl as a Smad-independent component of TGFβ signaling pathways and mediator of TGFβ-driven proliferation in human fibroblasts [87] , [104] , [105] . Our previous in vitro studies exhibited strong effects of Imatinib on schistosome morphology , physiology and survival in vitro suggesting that this compound may be one of the candidates for the design of alternative strategies to fight schistosomiasis [38] , [42] , [43] . This was confirmed by an independent approach recently , which reproduced similar phenotypes in vitro , although a first in vivo experiment failed [106] . Nonetheless , the data obtained in this study support the conclusion that Imatinib exerts broad negative effects on worm physiology substantiating the hypothesized role of PKs as potential targets [25] , [28] , [36] , [42] , [43] . In this respect it was encouraging to note that according to our microarray analysis also multidrug resistance ( MDR ) genes ( Smp_089200 , Smp_151290 ) were among the significantly down-regulated genes following Imatinib treatment . Thus they may represent additional targets of Abl-kinase-containing pathways . Future studies could also aim at analyzing the molecular networks controlling the expression of such MDR genes and their substrate specificities in more detail . Depending on the substrates transported by these MDRs , and with respect to treatment strategy and efficacy , the suppression of MDR genes as an additional consequence of inhibitor application would represent a potential side effect that is most welcome .
|
Schistosomiasis is an infectious disease caused by schistosome parasites , affecting millions of people worldwide . The pathogenic consequences of schistosomiasis are caused by the eggs inducing severe organ inflammations . Praziquantel is widely used to treat schistosomiasis; however , there is fear of resistance developing . Research in the last decades has provided strong evidence for the importance of protein kinases controlling physiological processes in schistosomes . Two Abl-kinases were discovered , whose activities are blocked by Imatinib , an inhibitor known as Gleevec/Glivec from human cancer therapy . In vitro , Imatinib treatment led to dramatic effects on morphology and physiology and to the death of adult schistosomes . Besides modeling of the schistosome Abl-kinases we investigated the effect of Imatinib on gene expression in adult S . mansoni by performing transcriptomics and discovered a wide influence on the transcription of genes involved in surface- , muscle- , gut- and gonad-associated processes . Comparative in silico analyses with data from a previous study indicated a yet unknown association of TGFβ and Abl-kinase signaling in schistosomes . Among others the gynecophoral canal protein gene GCP was identified as a common target . The data obtained demonstrate a substantial influence of Imatinib on physiological processes in adult schistosomes supporting the role of protein kinases as interesting targets .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"genome",
"expression",
"analysis",
"medicine",
"and",
"health",
"sciences",
"functional",
"genomics",
"animal",
"genetics",
"pathology",
"and",
"laboratory",
"medicine",
"tropical",
"diseases",
"microbiology",
"parasitic",
"diseases",
"parasitology",
"parasite",
"physiology",
"genomic",
"databases",
"genome",
"analysis",
"quantitative",
"parasitology",
"veterinary",
"science",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"veterinary",
"diseases",
"veterinary",
"parasitology",
"gene",
"expression",
"comparative",
"genomics",
"pathogenesis",
"signal",
"transduction",
"computer",
"modeling",
"cell",
"biology",
"gene",
"regulatory",
"networks",
"transcriptome",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"molecular",
"cell",
"biology",
"computational",
"biology"
] |
2014
|
Imatinib Treatment Causes Substantial Transcriptional Changes in Adult Schistosoma mansoni In Vitro Exhibiting Pleiotropic Effects
|
Following injury , axons of the peripheral nervous system have retained the capacity for regeneration . While it is well established that injury signals require molecular motors for their transport from the injury site to the nucleus , whether kinesin and dynein motors play additional roles in peripheral nerve regeneration is not well understood . Here we use genetic mutants of motor proteins in a zebrafish peripheral nerve regeneration model to visualize and define in vivo roles for kinesin and dynein . We find that both kinesin-1 and dynein are required for zebrafish peripheral nerve regeneration . While loss of kinesin-1 reduced the overall robustness of axonal regrowth , loss of dynein dramatically impaired axonal regeneration and also reduced injury-induced Schwann cell remodeling . Chimeras between wild type and dynein mutant embryos demonstrate that dynein function in neurons is sufficient to promote axonal regrowth . Finally , by simultaneously monitoring actin and microtubule dynamics in regenerating axons we find that dynein appears dispensable to initiate axonal regrowth , but is critical to stabilize microtubules , thereby sustaining axonal regeneration . These results reveal two previously unappreciated roles for dynein during peripheral nerve regeneration , initiating injury induced Schwann cell remodeling and stabilizing axonal microtubules to sustain axonal regrowth .
Axons of the mature peripheral nervous system have retained a remarkable ability for regeneration . Although simple in concept , peripheral nerve regeneration is a complex process that requires extrinsic as well as intrinsic mechanisms . Chief amongst the intracellular mechanisms that contribute to axonal regeneration are microtubule organization and dynamics as well as axonal transport . It has long been known that following injury the pool of dynamic microtubules at the lesion site , as well as axonal transport , increase [1–3] . Given the central role of both microtubule dynamics and axonal transport in promoting axonal regeneration , factors that regulate both processes are prime candidates for regulating peripheral nerve regeneration . The molecular motor proteins kinesin-1 and dynein are key regulators of both microtubule organization and axonal transport and have both been implicated in peripheral nerve regeneration . Kinesin-1 is an anterograde motor that is essential for maintaining neuronal homeostasis by transporting cargos , including organelles and mRNA , from the cell body toward synaptic terminals . Kinesin-1 has also been shown to drive axonal outgrowth during development and after injury [4 , 5] . Dynein has similarly been studied for its role in maintaining homeostasis by transporting cargo , however dynein moves cargo retrogradely towards the cell body . Dynein also plays an important role in axonal injury by trafficking injury signals , including components of JNK and ERK MAPK pathways , which are generated at the lesion site and actively transported to the cell body [6 , 7] . There these injury signals initiate a regenerative response , characterized first by upregulation of regeneration-associated genes that prevent neuronal cell death , and by initiating a genetic program that promotes regrowth of injured axons back to their original targets [8 , 9] . More recently it has become clear that in addition to its role in retrograde transport , dynein also functions in cytoskeletal organization and maintenance . For example , in C . elegans dynein regulates local microtubule dynamics in dendrites to promote microtubule stabilization [10] . Additionally , in the axon dynein transports microtubules to establish and maintain microtubule polarity [11–13] . Finally , besides its preeminent role in axonal homeostasis , dynein is also required for Schwann cell development and myelination [14] . Yet despite dynein’s well documented roles in both axons and glial cells , the effects of dynein on the cellular behaviors of regenerating axons and their associated glial cells in intact animals have not been examined . In order to examine the diverse cellular functions of molecular motors in multiple cell types , we combined genetic mutants with live imaging of nerve regeneration in larval zebrafish , as previously described [15] . This allowed us to study the real-time dynamics of regenerating axons and surrounding Schwann cells in a whole organism context . We find that the molecular motors kinesin-1 and dynein , albeit to different degrees , are both required for axonal regrowth in vivo . Furthermore , we find that dynein is also required to initiate injury-induced morphology changes in Schwann cells , however wild type neurons transplanted into otherwise dynein mutant animals are able to regrow robustly , indicating that neuronal dynein is sufficient to promote axonal regrowth . Finally , we find that dynein is dispensable for initiation of axonal regrowth but is required to stabilize microtubules in injured axons to generate persistent , long-range regrowth . These findings elucidate previously unknown roles for dynein in the initiation of injury-induced Schwann cell behaviors , and identify a distinct role for dynein in promoting axonal regeneration through persistent axonal regrowth via microtubule stabilization .
To determine the in vivo roles of molecular motors in peripheral nerve regeneration , we first assessed regeneration in mutants lacking kif5aa , which encodes the neuron-specific Kif5A heavy chain of the conventional anterograde motor Kinesin-1 . We have previously shown that laser mediated transection of motor nerves in larval zebrafish initiates a Schwann cell dependent peripheral nerve regeneration program reminiscent of what is observed in adult vertebrates [16] . Following their complete transection at 5 days post-fertilization ( dpf ) , ventral motor nerves exhibit Schwann cell dependent functional regeneration by 48 hours post-transection ( hpt ) [15] ( Fig 1A and 1B ) . Prior to transection , kif5aa-/- motor nerves were indistinguishable from wild type nerves ( Fig 1C ) . By 48 hpt , motor axons in kif5aa-/- mutants had regrown across the full extent of the ventral myotome , although when compared to wild type siblings the number of fascicles that reached their ventral targets was reduced ( Fig 1D and 1E ) . Using a previously established semi-quantitative scoring index ( for details see Materials and methods and [15] we confirmed that compared to wild type siblings , motor axons in kif5aa mutants exhibited reduced regeneration ( p = 0 . 0487 , Fisher’s exact test ) . We next assessed motor axon regeneration in genetic mutants for the dynein heavy chain gene ( dync1h1 ) which encodes a core component of the retrograde motor dynein . Prior to injury at 5 dpf , dync1h1-/- motor axons exhibit normal architecture , presumably due to the large maternal load sufficient to promote axonal development [17] ( Fig 1F ) . In contrast , following transection , motor axons in dync1h1-/- mutant animals frequently failed to extend beyond the transection site ( Fig 1G , quantified in Fig 1H ) . Analysis of dynein heterozygotes revealed a less severe , although still significant , defect in axonal regrowth , demonstrating a dose-dependent requirement for dynein in promoting axonal regrowth . The severity of the regeneration phenotype we observed in homozygous dync1h1-/- mutants was significantly stronger than that present in kif5aa-/- mutants . This is consistent with the notion that other heavy chains of Kinesin-1 as well as other Kinesin family motors might compensate for the absence of kif5aa [18 , 19] . In contrast , dynein is the sole protein responsible for microtubule-associated retrograde transport , and therefore the regeneration phenotype we observe in dync1h1-/- mutants likely represents a complete block of retrograde transport . We therefore focused on further defining the role of dynein in peripheral nerve regeneration . In addition to its important and well-studied function in neurons , dynein is also required for proper differentiation and myelination of Schwann cells during development [14] . Furthermore , in zebrafish lacking Schwann cells , regenerating axons sprout from the proximal nerve stump but fail to grow across the injury gap [20] , somewhat reminiscent of the phenotype we observe in dynein mutants . Given the importance of Schwann cells for peripheral nerve regeneration and the role of dynein in Schwann cell development , we sought to determine whether dynein is also required for the Schwann cell response to injury , characterized by stereotyped changes in Schwann cell morphology . We have previously shown that before injury , Schwann cell membranes ensheathe individual motor axons , and that following nerve transection when axons fragment , Schwann cell membranes reorganize , changing from a smooth , tube-like appearance to a more rounded and granular morphology [20] , indicative of their transition to an activated , dedifferentiated state , known as the repair cell state that promotes axonal regeneration . Previous studies revealed that in dynein mutants , Schwann cells developmentally arrest at the promyelinating stage [14] . We therefore first wanted to determine whether immature Schwann cells are able to respond appropriately to injury . For this we examined a mutant for the G-protein coupled receptor GPR126 , in which Schwann cells also arrest at the promyelinating stage [21] , similar to what has been reported for dync1h1 mutants . Importantly , in contrast to dynein mutants , gpr126 mutants do not exhibit an obvious deficit in axonal regeneration ( n = 37 nerves from 12 +/+ or +/- wild type larvae and 33 nerves from 11 gpr126-/- mutants , respectively; p>0 . 85 , Fisher’s exact test ) . Analysis of Schwann cells dynamics in gpr126 mutants revealed that Schwann cells respond to injury by extending their membranes dramatically compared to their pre-injury state , indistinguishable from wild type Schwann cells ( Fig 2A–2D ) . This demonstrates that developmentally arrested Schwann cells are still able to respond appropriately to nerve injury . Having determined that promyelinating Schwann cells are competent to respond appropriately to nerve injury , we next examined the behavior of dync1h1-/- mutant Schwann cells . Unlike wild type and gpr126 mutant Schwann cells , we find that following nerve transection dync1h1-/- mutant Schwann cells fail to initiate any morphological changes , and instead retain their pre-injury morphology and membrane position for the duration of the imaging period ( up to five hours ) , arguing against a delay in onset but rather for a complete lack of Schwann cell injury response ( Fig 2E and 2F ) . To quantify this phenotype , we measured the changes in Schwann cell width following nerve transection as a simpler proxy for the complex changes in Schwann cell morphology ( Fig 2G ) . This revealed that while wild type and gpr126-/- Schwann cells significantly increase in width after injury , dync1h1-/- Schwann cells show no significant change . Thus , while dync1h1-/- mutant axons initiate fragmentation following injury , their associated Schwann cells fail to respond , consistent with the idea that dynein is critical for injury-induced Schwann cell remodeling . Our results reveal injury-induced phenotypes in two cell types after injury in dynein mutants , and we therefore wondered whether dynein functions in neurons or Schwann cells to promote axonal regrowth . To determine the cell type in which dynein functions to promote axonal regrowth , we generated chimeras at the blastula stage [22] that contained wild type motor neurons and axons in otherwise dync1h1-/- larvae ( Fig 3A and 3B ) . Control transplantations have previously shown that wild type cells transplanted into wild type embryos generate motor neurons that are morphologically and functionally unaffected by transplantation [23] . Following development and subsequent transection in a dync1h1-/- environment , wild type axons were able to regenerate robustly for the first 9 hours after sprouting ( Fig 3C–3F ) , in a manner indistinguishable from wild type axons in a fully wild type environment . This indicates that restoring dynein specifically in neurons in a dynein mutant is sufficient to promote axonal regrowth , demonstrating a neuron-intrinsic role for dynein during peripheral nerve regeneration . Interestingly , we found that dync1h1-/- axons that had wild type axons in the same nerve regrew more robustly than dync1h1-/- axons in nerves with no transplanted cells ( 14 . 23 ± 2 . 06 μm growth in dync1h1-/- larvae without transplants , see below; 39 . 33 ± 4 . 72 μm growth in dync1h1-/- larvae with transplants , Fig 3F ) . In several instances , we observed dync1h1-/- axons growing along previously extended wild type axons ( Fig 3G–3I ) . This indicates that the presence of wild type axonal regrowth is able to partially rescue the dync1h1-/- axonal regrowth defects , likely through cell-cell adhesions between the dynein mutant axon and wild type axons . We next asked how dynein promotes axonal regeneration within peripheral nerves . Peripheral nerve regeneration is a dynamic process composed of several defined stages , starting with growth cones emerging from the proximal stump and starting to probe the injury gap environment . This is followed by stabilization of axonal regrowth across the injury gap and along the correct trajectory , and finally rapid , sustained axonal regrowth towards their original targets [24] . We used live cell imaging after nerve transection to quantify axonal dynamics in dynein mutants and determine which of these stages require dynein . In wild type siblings , we observed growth cones emerging from the proximal stump extending ( 3 . 54 events per 8 hours ) and retracting ( 1 . 08 events per 8 hours ) repeatedly , consistent with the idea that these growth cones are probing the injury gap for a path towards their original targets ( Fig 4A and 4B ) . We found that dync1h1-/- axons exhibit similar frequencies of axonal extensions and retractions ( Fig 4C and 4D ) , suggesting that they probe the injury gap as actively as their wild type siblings ( Fig 4E ) . We next examined the second stage of axonal regeneration when axons become stabilized and then extend toward their original targets . To quantify this process we measured the overall displacement of regenerating growth cones over the first ~8 hours after sprouting began . We found that the majority of regenerating wild type axons grew beyond the transection site within 8 hours of sprouting ( Fig 4F ) , travelling an average of 41 . 49 μm ( SEM ± 5 . 84 ) over this time period . In contrast , regenerating dync1h1-/- axons rarely extended beyond the transection site ( Fig 4G ) , travelling an average of 14 . 23 μm ( SEM ± 2 . 06 ) and never exceeding 21 . 94 μm in growth . Moreover , quantification of growth cone displacement at 8 hours post transection revealed that compared to regenerating wild type axons , dync1h1-/- axons exhibited a significant decrease in axonal extension ( Fig 4H ) . Combined these results argue that rather than initiating growth cone sprouting and short range axonal extensions , dynein predominantly acts early during axonal regeneration to stabilize regenerating axons thereby promoting persistent , long-range regrowth . Dynein has recently been shown to play a critical role in generating and maintaining microtubule organization , both processes central to axonal growth [10 , 11 , 25 , 26] . To determine whether dynein regulates microtubule dynamics in axons during regeneration , we used a transgenic line that simultaneously labels actin and microtubules in motor neurons ( mnx1:Gal4; UAS:lifeact-GFP-v2a-EB3-RFP ) . Growth cone extension occurs in three stages: first , protrusion driven by F-actin , then engorgement driven by microtubule-based transport of organelles and vesicles , and finally consolidation in which the growth cone contracts and stabilizes to form a cylindrical axon shaft [27] . In regenerating wild type axons , filopodia extend at the growth cone and microtubules follow behind , stabilizing and consolidating newly formed protrusions ( Fig 5A–5D ) . The majority of regenerating dync1h1-/- axons ( n = 30/37 ) displayed one of two phenotypes characteristic for microtubule disruption . In 59 percent ( n = 22/37 ) we observed filopodia extension followed briefly by microtubule extension ( Fig 5E and 5F ) and then arrest at the engorgement stage before finally retracting ( Fig 5G and 5H ) . In 22 percent ( n = 8/37 ) of regenerating dync1h1-/- axons , microtubules faithfully followed filopodia extending at growth cones . However , rather than consolidating in the proximal growth cone , they adopted a looped conformation at the leading edge of the growth cone , leading to stalling and retraction ( Fig 5I and 5M ) . The remaining 19 percent had straight , ordered microtubules ( n = 7/37 ) . This suggests that a lack of dynein may lead to loss of microtubule organization at regenerating growth cones and stalling of regenerating axons early during the regeneration process . To determine whether stabilizing microtubules during axonal regrowth could compensate for a lack of dynein , we transected nerves in dynein mutant larvae and subsequently treated the larvae with taxol . We used timelapse imaging to assess axonal regrowth dynamics and found that in dynein mutant embryos stabilizing microtubules with 5 μm taxol partially rescued axonal regrowth ( Fig 5N ) . Combined , these findings support a model by which dynein plays a critical role in regulating microtubule dynamics , thereby stabilizing growth of regenerating axons as they initiate their trajectory across the injury gap and towards their original targets . Thus , we demonstrate a role for dynein in promoting axonal extension via microtubule stabilization , as well as a previously uncharacterized role in initiating Schwann cell response to injury .
All experiments were conducted according to an Animal Protocol fully approved by the Uni- versity of Pennsylvania Institutional Animal Care and Use Committee ( IACUC ) on January 24 , 2014 , protocol number 803446 . Veterinary care is under the supervision of the University Laboratory Animal Resources ( ULAR ) of the University of Pennsylvania . All transgenic lines were maintained in the Tübigen or Tupfel long fin genetic background and raised as previously described [22] . The Tg ( mnx1:GFP ) ml2 line [28] was used to label spinal motor nerves and the Tg ( sox10 ( 7 . 2 ) :mRFP ) vu234 line [29] was used to label Schwann cells . The Tg ( UAS:lifeact-GFP-v2a-EB3-RFP ) line was used to label microtubules and actin . The dync1h1hi3684Tg [30] and gpr126stl47 [31] mutant strains were used and genotyping protocols were performed as previously described . Nerve transection and live imaging were performed as previously described [15] . Axon growth extent quantification was performed as previously described [20] . Transected nerves in which axons failed to regrow or did not extend through the entire length of the ventral myotome are categorized as “no/weak regeneration . ” Nerves with at least one fascicle that extended through the entire length of the ventral myo- tome are categorized as “moderate regeneration . ” Finally , nerves with two or more fascicles extending through the entire length of the ventral myotome are categorized as “strong regeneration . ” Axons were imaged every 15 minutes from ~7 to ~16 hpt . Extensions and retractions were defined as growth or retraction of >1 μm between timelapse frames and number of extension and retraction events was counted . Continued movements of the same direction in a subsequent frame were not counted as new events . Measurements were performed on each visibly distinct axon in a nerve . Axons imaged at ~16 hpt were measured by drawing a line from the spinal cord exit point to the growth cone . Measurements were performed on each visibly distinct axon in a nerve . Axons and Schwann cells were imaged before transection and every 15 minutes from ~1 to ~5 hpt . Schwann cell width was measured at the widest point in pre- and post-transection images . Using ImageJ , a line was drawn from one edge of the Schwann cell membrane to the other in an orientation perpendicular to the motor nerve and was measured in microns . Cell transplantations were performed as previously described [32] . Wild type cells were transplanted into dync1h1-/- embryos in areas known to develop into motor neurons . Larvae were screened at 5 dpf to identify nerves that contained transplanted motor neurons and no other transplanted cell types along the path of the ventral motor nerve . Transection , imaging , and quantification of growth cone displacement in identified nerves were performed as described above . Larvae were transected according to the above protocol . 3 hours after transection , embryos were bathed in embryo medium with either 1% DMSO or 1% DMSO and 5 μM taxol ( paclitaxel , Life Technologies #P3456 ) . When mounted for overnight timelapse imaging , larvae were bathed in Ringer’s solution with either 1% DMSO or 1% DMSO and 5 μM taxol for the duration of imaging . Image stacks were compressed into maximum intensity projections ( MIPs ) in Slidebook 6 then processed using ImageJ and Photoshop to normalize brightness and contrast . Fisher’s exact and Student’s t tests were performed on all applicable datasets .
Nerve injury induces a local signaling cascade that leads to the production of axon intrinsic signals at the lesion site [33] . There is overwhelming evidence that dynein is critical to transport these injury signals from the lesion site to the cell body where they initiate a neuronal injury response [34–36] . We find that in presumptive dynein null mutants , injured neurons robustly respond to the injury and within ~8–10 hours , regenerating axons sprout from the proximal stump , indistinguishable from what we observe in wild type animals . This raises the question whether axonal sprouting can occur independently of dynein-dependent injury induced signals , or whether in our zebrafish model dynein-mediated retrograde transport is less important to mount an injury response ? One clear difference between rodent models and our model is the distance between the injury site and the neuronal cell bodies . In rodent sciatic nerve models lesions are introduced millimeters away from neuronal cell bodies [36] , while in larval zebrafish—due to the smaller animal size—lesions are generated about 10–50 μm away from neuronal cell bodies [15] Thus , it is conceivable that due to the almost 100-fold reduction in distance between lesion site and cell body , injury signal propagation from the lesion site to the cell body is less dependent on dynein function . Although it remains unclear how injury signal propagation can occur independent of dynein , this provided us with the unique opportunity to examine dynein’s role in peripheral nerve regeneration beyond its role in injury signal transport . Endpoint analyses at 48 hpt uncovered a clear role for dynein in peripheral nerve regeneration , with clear effects on both axonal regrowth as well as on injury-induced Schwann cell remodeling ( Figs 1 , 2 and 4 ) . Using live-imaging to visualize the early stages of the regeneration process , we found that dynein promotes the stabilization and growth of long-range axonal projections , providing compelling evidence that apart from its well-documented role in retrograde injury signal transport , dynein also plays a critical role in sustaining axonal regrowth . Moreover , simultaneously visualizing the cellular behavior of both axons and Schwann cells revealed that loss of dynein prevented injury-induced Schwann cell remodeling . The transition of Schwann cells from their fully differentiated state to a repair cell state is a well-documented and integral aspect of peripheral nerve regeneration [37 , 38] , accompanied by dramatic morphological changes to the Schwann cell , as the cell breaks down its myelin and extends its membrane to engulf axonal debris [39 , 40] . Dynein regulates several steps of membrane trafficking , including ER to golgi transport , as well as endosomal trafficking [41] , so it is conceivable that dynein plays a direct , cell-autonomous role in this process . Alternatively , the inability of Schwann cells to initiate the remodeling process might be a consequence of strongly reduced axonal regrowth , and future experiments will be required to test a possible Schwann cell-specific role for dynein in the remodeling process . Given that dynein mutants exhibit defects in axonal regrowth and Schwann cell morphology , we performed chimeric analysis experiments . These experiments revealed that dynein function in injured neurons is sufficient to sustain axonal regeneration . Importantly in our chimera experiments , of the roughly 60 axons contributing to an individual motor nerve [42] , on average only 1–3 transplanted wild type axons were present . This low level of chimerism was critical to evaluate regrowth capacity of individual wild type axons . This also revealed that the presence of individual wild type axonal regrowth facilitated regrowth of individual , neighboring dynein deficient axons ( Fig 5G–5I ) . At the same time , the low level of chimerism precluded us from asking whether neuronal dynein restored all aspects of peripheral nerve regeneration , including the overall robustness of axonal regrowth for a whole nerve and injury-induced Schwann cell remodeling . Thus , while neuronal dynein plays a critical role in sustaining axonal regrowth , we cannot exclude the possibility that dynein function in Schwann cells also contributes to peripheral nerve regeneration . Cytoskeletal dynamics are critical to growth cone formation [43] , axonal outgrowth during development [44] , and axonal regeneration [45] . Previous studies have revealed that microtubule stabilization promotes axonal regrowth after injury both in vitro and in vivo [46–48] . Interestingly , studies of C . elegans dynein heavy chain mutants recently revealed that dynein acts locally in dendrites to stabilize microtubules [10] . This raised the possibility that dynein may also act locally in regenerating axons to stabilize microtubules . We assessed cytoskeletal dynamics during regeneration using a transgene that allowed us to visualize actin and microtubules simultaneously in live , regenerating axons . This revealed that while actin dynamics were grossly unaffected in dynein mutant axons , microtubules often appeared unstable and disordered , with some axons exhibiting looping microtubule configurations reminiscent of those seen in the dendrites of C . elegans dynein heavy chain mutants [10] . Thus , our results provide compelling evidence that besides its well-documented role in retrograde transport , dynein also promotes microtubule stability critical for growth cone advancement [49] , providing a potential mechanism for the rapid and sustained extension observed during wild type axonal regrowth , and deficient in dynein mutants ( Fig 2 ) . Dynein is also known to modulate microtubule dynamics is through microtubule sliding [50] , providing an alternative mechanism . This might be a direct effect or may affect microtubule sliding indirectly via modulation of kinesin-1 as these motors have been shown to transport each other directly with one another [51] . Taken together , our results suggest that beyond its function in retrograde injury signaling dynein has a multifaceted role in nerve regeneration that warrants further studies .
|
Nerve regeneration requires coordinated responses from multiple cell types after injury . Axons must extend from the neuronal cell body back towards their targets , while surrounding Schwann cells enter a repair cell state in which they promote regeneration . While nerves of the peripheral nervous system can regrow , it is estimated that fewer than 10 percent of patients fully recover function after nerve injury . In order to understand the mechanisms by which peripheral nerves regrow , we used live cell imaging in the zebrafish to observe the process of nerve regeneration , monitoring axons and Schwann cells simultaneously during this process . Using genetic mutants , we identified a role for the molecular motors kinesin-1 and dynein in promoting axonal regrowth . Furthermore , we found that dynein plays an additional role in Schwann cell response to injury . Thus , we demonstrate that molecular motors are required in multiple cell types to promote nerve regeneration .
|
[
"Abstract",
"Introduction",
"Results",
"Materials",
"and",
"methods",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"microtubules",
"nervous",
"system",
"cell",
"processes",
"neuroscience",
"macroglial",
"cells",
"motor",
"neurons",
"nerve",
"regeneration",
"developmental",
"biology",
"dyneins",
"molecular",
"motors",
"organism",
"development",
"schwann",
"cells",
"nerve",
"fibers",
"morphogenesis",
"cellular",
"structures",
"and",
"organelles",
"cytoskeleton",
"animal",
"cells",
"proteins",
"axons",
"glial",
"cells",
"axonal",
"transport",
"biochemistry",
"cytoskeletal",
"proteins",
"cellular",
"neuroscience",
"cell",
"biology",
"regeneration",
"anatomy",
"nerves",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"microtubule",
"motors"
] |
2019
|
Dynein promotes sustained axonal growth and Schwann cell remodeling early during peripheral nerve regeneration
|
While C57BL/6 mice infected in the ear with L . major mount a vigorous Th1 response and resolve their lesions , the Th1 response in C57BL/6 mice infected with L . mexicana is more limited , resulting in chronic , non-healing lesions . The aim of this study was to determine if the limited immune response following infection with L . mexicana is related to a deficiency in the ability of monocyte-derived dendritic cells ( mo-DCs ) to prime a sufficient Th1 response . To address this issue we compared the early immune response following L . mexicana infection with that seen in L . major infected mice . Our data show that fewer monocytes are recruited to the lesions of L . mexicana infected mice as compared to mice infected with L . major . Moreover , monocytes that differentiate into mo-DCs in L . mexicana lesions produced less iNOS and migrated less efficiently to the draining lymph node as compared to those from L . major infected mice . Treatment of L . mexicana infected mice with α-IL-10R antibody resulted in increased recruitment of monocytes to the lesion along with greater production of IFN-γ and iNOS . Additionally , injection of DCs into the ear at the time of infection with L . mexicana also led to a more robust Th1 response . Taken together , these data suggest that during L . mexicana infection reduced recruitment , activation and subsequent migration of monocytes and mo-DCs to the draining lymph nodes may result in the insufficient priming of a Th1 response .
Infection with Leishmania results in a variety of outcomes , depending on the parasite species and immune response mounted by the host [1] . Murine disease models resemble human disease , with some infections being self-healing and others chronic . Resolution of leishmaniasis requires the production of IFN-γ by Th1 cells; the absence of a strong Th1 response results in chronic disease with non-healing lesions [2] , [3] . Th1-mediated protection is promoted by IFN-γ-induced production of nitric oxide ( NO ) in infected cells , which ultimately leads to parasite killing [2] , [3] . In C57BL/6 mice , infection with L . major results in a strong Th1 response with self-resolving lesions , in contrast , L . mexicana lesions fail to resolve [4] . The chronic nature of L . mexicana lesions is most likely due to their inability to stimulate an effective Th1 response [4] , [5] , [6] . Similarly , L . amazonensis fails to induce a strong Th1 response and leads to chronic lesions in mice [7] , [8] , [9] , [10] . However , the immune mechanisms limiting Th1 responses following either L . mexicana or L . amazonensis infection are not yet fully defined . Several demonstrations that infection with L . mexicana suppresses IL-12 production by macrophages and dendritic cells ( DCs ) [11] , [12] , [13] suggested that failure to produce IL-12 may limit the Th1 response , resulting in the observed susceptibility to L . mexicana [14] , [15] , [16] . However , we found that administration of IL-12 failed to promote disease resolution , suggesting that the inability of L . mexicana mice to resolve their infection is not solely dependent upon lack of IL-12 [17] . Therefore , we hypothesized that a more generalized deficit in DC function may contribute to the chronic lesions that develop following L . mexicana infection . Monocyte-derived DCs ( mo-DCs ) play an important role in the development of protective immunity [18] , [19] , [20] , [21] . Mo-DCs differentiate from inflammatory monocytes ( CD11b+ , Ly6C+ , CCR2+ and CX3CR1lo ) recruited to sites of inflammation . Once activated , mo-DCs produce inducible nitric oxide synthase ( iNOS ) [22] . Indeed , mo-DCs appear to be the major producers of iNOS during L . major infection [23] and are therefore likely essential for reducing the parasite burden . In addition to iNOS production , mo-DCs contribute to immunity following infection with L . major by migrating to draining lymph nodes ( dLN ) where they stimulate antigen-specific Th1 T cell responses [24] . Moreover , we recently found that L . major-activated DCs induce lymph node hypertrophy , which promotes additional recruitment of naïve T cells into the lymph node to enhance the protective response [25] . Taken together , these data indicate that mo-DCs play an important role in the development of protective immunity to . L . major , and that a deficit in their recruitment or activation might limit a protective Th1 response . In the present study , we investigated whether the meager Th1 response observed during L . mexicana infection is due to limitations in the: 1 ) recruitment of monocytes from the blood to the site of infection; 2 ) differentiation of monocytes into iNOS-producing mo-DCs; and/or 3 ) migration to the draining lymph node . We found that monocyte recruitment to the site of infection was reduced in L . mexicana infected mice compared to L . major infected mice . Moreover , while monocytes in L . mexicana lesions upregulated expression of CD11c , they produced significantly less iNOS and migrated less efficiently to the draining lymph node relative to monocytes in L . major infected mice . Following treatment of L . mexicana infected mice with α-IL-10R antibody , there was increased recruitment of monocytes to the lesion , as well as increased production of IFN-γ and iNOS . Additionally , when DCs were injected into the ear at the time of infection with L . mexicana , there was a more robust Th1 response . These data imply that the poor Th1 response observed during L . mexicana infection results from both reduced monocyte recruitment to the lesions , and a relative deficit in their differentiation into functional effector populations .
All animal studies were carried out in compliance with the guidelines of the Institutional Animal Care and Use Committee ( IACUC ) of the University of Pennsylvania and in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The animal protocol was approved by the IACUC of the University of Pennsylvania , Philadelphia PA . Female C57BL/6 ( B6 ) and B6-Ly5 . 2/Cr ( CD45 . 1 ) mice were purchased from the National Cancer Institute ( Fredricksburg , MD ) . Animals were maintained and experiments were carried out in a specific pathogen-free environment . L . major V1 parasites ( MHOM/IL/80/Friedlin ) or L . mexicana parasites ( MNYC/BZ/62/M379 ) were grown until stationary phase in Schneider's Drosophila medium ( Gibco , Grand Island , NY ) supplemented with 20% heat-inactivated FBS ( Gibco ) and 2 mM l-glutamine ( Sigma ) . Metacyclic promastigotes were isolated by density gradient [26] . For infection of mice , 1×105 metacyclic parasites were injected into the ear . For flow cytometry , cells were isolated from ears or draining lymph nodes . Dermal ear sheets were separated and incubated in Liberase TL ( Roche , Indianapolis , IN ) for 1 hr at 37°C . Ears and draining lymph nodes were made into single cell suspensions and washed with PBS . Fixable Aqua dye ( Invitrogen , Carlsbad , CA ) was added to assess cell viability . Cells were then incubated with Block ( CD16/32 , inactivated mouse sera and Rat IgG ) followed by fluorochrome-conjugated antibodies for surface markers ( CD11c , CD11b , CD45 . 1 , CD45 . 2 , MHCII ( e-Bioscience , San Diego , CA ) , and Ly6C ( BD Pharmingen , San Diego , CA ) . Intracellular staining was performed for iNOS using an unconjugated anti-iNOS/NOS II rabbit polyclonal IgG ( Millipore , Temecula , CA ) followed by flourochrome-conjugated donkey-anti-rabbit IgG ( e-Bioscience ) . Briefly , surface-stained cells were fixed in PBS with 2% paraformaldehyde and then permeabilized with 0 . 2% saponin in FACS staining buffer ( PBS containing 0 . 1% BSA ) . Cells were fixed by using 2% paraformaldehyde and samples were acquired on a FACS Canto flow cytometer ( BD Pharmingen ) . Analysis was performed using FlowJo software ( Tree Star , Ashland , OR ) . Bone marrow was harvested from tibias and femurs of naïve mice . Following lysis of red blood cells with ACK lysis buffer ( Lonza , Walkersville , MD ) , Miltenyi MACS columns were used to purify monocytes . Briefly , anti-Ly6G-biotin antibody and biotin microbeads ( Miltenyi , Auburn , MD ) were added to the cells , which were then placed over a LS column . Ly6G+ cells on the column were discarded and anti-PE CD11b ( e-Bioscience ) and PE microbeads were added to the flow through , which was placed over a 2nd LS column . CD11b+ cells attached to the column were washed off . Monocytes were enriched to 50% as evaluated by flow cytometry . 1×106 total cells from the final column were transferred into mice by intradermal injection into the lesions . Mice were anesthetized using ketamine/xylazine and the ventral side of each ear was painted with FITC isomer ( Sigma , St . Louis , MO ) . FITC ( 8 mg/mL ) was dissolved in an equal volume of acetone and dibutyl phthalate ( Sigma ) and 25 µL of the mixture was applied to the skin . Migration of FITC+ cells was assessed in the draining lymph node 48 hours following application . Mice were injected intraperitoneally with 500 µg of α-IL-10R antibody ( 1B1 . 3A , BioXcell , West Lebanon , NH ) one day prior to intradermal infections of 1×105 L . mexicana metacyclics in both ears . Mice were subsequently treated with 500 µg of α-IL-10R antibody on day 3 and then 250 µg every 3 days until the final harvest at 2 weeks post-infection . Supernatants from draining lymph node cultures stimulated with L . mexicana freeze-thaw antigen for 72 hrs were collected and assayed by sandwich ELISA using paired monoclonal antibody to detect IFN-γ . DCs were generated as previously described [27] . Briefly , bone marrow cells from C57BL/6 mice were isolated from femurs and tibias of mice by syringe flushing . Bone marrow cells were counted and seeded into 6-well plates at 5×105 cells/mL in 3 mLs of media - RPMI 1640 ( Gibco ) supplemented with 10% heat-inactivated FBS ( Gibco ) , 2 mM glutamine ( Sigma ) , 50 µM 2-ME ( Gibco ) , 100 U/mL penicillin ( Sigma ) , 100 µg/mL streptomycin ( Sigma ) and 20 ng/mL GM-CSF ( Peprotech , Rocky Hill , NJ ) per well . Cells were maintained at 37°C with 5% CO2 and fed on days 3 , 6 and 8 with 3 mLs of fresh media . Cells were harvested on day 10 and injected into the ear of C57BL/6 mice at the time of infection . Briefly , 1×106 DCs and 1×105 L . mexicana metacyclics were mixed immediately prior to injecting into the ear . Statistical significance was determined using unpaired , two-tailed Student's t test . Results with a p value ≤0 . 05 were considered significant .
The development of a protective response following L . major infection is associated with the recruitment of monocytes into the lesion , which are believed to differentiate into mo-DCs ( defined as CD11bhi , CD11c+ , Ly6C+ ) to prime a strong Th1 response [24] . Since L . mexicana infection promotes chronic , non-healing lesions and a minimal Th1 response , we hypothesized that fewer monocytes would be recruited to lesions following infection with L . mexicana compared to L . major . To test this , we infected C57BL/6 mice with either L . major or L . mexicana parasites and assessed the cellular composition of the lesions at 3 and 14 days post-infection . Expression of CD11b , a subunit of αMβ2 ( also known as Mac-1 and CR3 ) , was used to detect infiltrating leukocyte populations , including monocytes , macrophages , and granulocytes [28] . At 3 days post-infection , there was a significant increase in the percentage of CD11bhi cells in dermal lesions from L . major infected mice compared to normal skin ( Fig . 1A ) . In contrast , no increase in CD11bhi cells was observed in lesions from L . mexicana infected mice . Moreover , the difference in percentage of CD11bhi cells between L . major and L . mexicana lesions was still evident two weeks after infection ( Fig . 1A ) . In contrast , neutrophil ( CD11bhi Ly6G+ ) frequency increased equally in lesions of both L . major and L . mexicana infected mice by day 14 as compared to normal skin ( data not shown ) . Consistent with the observed alterations in CD11bhi cells , there was an increase in inflammatory monocytes ( CD11bhi CD11c− Ly6C+ ) in the lesions from 3-day and 14-day L . major infected mice as compared with normal skin ( Fig . 1B ) while no such increase was observed following L . mexicana infection ( Fig . 1B ) . By day 14 , mo-DCs ( CD11bhi CD11c+ Ly6C+ ) were evident in lesions of both L . major and L . mexicana infected mice , however , mo-DCs were preferentially represented in L . major lesions ( Fig . 1C ) . Taken together , these results suggest that L . mexicana fails to promote the recruitment of monocytes , reducing the number of cells available for subsequent differentiation into mo-DCs capable of controlling parasite numbers . Although there were fewer monocytes recruited to lesions from L . mexicana infected mice compared with those from L . major infected mice , the ratio of monocytes to mo-DCs was similar ( Fig . 2 ) . In these experiments we did not determine if the CD11c+ Ly6C+ cells were derived from monocytes , but based on previous findings [24] , this is our assumption . However , mo-DCs in the lesions from L . mexicana infected mice expressed significantly less iNOS compared with mo-DCs from L . major lesions . Thus , while approximately 20% of the mo-DCs in lesions from L . major infected mice were iNOS+ , only 3% were iNOS+ in lesions from L . mexicana infected mice ( Fig . 3A ) . Similarly , the number of iNOS-producing mo-DCs was significantly reduced in lesions from mice infected with L . mexicana ( Fig . 3B ) . CD11bhi CD11c− Ly6C+ , inflammatory monocytes , did not make iNOS in either L . major or L . mexicana infected mice ( data not shown ) . These data demonstrate that there are fewer iNOS-producing mo-DCs in L . mexicana infected mice , potentially contributing to the inability of these mice to resolve their infection . In addition to killing parasites at the site of infection through iNOS-dependent mechanisms , mo-DCs also migrate to dLNs where they orchestrate the developing immune response through antigen presentation and regulation of cytokine production [29] . Recently , we have also shown that L . major-activated DCs promote lymph node hypertrophy following infection [25] and the impaired lymph node expansion following L . mexicana infection [30] led us to investigate if a reduction in mo-DCs migration to the draining lymph node during L . mexicana infection limits the Th1 response and impairs lymph node expansion . To evaluate the ability of DCs to migrate from the site of infection , C57BL/6 mice were infected in the ear with L . major or L . mexicana and two weeks post-infection the ears were FITC painted . After 48 hours we compared the FITC+ DCs ( CD11c+ MHCIIhi ) in the draining lymph node from naïve , L . major infected or L . mexicana infected mice . Notably , there were significantly more FITC+ DCs in L . major infected mice when compared to either naïve or L . mexicana infected mice . In contrast , there was no difference in the number of FITC+ DCs between naïve and L . mexicana infected mice ( Fig . 4 ) , indicating that mo-DCs migration to the dLN is compromised in L . mexicana infected mice . We next wanted to determine if the microenvironment within L . mexicana lesions actively inhibited DC migration . Therefore , we injected the same number of CD45 disparate monocytes into L . major or L . mexicana lesions and evaluated their migration to the draining lymph node . Figure 5A shows an equivalent number of CD11b+ CD45 . 1+ cells in the ear of L . major or L . mexicana infected mice approximately 18 hours following monocyte transfer . Interestingly , the expression of Ly6C on the donor monocytes was lower in L . major infected mice as compared to L . mexicana infected mice ( Fig . 5B ) . As downregulation of Ly6C is associated with activation of mo-DCs [24] , [31] , these data suggest that mo-DCs in L . mexicana infected mice do not differentiate as efficiently as mo-DCs from L . major infected mice . Even more strikingly , there is a dramatic increase in both the frequency and absolute number of transferred cells in the draining lymph node of L . major infected as compared to L . mexicana infected mice ( Fig . 5C and D ) , indicating that L . mexicana infection does not increase mo-DC trafficking to dLNs . However , we cannot exclude the possibility that there may be a difference in retention in the dLN of L . major versus L . mexicana infected mice . Together , these data indicate that a lack of mo-DCs migration from the site of L . mexicana infection to the draining lymph node may prevent T cell priming and impair lymph node expansion , precluding the induction of a protective Th1 response and resulting in the development of chronic disease . IL-10 has been described as having anti-inflammatory effects during infection by inhibiting cytokine production and antigen presentation [32] , however , more recently it was shown that IL-10 also limits recruitment of CD11b+ Ly6C+ monocytes following T . brucei infection [33] . Since IL-10−/− mice infected with L . mexicana resolve their lesions [6] , we wanted to investigate whether blocking interaction of IL-10 with its receptor would lead to increased monocyte recruitment . We infected C57BL/6 mice as before with L . mexicana and treated one group with α-IL-10R antibody . We evaluated monocyte recruitment to lesions on days 7 and 14 following infection and found that there was a greater percentage and number of monocytes recruited to L . mexicana lesions in mice treated with α-IL-10R ( Fig . 6A ) . Similarly , the percentage and number of mo-DCs in the lesions of L . mexicana infected mice was also significantly increased when IL-10R was blocked ( Fig . 6A ) . Moreover , there were increased levels of IFN-γ in the draining lymph nodes ( Fig . 6B ) , as well as a greater percentage and number of iNOS-producing mo-DCs in the lesions of L . mexicana infected mice treated with α-IL-10R ( Fig . 6C ) . These data suggest that IL-10 is a key factor contributing to the limited number of monocytes observed during L . mexicana infection since blocking the interaction of IL-10 with its receptor results in a dramatic increase in monocytes and mo-DCs in the lesion . Surprisingly , we did not see a difference in the parasite burden in treated and untreated L . mexicana infected mice at this early time point , in spite of the fact that we have previously shown that IL-10−/− mice eventually resolve their L . mexicana lesions [6] . Our assumption is that the effect on parasite burden is simply delayed and will develop later . The previous experiment , where L . mexicana infected mice were treated with α-IL-10R antibody , suggests that the increase in mo-DCs in the lesion may result in the priming of an improved Th1 response . Here , we test whether there is a correlation between increased numbers of DCs in the lesion and a more robust Th1 response . We injected DCs into the ear at the time of infection with L . mexicana and we compared the Th1 response 14 days post-infection to L . mexicana infected mice receiving no DCs . As predicted , L . mexicana infected mice receiving DCs produced greater levels of IFN-γ ( Fig . 7A ) , and had a greater percentage and number of iNOS-producing DCs ( Fig . 7B ) . Moreover , the impaired lymph node expansion that occurs during infection with L . mexicana was overcome in mice that received DCs ( Fig . 7C ) . Taken together , these data suggest that the limited Th1 response observed in L . mexicana infected mice can be overcome if a greater number of DCs can be established in the lesion .
Infection of C57BL/6 mice with either L . major or L . mexicana results in cutaneous lesions . However , while L . major-induced lesions heal , those induced by L . mexicana infection do not . The chronicity of L . mexicana infections is attributable to the limited Th1 response mounted by the host to the parasite [5] , [6] , [34] . Since the development of a Th1 response in leishmaniasis depends upon IL-12 production by DCs [35] , [36] , [37] , [38] , it was originally thought that L . mexicana fails to induce a healing response due to its inability to stimulate IL-12 production [11] , [12] , [14] , [16] . However , the limited Th1 response in L . mexicana infected mice is not reversed by treatment with rIL-12 [17] , suggesting that there is a more generalized impairment in DC function . Differentiation of mo-DCs from inflammatory monocytes at the site of infection plays an essential role in immune protection in a number of infectious diseases [19] , [21] . Monocytes are recruited to L . major infected skin [23] , [24] , [39] and mo-DCs are thought to be essential for the induction of the Th1 response in L . major infection [24] , suggesting that limitations in monocyte recruitment and differentiation ( or both ) may lead to chronic disease following L . mexicana infection . In support , our current studies demonstrate that fewer monocytes are recruited during infection with L . mexicana when compared to L . major , and the consequent reduction in differentiated mo-DCs present in L . mexicana lesions likely compromises generation of a protective Th1 response . In addition , reduced monocyte recruitment and the observed decrease in iNOS expression will limit the killing capacity of these cells [39] , presumably leading over time to increased parasite burden . Together , the limited Th1 response and enhanced parasite burden could promote the chronic exacerbated disease observed following L . mexicana infection . The importance of monocyte recruitment in limiting the progression of infectious diseases has been most clearly demonstrated in CCR2 deficient ( CCR2−/− ) mice . CCR2 is a chemokine receptor expressed on inflammatory monocytes that mediates monocyte chemotaxis . In CCR2−/− mice , Ly6Chi monocytes accumulate in the bone marrow due to their inability to emigrate from this site . Limited recruitment of monocytes to the site of infection likely contributes to the enhanced susceptibility of CCR2−/− mice to Listeria infection [40] . In addition , following oral Toxoplasma gondii infection of CCR2−/− mice , monocytes fail to be recruited to the illeum , allowing for uncontrolled parasite growth . However , adoptive transfer of CCR2-expressing monocytes into T . gondii infected CCR2−/− mice protected them from this otherwise lethal infection [41] . CCR2−/− mice infected with L . major are also more susceptible to infection due to an attenuated Th1 response [42] . Interestingly , treatment of CCR2−/− mice with rIL-12 is able to reverse the susceptibility to L . major infection [43] . Since mo-DCs have been described as the major producers of IL-12 during L . major infection [24] , these data support our hypothesis that compromised recruitment of monocytes to the lesion influences the development of a Th1 response in L . mexicana infected mice . As IL-10 has been shown to limit the recruitment of CD11b+ Ly6C+ monocytes during infection with T . brucei [33] and we have previously shown that IL-10−/− mice infected with L . mexicana resolve their lesions [6] , we hypothesized that monocyte recruitment following L . mexicana infection is impacted by IL-10 production at the lesion site . In fact , we showed that by blocking IL-10R , there was increased recruitment of CD11bhi Ly6C+ monocytes to L . mexicana infected lesions . Moreover , L . mexicana infected mice treated with α-IL-10R produced significantly more iNOS and IFN-γ than L . mexicana infected C57BL/6 mice . As during T . brucei infection [33] , it is likely that production of IL-10 in L . mexicana-induced lesions may work on several levels . IL-10 could lead to decreased levels of CCL2 , which would explain the limited recruitment of monocytes into the lesions . IL-10 is also capable of dampening Th1 responses , which would result in lower levels of iNOS and IFN-γ . Therefore , these data provide a mechanism as to why there is limited recruitment of monocytes to the lesion during infection with L . mexicana . Finally , while DCs are clearly needed to prime T cells in the draining lymph node , they also promote lymph node hypertrophy . We previously demonstrated that lymph node hypertrophy is associated with the protective response to L . major infection [30] and have more recently revealed that L . major-activated DCs are responsible for lymph node expansion [25] . During infection with L . mexicana , lymph node hypertrophy is greatly reduced , potentially limiting the immune response [30] . Here we have used two methods to track migration of mo-DCs from the lesion to the draining lymph node; one method marked endogenous mo-DCs in the lesion and the other utilized injection of CD45 disparate monocytes directly into the lesion . While fewer endogenous mo-DCs from the L . mexicana lesion migrated to the dLN as compared to L . major , this may have been due to the relatively low numbers of monocytes initially present within the lesions of L . mexicana infected mice . To address this problem , we injected equal numbers of monocytes into L . major or L . mexicana lesions , and found that there was still a deficit in the migration of mo-DCs to the dLN from L . mexicana lesions . An inability of mo-DCs to migrate to the dLN could prevent both antigen-specific responses , as well as mo-DC-driven lymph node hypertrophy , providing a potential explanation for the reduction in lymph node size in L . mexicana infected mice . Interestingly , if DCs are injected into the ear at the same time of infection with L . mexicana , mice have significantly larger lymph nodes and are able to mount a more robust Th1 response compared to mice that did not receive DCs . These data clearly demonstrate that mo-DCs are important in initiating an appropriate immune response against Leishmania and that the limited recruitment of monocytes observed during L . mexicana infection could lead to the chronic nature of the disease . In summary , we have demonstrated that 1 ) fewer monocytes are recruited to lesion during infection with L . mexicana as compared to L . major , 2 ) fewer iNOS producing mo-DCs are present in the lesions of L . mexicana infected mice 3 ) fewer mo-DCs migrate to the dLN node during L . mexicana infection , 4 ) blocking IL-10R leads to increased monocyte recruitment and a more robust Th1 response during L . mexicana infection , and 5 ) injection of DCs into the ear at the time of infection with L . mexicana also leads to increased levels of iNOS and IFN-γ . Together , these findings provide a mechanistic basis for the limited Th1 response , and lack of lymph node hypertrophy observed in L . mexicana infected mice and offer a better understanding of the important role that monocytes play during infection with Leishmania .
|
Leishmaniasis , caused by protozoan parasites belonging to the genus , Leishmania , exhibits clinical symptoms ranging from mild cutaneous lesions to more severe cutaneous or visceral disease . Here , we focus on L . major and L . mexicana , two species that lead to self-resolving and chronic cutaneous lesions , respectively . A strong Th1 response is necessary for resolution of disease following L . major infection . However , L . mexicana infection induces a limited Th1 response resulting in chronic disease . Monocyte-derived dendritic cells are believed to be important in priming the Th1 response during L . major infection , and therefore in this study we evaluated whether there are quantitative and/or qualitative differences in monocyte-derived dendritic cells following L . mexicana infection . We found that fewer monocytes were recruited to the lesions of L . mexicana infected mice as compared to mice infected with L . major . In addition , there were fewer iNOS producing monocyte-derived dendritic cells in the lesions of L . mexicana infected mice and less migration of monocyte-derived dendritic cells to the draining lymph node . Manipulations that allow for increased monocytes in the lesions of L . mexicana infected mice also resulted in a more robust Th1 response . Thus , these findings provide a mechanistic basis for the limited Th1 response observed during L . mexicana infection and also offer a better understanding of the important role that monocytes play during infection with Leishmania .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"adaptive",
"immunity",
"immune",
"cells",
"monocytes",
"immunity",
"innate",
"immunity",
"antigen-presenting",
"cells",
"parasitology",
"immunology",
"biology",
"microbiology",
"host-pathogen",
"interaction",
"immune",
"response"
] |
2012
|
Leishmania mexicana Induces Limited Recruitment and Activation of Monocytes and Monocyte-Derived Dendritic Cells Early during Infection
|
Allele-specific DNA methylation ( ASM ) is well studied in imprinted domains , but this type of epigenetic asymmetry is actually found more commonly at non-imprinted loci , where the ASM is dictated not by parent-of-origin but instead by the local haplotype . We identified loci with strong ASM in human tissues from methylation-sensitive SNP array data . Two index regions ( bisulfite PCR amplicons ) , one between the C3orf27 and RPN1 genes in chromosome band 3q21 and the other near the VTRNA2-1 vault RNA in band 5q31 , proved to be new examples of imprinted DMRs ( maternal alleles methylated ) while a third , between STEAP3 and C2orf76 in chromosome band 2q14 , showed non-imprinted haplotype-dependent ASM . Using long-read bisulfite sequencing ( bis-seq ) in 8 human tissues we found that in all 3 domains the ASM is restricted to single differentially methylated regions ( DMRs ) , each less than 2kb . The ASM in the C3orf27-RPN1 intergenic region was placenta-specific and associated with allele-specific expression of a long non-coding RNA . Strikingly , the discrete DMRs in all 3 regions overlap with binding sites for the insulator protein CTCF , which we found selectively bound to the unmethylated allele of the STEAP3-C2orf76 DMR . Methylation mapping in two additional genes with non-imprinted haplotype-dependent ASM , ELK3 and CYP2A7 , showed that the CYP2A7 DMR also overlaps a CTCF site . Thus , two features of imprinted domains , highly localized DMRs and allele-specific insulator occupancy by CTCF , can also be found in chromosomal domains with non-imprinted ASM . Arguing for biological importance , our analysis of published whole genome bis-seq data from hES cells revealed multiple genome-wide association study ( GWAS ) peaks near CTCF binding sites with ASM .
Evidence from genome-wide association studies ( GWAS ) and cross-species comparisons suggests that many inter-individual phenotypic differences result from genetic variants in non-coding DNA sequences . Thus a major challenge in the post-genomic era is to define the mechanisms by which non-coding sequence polymorphisms and haplotypes result in differences in biological phenotypes . One hypothesis comes from recent work that has revealed strong cis-acting influences of simple nucleotide polymorphisms ( SNPs ) , and the haplotypes in which these SNPs are embedded , on epigenetic marks , leading to allele-specific DNA methylation ( ASM ) and allele-specific chromatin structure [1] . Historically , ASM has been most intensively studied in the context of genomic imprinting - a parent-of-origin dependent phenomenon that affects about 80 human genes . Mechanistic principles that have emerged from studying imprinted domains include the presence of discrete small ( one to several kb ) DNA intervals , called differentially methylated regions ( DMRs ) , which show strong parent-of-origin dependent asymmetry in CpG methylation between the two alleles and which control the allele-specific expression ( ASE ) of one or more nearby genes in cis , in some cases via methylation-sensitive binding of the insulator protein CTCF [2]–[6] . While genomic imprinting is a potent mode of gene regulation , it affects fewer loci than the more recently recognized phenomenon of non-imprinted haplotype-dependent ASM . Mapping haplotype-dependent ASM , and related phenomena such as ASE , will be useful for interpreting the biological meaning of statistical peaks from GWAS ( and non-coding variants from post-GWAS exome sequencing studies ) , as bona fide regulatory sequence variants can reveal their presence by conferring a measurable epigenetic asymmetry between the two alleles . However as yet there have not been many insights to the molecular mechanisms underlying non-imprinted ASM . This situation raises an interesting question – could any principles from studies of genomic imprinting also be relevant for understanding haplotype-dependent , non-imprinted , ASM ? To begin to address this issue , and to identify and characterize new examples of loci with imprinted and non-imprinted ASM , we have searched for “index regions” in the human genome showing strong and highly recurrent ASM and used these locations as starting points for intensive local mapping of DNA methylation patterns in multiple human tissues . Here we describe these data for new examples of loci with imprinted and non-imprinted ASM , which reveal epigenomic features that are shared between these two allele-specific phenomena .
The MSNP procedure , an adaptation of SNP arrays for detecting ASM , was described in our initial report of haplotype-dependent ASM in human tissues [7] . As a starting point for comparing the structures of chromosomal domains with imprinted ( parent-of-origin dependent ) versus non-imprinted ( haplotype dependent ) ASM we applied higher resolution MSNP to several human tissue types from multiple individuals and identified 4 additional SNP-tagged StyI or NspI restriction fragments with CpG dinucleotides in HpaII sites that showed strongly asymmetrical methylation between the two alleles in multiple heterozygous samples ( examples in Figure S1 ) . These index fragments are tagged by SNPs rs1530562 between the STEAP3 and C2orf76 genes ( chromosome band 2q14 ) , rs2346019 near the VTRNA2-1 vault family small RNA gene ( 5q31 . 1 ) , rs2811488 between the C3orf27 and RPN1 genes ( 3q21 ) , and rs2302902 in the ELK3 gene ( 12q23 . 1 ) . We used Sanger bisulfite sequencing ( bis-seq ) of amplicons containing these SNPs and multiple ( non-polymorphic ) adjacent CpG dinucleotides to confirm ASM in heterozygous individuals in a panel of primary tissues chosen for relevance to complex diseases and representation of multiple cell lineages: peripheral blood ( PBL ) , whole fetal and adult lung and adult bronchial epithelial cells , adult liver , adult brain ( cerebral cortical grey matter ) , placenta ( chorionic villi taken from near the fetal surface ) , human mammary epithelial cells ( HMEC ) , and sperm . These analyses showed that the ASM was reproducible for all 4 index fragments , with the allelic asymmetry in methylation affecting multiple ( non-polymorphic ) CpGs around and including the index HpaII sites ( Fig . 1 and Figs . S2 , S3 , S4 ) . Among these 4 regions , moderate ASM was observed in the ELK3 gene , while the other index regions showed stronger allelic asymmetries . ASM was seen in two or more tissues for each of the index regions except the C3orf27-RPN1 region , for which the index amplicon revealed highly tissue-specific ASM , which was very strong in the large majority of human placentas examined ( 33/34 cases; all tested using tissue from the fetal side of the organ , consisting of the free chorionic villi without maternal decidua; Table 1 ) , but absent in the other tissues tested ( Fig . 1 , Table 1 and Figure S3 ) . We next asked whether the ASM in each of these 4 regions was due to parental imprinting or , alternatively , to cis-acting effects of the local DNA sequence or haplotype . Sanger and bis-seq in a total of 198 and 129 individuals with all 3 possible genotypes , plus HpaII-pre-digestion/PCR/RFLP assays [7] in 76 and 30 heterozygotes , respectively , showed unequivocally that ASM both in the STEAP3-C2orf76 intergenic region and in the ELK3 intragenic region is haplotype-dependent: the simple genotypes at nearby SNPs consistently predicted the methylation status of 10 and 6 clustered CpGs , respectively , in these two index amplicons ( Fig . 1 and Table 1 ) . The ASM at both loci showed some tissue-specificity: for the STEAP3-C2orf76 intergenic amplicon we found strong ASM in multiple tissues including brain , placenta , liver , HMEC , and PBL; with sperm DNA showing biallelic hypermethylation , while for the ELK3 intergenic amplicon we found strong ASM in PBL and weak ASM in liver , with HMEC DNA showing biallelic hypermethylation ( Fig . 1 and Table 1 ) . In contrast to the results for the STEAP3-C2orf76 and ELK3 index regions , Sanger and bis-seq in 192 and 325 individuals with all 3 possible genotypes and restriction pre-digestion/PCR/RFLP assays in 70 and 75 heterozygotes for the C3orf27-RPN1 and VTRNA2 index regions respectively showed that either allele could be relatively hypermethylated with the genotype of the index SNP having no predictive value - a situation suggestive of imprinting , in which methylation is dictated by parent-of-origin , not by haplotype ( Table 1 ) . To test for bona fide parental imprinting at these two loci we analyzed trios of maternal and paternal PBL and placenta DNA . In each of 13 trios informative for the C3orf27-RPN1 index SNP and in each of 10 trios informative for the VTRNA2-1 index SNP we found relative hypermethylation of the maternal allele in the placental DNA ( p = 0 . 000311 for parental imprinting of the C3orf27-RPN1 locus and p = 0 . 001565 for the VTRNA2 locus , Chi-Square test; Fig . 2 ) . While this manuscript was in preparation Treppendahl et al . reported ASM at VTRNA2-1 in human hematopoietic cells [8] but imprinting was not tested in that study . Imprinting has not been previously reported in the C3orf27-RPN1 region . As noted above , in the RPN1 downstream DMR the ASM was restricted to placenta DNA samples , and in both loci sperm DNA showed biallelic hypomethylation , suggesting that the methylation imprint is acquired in the oocyte or early post-zygotically . The parent-of-origin dependence of the ASM in this region indicates that at least the initial “signal” for the imprint must be gametic , not somatic . However , we do not yet know the pattern of methylation in oocytes , so whether the densely methylated pattern of the maternal allele is established in the oocyte , or acquired early post-zygotically , remains to be determined . By analogy with other imprinted genes [9] , [10] , the placenta-specific methylation imprinting of the RPN1 3′ region could be relevant to placental and fetal growth , so for planning future studies it is relevant to ask whether imprinting of this region also occurs in mice . The mouse genome shows conservation of synteny with regard to the linkage and order of the Rab7-Rpn1-Gata2 genes , but an orthologue of human C3orf27 ( which is located between the human RPN1 and GATA2 genes ) is not present in this region of the mouse genome . Moreover , the full-fledged CGI at the 3′ end of the RPN1 gene in humans is not present in this position of the mouse Rpn1 gene . Despite this lack of conservation of the CGI , and indeed the lack of conservation of the DNA sequence in this region , we tested for ASM in a CG-rich sequence located immediately downstream of Rpn1 in the mouse , in a position analogous to the human DMR . Using extraembryonic tissues ( yolk sac and placenta ) from reciprocal inter-strain mouse crosses we found that CpG methylation in the orthologous region is somewhat allele-specific , but with much less allelic asymmetry than the human locus ( Figure S5A ) . The methylation patterns in the mouse yolk sac and placenta are most simply explained by weak parental imprinting superimposed on a cis-acting haplotype effect . After accounting for the cis-effect ( CAST allele more methylated than B6 allele ) , the direction of the weak imprinting matches that in humans , with the maternal allele relatively hypermethylated ( Figure S5B ) . The main finding from this analysis - that the parental imprint is much weaker in mice - suggests that non-conserved , human-specific , sequences are important for establishing or stabilizing the methylation imprint in this chromosomal region . A well-studied group of growth-regulating genes in fact show conserved imprinting in the placentas of humans and mice [10] , [11] , but the poor conservation of imprinting of the RPN1 downstream region is not surprising , as gene regulation in the placenta is known to be rapidly and continually evolving in mammals along with changes in placental anatomy [12] , [13] . Given that genomic imprinting appears to have evolved in mammals with placentation , a prediction is that the C3orf27-RPN1 intergenic DMR , and its associated RNA transcripts , may play a “human-specific” role in placental growth and development that is not conserved in mice . Overall , these data add two new loci to the current list of approximately 80 human genes with robust methylation imprinting characterized to date . To compare the structures of chromosomal domains with imprinted versus non-imprinted ASM we used high throughput bisulfite PCR ( Fluidigm AccessArray ) with sample bar-coding , followed long-read 454 Pyrosequencing of amplicons bracketing 3 of the index regions; two with imprinted ASM ( VTRNA2-1 and C3orf27-RPN1 ) and one with non-imprinted/haplotype dependent ASM ( STEAP3-C2orf76 ) . To maximize the information from this approach we designed the amplicons to contain SNPs with heterozygosities ≥0 . 2 and to span DNA segments containing ≥4 CpG dinucleotides . For VTRNA2-1 region , 6 additional amplicons covering 3 CpGs were included . The long-read sequencing platform was optimal for analyzing ASM as it allowed us to capture the CpG methylation pattern and SNP genotype in each read , with no ambiguity as to phase , while the primer bar-coding strategy facilitated the analysis of a large series of individuals , with DNA samples from various tissue types including placenta , PBL , lung , liver , brain , heart , peripheral blood mononuclear cells ( PBMC ) and polymorphonuclear leukocytes ( PMN ) . The resulting dataset for the 3 chromosomal regions consisted of 104 amplicons analyzed in 96 biological samples , totaling 354 Mb of bis-seq , with a mean depth of 90 sequences per sample per amplicon and with a mean read length of 304 bp . Also included are data from >3500 full-length Sanger bis-seq reads of from 300 to 500 bp , which we carried out on multiple amplicons and tissue samples to supplement and verify the 454 data . Our strategy allowed us to score net and allele-specific methylation for each informative amplicon over multiple CpGs with complete genetic phase information in each read . The data are summarized by color-coding for ASM ( difference in percent methylation of allele A versus allele B averaged for heterozygous samples by tissue ) in Figure 3 and Figure 4 . These “ASM heat maps” illustrate the amplicons for each chromosomal region , aligned to UCSC genome browser tracks [14] , showing CpG islands ( CGIs ) , histone modifications and CTCF binding sites . The most striking finding from this long-range analysis , and from additional intensive short-range mapping by Sanger bis-seq ( Figs . 5 and 6A ) , is the discrete nature of all three DMRs; for all three regions , two with imprinted ASM and one with non-imprinted haplotype-dependent ASM , the allelic asymmetry in DNA methylation is restricted to one or two adjacent amplicons , with all three DMRs being less than 2 kb in length ( Figs . 5 and 6 and Table 1 ) . Within each of the regions examined , the flanking DNA outside of the DMRs showed varying levels of net CpG methylation , without asymmetry between the two alleles . Our mapping does not rule out additional DMRs farther away , but for the STEAP3-C2orf76 region , which shows haplotype-dependent ASM , any DMRs farther away would be in separate haplotype blocks and therefore be independently regulated domains . We additionally carried out short-range mapping of ASM around the ELK3 intragenic index fragment , which showed that it too is discrete and <2 kb in size ( Figure S6 ) . Alignment of our methylation sequencing data with ChIP-Seq data from ENCODE and related projects , as displayed on the UCSC genome browser [14] , showed that for both of the 2 imprinted domains , and for the non-imprinted STEAP3-C2orf76 intergenic region , though not for the ELK3 intragenic DMR , the DMRs with ASM overlapped precisely with empirically determined CTCF binding sites ( Figs . 5 , 6A and Figure S6 ) . To ask whether there are additional examples of small discrete DMRs with strong and recurrent haplotype-dependent ASM that overlap with CTCF binding sites , we returned to an interesting locus with haplotype-dependent ASM and ASE that we had characterized in our initial report on this phenomenon [7] , namely , the CYP2* gene cluster in chromosome band 19q13 . 2 . As shown in Figure 7 , the intragenic DMR in CYP2A7 , tagged by SNP rs3815710 , in fact shows discrete borders and precise co-localization with a CTCF binding site . Another DMR in this large gene cluster , tagged by SNP rs3844442 and located between 2 CYP2* pseudogenes [7] , overlaps with a weaker CTCF binding site . Maps of each of these regions with haplotype-dependent ASM are shown aligned to haplotype blocks in Figure S7 . To test directly for allele specific binding of CTCF to the STEAP3-C2orf76 DMR , we carried out CTCF ChiP on a freshly obtained blood sample from a heterozygous individual that had shown strong sequence dependent ASM by previous bis-seq . By comparison with the IgG control IP , end-point PCR amplification of a region of the STEAP3-C2orf76 DMR covering the index SNP rs1530562 confirmed CTCF binding to this region , and Sanger sequencing of the PCR product confirmed that only the hypomethylated allele was present in the CTCF enriched product , while both alleles were observed in the PCR product from the input non-immunoprecipitated DNA ( Fig . 6B ) . Thus , CTCF binds specifically to the hypomethylated allele of the STEAP3-C2orf76 intergenic DMR . To test for a possible genome-wide non-random association of ASM with CTCF sites , we used previously published ASM site lists [15] from whole genome bis-seq of a single cell line , namely H1 hES cells [16] . Chen et al . [15] found 83 , 000 CpG sites with ASM , representing 14% of the total CpGs tested ( a high percentage that partly reflects incomplete filtering-out of CpG SNPs in their analysis . ) Based on their analysis , we first asked whether the ASM regions that we have described here were also detectable in H1 hES cells . The multi-tissue ASM we identified in the STEAP3-C2orf76 and VTRNA2-1 index regions was also present in the H1 hES cells , while the tissue-specific ASM found in C3orf27-RPN1 was not present . In addition , the ASM we found in the CYP2A7 region was not mirrored in H1 hES cells , consistent with tissue-specificity for liver . We then looked for CTCF binding peaks in 500 bp windows centered on each of the CpGs tested by Chen et al . for ASM , using ChIP-seq data generated in H1 hES cells by the Broad/MGH , Hudson and UT-Austin ENCODE groups . We found 13 , 485 CTCF binding sites in the vicinity of the informative CpG sites . Most of these CTCF binding sites were in intragenic or intergenic regions ( 47% ) ; fewer were in gene promoter regions ( 5% ) or next to CGIs ( 7% ) . Overall , CTCF sites were far more frequently associated with unmethylated CpGs than with fully methylated , partially methylated or ASM CpGs ( Figure S8 ) . However , when considering only CpG sites in gene promoter regions ( within 1 kb upstream and downstream of transcription starting sites ) , we find a 1 . 7-fold increase in the percent of ASM CpGs located near CTCF binding sites ( that is , in a 500 bp window centered on the CpGs ) , compared to fully methylated CpG sites , but not unmethylated CpGs or partially methylated CpGs . When considering only the 99 ASM CpGs located in CGI-associated gene promoter regions ( promoter CGIs ) 40 . 4% were near CTCF sites , while less than 30% of fully methylated and partially methylated and only 23% of completely unmethylated CpGs in promoter CGIs were near CTCF sites ( Figure S8A ) . Conversely , the distribution of the different classes of CpG sites ( ASM , fully methylated , partially methylated , fully unmethylated ) with the presence of CTCF binding sites in a 500 bp window centered on the CpG sites compared to their distribution without CTCF binding sites in such windows , reflected this association between ASM and CTCF binding sites in promoter regions and CGIs , with an increased proportion of ASM in these regions ( Figure S8B ) . Next , to more confidently enumerate the CTCF sites overlapping or very close to regions of ASM , we looked specifically for CTCF binding sites with ≥2 ASM CpGs ( as defined by Chen et al . [15] ) in a 1 kb window centered on the CTCF peak . A total of 158 CTCF sites met this criterion; among them , 142 were gene associated , with at least one named gene within 100 kb of the CTCF sites . Of these genes , we were able to cross-tabulate 75 to the current GWAS catalog ( http://www . genome . gov/gwastudies ) . Interestingly , 61 of them were associated with a major human trait or disease susceptibility signal , and for 25 genes the CTCF binding sites , ASM and the GWAS peak SNP were all located in the same haplotype block - suggesting that these GWAS signals reflect a bona fide ( functionally active ) regulatory haplotype ( Table S4 , including our annotations for CpG SNPs , and examples in Fig . 6C ) . Thus , while CTCF sites are not enriched overall near regions of ASM , ASM-associated CTCF sites have a somewhat different distribution than other CTCF sites and , more importantly , are often associated with GWAS peaks for human traits and diseases . One of the biologically important consequences of ASM is allele-specific RNA expression ( ASE ) . We previously demonstrated strong haplotype-dependent ASE of the CYP2A7 gene [7] , and we were interested to ask whether ASE could be detected in the other chromosomal regions with ASM analyzed in this report . Assays comparing the representation of SNPs in genomic versus cDNA PCR products showed a definite and recurrent bias in allele-specific mRNA expression of the C2orf76 gene , associated with the STEAP3-C2orf76 DMR ( Fig . 8 ) . The untranslated AK097792 RNA transcript , associated with RPN1-C3orf27 DMR also showed a recurrent bias in ASE: from 13 tested individuals informative for ASE , 6 showed a definite but not complete allele-specific bias , four had complete ASE and three were biallelically expressed ( Fig . 8 and Fig . S3 ) . In each of 6 informative samples with definite or complete ASE , and data for ASM , the relatively hypermethylated allele was the repressed one , suggesting a functional link between ASE and ASM at this locus . Interestingly , current ENCODE data support our findings from RT-PCR of a long non-coding RNA ( lncRNA ) traversing the DMR , and also suggest that a micro-RNA , as yet unnamed , arises from very close to this DMR ( Figure S3 ) . The absence of transcribed SNPs prevented ASE assays for VTRNA2-1 , but as was found by Treppendahl et al . [8] using a different cell line , our assays for activation of this gene by the demethylating drug 5aza-dC provided evidence for methylation-dependent expression of VTRNA2-1 in HL60 , ML3 , and Jurkat leukemia cells ( Figure S9 ) . Additional analyses showed that the STEAP3 gene is biallelically expressed in PBL and liver , while TGFBI and SMAD5 , flanking VTRNA2-1 , exhibit only a slight allelic expression bias in PBL , as do RPN1 and C3orf27 in placental samples ( data not shown ) .
ASM is a well-studied hallmark of genomic imprinting . However , in the past several years this same type of allelic asymmetry has been found in and around many non-imprinted genes , both in normal human tissues and in F1 progeny of inter-strain mouse crosses [1] , [17] . In contrast to imprinted ASM , in which the methylation status is dictated by the parent of origin of the allele , when ASM occurs at non-imprinted loci it is usually haplotype-dependent . Thus non-imprinted ASM is controlled by genetic polymorphisms , such as SNPs , indels and copy number variants ( CNVs ) , which act in cis to set up the allelic asymmetries . The mechanism ( s ) by which this occurs are not yet understood . Research on this topic is important because haplotype-dependent ASM , with its functional correlate of genotype-specific gene expression manifesting as ASE and expression quantitative trait loci ( eQTLs ) , is a major route by which inter-individual genetic differences in non-coding sequences can lead to differences in phenotypes , including disease susceptibility . Here we have described new examples of genes and regulatory sequences with imprinted and non-imprinted ASM , which have revealed epigenomic features that are shared between these two epigenetic phenomena . Genes that are regulated by parental imprinting frequently have important roles in cell proliferation and tissue growth , so it is interesting that the VTRNA2-1 gene may have a suppressor role in human acute myeloid leukemias and other cancers [8] , [18] , and that vault RNAs are strongly upregulated during ES cell differentiation [19] . Similarly , given the clear importance of multiple imprinted genes in placental and fetal growth [10] , [11] , and the accumulating data for rare but recurrent alterations of maternal imprints in offspring conceived by assisted reproductive technologies , the imprinted C3orf27-RPN1 chromosomal domain , including the lncRNA corresponding to the AK097792 EST , as well as other RNAs , such as the as yet unnamed micro-RNA precisely localized to this region by data from the ENCODE/CSH project ( Figure S3 ) will be interesting for future studies of placental growth and function . The loci with non-imprinted ASM that we have described here are also potentially relevant to human traits . CYP2A7 is located immediately adjacent to and is highly homologous to the CYP2A6 gene , and both genes are predicted to encode cytochrome P450 proteins responsible for coumarin and nicotine metabolism and smoking behavior [20] . The C2orf76 gene encodes a protein with unknown function , but a non-synonymous SNP in this gene was tentatively implicated in HIV susceptibility in African-Americans by GWAS at a nominal p-value of 0 . 0001 [21] . As we have previously pointed out [1] , a major utility of mapping non-imprinted , haplotype-dependent ASM is to reveal bona fide regulatory haplotypes underlying human complex disease susceptibility . In this regard , our current analysis of prior whole genome bis-seq data has already yielded an interesting set of additional loci in which GWAS signals for important human traits and disease susceptibility coincide with both ASM and CTCF sites . Lastly , our data lay an essential groundwork for testing mechanisms of haplotype-dependent ASM . One feasible application will be in determining the minimal sequence requirements that allow local haplotypes to dictate methylation patterns in cis: based on our mapping data it should be possible to design allelic series of large bacterial artificial chromosomes ( BACs ) containing the chromosomal regions described here , which can be subjected to targeted site-directed mutagenesis , including the CTCF sites as well as haplotype-specific sequences outside of these sites , and then analyzed for cis-effects on CpG methylation patterns in vivo in BAC-transgenic mice .
Tissues from human organs were obtained , without patient identifiers , from the Molecular Pathology Shared Resource of the Herbert Irving Comprehensive Cancer Center and blood samples from volunteers were obtained with informed consent under protocols approved by the internal review board of Columbia University Medical Center . MSNP assays were performed as previously described [7] . Affymetrix 250K StyI and 6 . 0 SNP arrays were used , with initial complete digestions of the genomic DNA samples with StyI ( 250K arrays ) or NspI ( 6 . 0 arrays ) , with the methylation-sensitive restriction enzyme HpaII or with its methylation-insensitive isoschizomer MspI , followed by probe synthesis and labeling as recommended by Affymetrix . The MSNP data were processed numerically as described [7] , followed by visualization of the allele-specific hybridization data in dChip [22] to identify SNP-tagged StyI or NspI restriction fragments where call conversions were made from AB with no HpaII to AA or BB with HpaII pre-digestion . Based on recurrent call conversions in two or more samples and consistency of hybridization across multiple probes on visualization of the intensity data , we selected loci for validation of ASM by bis-seq of amplicons spanning the original index SNPs . We used Sanger bis-seq of multiple clones for intensive short-range mapping of ASM in each of the 5 chromosomal regions , as well as for filling gaps in the long-read 454 bis-seq . Genomic DNA ( 0 . 5 µg ) was bisulfite converted using the EpiTect Kit ( Qiagen ) following the manufacturer's protocol . Primers were designed in MethPrimer [23] to encompass >5 contiguous CpGs and one or more SNP , and the converted DNA was amplified for regions of interest using optimized PCR conditions ( Table S1 ) . PCR reactions were performed in duplicate then pooled and purified using the Wizard SV System ( Promega ) , followed by cloning using the TOPO TA Kit ( Invitrogen ) . Multiple clones were sequenced . Primer sets spanning 50–100 kb on each side of each of the three index fragments for long-range mapping were designed as described in [24] and in part empirically , using both strands of an in silico bisulfite-converted SNP-masked genome . Primer sequences were not allowed to overlap with any SNPs having >1 percent heterozygote frequency in dbSNP130 . Amplicons were allowed to range between 200–650 bp in size , with most between 200–450 bp . CpG content was allowed to range between 0–1 CpGs per primer with zero CpGs preferred . Tm was allowed to range from 56–60°C , primer GC content is allowed to range between 35–65% , and self-complementarity , primer-primer complementarity , and hairpin formation is restricted for stable structures with melting temperatures above 46°C . Primers generating off target amplicons of size 50–2000 bp ( with two mismatches allowed ) were filtered out by scanning against a bisulfite converted genome with NCBI reverse e-PCR . From these primer sets , 41 amplicons for the VTRNA2-1 region , 27 amplicons for the STEAP3-C2orf76 region and 25 amplicons for the C3orf27-RPN1 region were picked based on minimum cut-offs for the number of CpG sites encompassed in each amplicon ( ≥4 ) , and heterozygosity frequencies of annotated SNPs between the primers ( ≥ . 2 ) . For VTRNA2-1 region , 6 additional amplicons covering 3 CpGs were included . Primer sequences are in Table S2 . Bisulfite modification was performed for 96 samples using 2 ug of genomic DNA and the Epitect 96 Kit ( Qiagen ) , following the recommended protocol . The bisulfite modified DNA was re-precipitated and concentrated using Pellet Paint ( EMD Chemicals ) . To amplify the DNA from the 96 samples , the Fluidigm 48×48 Access Array was used , with JumpStart Taq Polymerase ( Sigma ) and a 60°C touchdown for ten cycles followed by 51°C for 29 cycles of amplification . Of the resulting PCR products , one microliter of a 1∶50 dilution was used in a second barcoding PCR reaction . Fluidigm barcodes designed with 454 adapters were added to samples 1–48 and 49–96 during this PCR step . Both the 454 FLX instrument and 454Junior instrument and their emPCR and Sequencing kits ( Roche ) were used . The barcoded PCR products were pooled and purified using Agencourt AMPure XP beads and a titration was performed ( described in 454 provided protocol ) in order to find the best bead∶DNA ratio required to remove DNA fragments smaller than 300 bp . The purified PCR products were then prepared for sequencing following the 454 emulsion PCR protocol ( Roche ) . A titration of bead∶DNA ratios was also performed during this step to ensure an ideal ratio was achieved in order to obtain the best sequencing results . The amplified PCR products were then sequenced using the sequencing kits designed for each instrument . The filter settings were adjusted on the 454 sequence analysis software to a vfTrimBack Scale Factor of . 5 . The sequences were then mapped , scored for each previously identified SNP , and each allele was analyzed for percent methylation , both at individual CpGs and averaged over all CpGs in the amplicon . Of the original 99 primer pairs , 91 gave a sufficient number of reads to allow genotyping ( minimum ten reads with at least 20% of reads representing the minor allele ) , and net methylation analysis in most samples , and in at least 3 samples per tissue ( minimum 7 reads ) , as well as scoring ASM in multiple heterozygous samples . Also included were data from >3500 full-length Sanger bis-seq reads of from 300 to 500 bp performed on several amplicons and tissues . T-tests were performed to determine significant ASM , p<0 . 05 , for samples with at least 5 reads per allele . Tissue and blood cell RNA was extracted using Trizol reagent ( Invitrogen ) following the manufacturer's protocol . The RNA was then reverse transcribed to cDNA using RT Kit ( AmbionRETROscript ) with random hexamer primers . PCR primers were designed for genomic DNA and cDNA to cover a SNP in the exon of the gene of interest using Primer3 software , with the cDNA primers spanning an intron of the gene , or in the case of the AK097792 non-translated RNA , not spanning an intron but verified as amplifying only cDNA in the samples tested , by lack of PCR product in the minus-RT control reaction ( Table S3 ) . Genomic DNA and cDNA were amplified by PCR and the PCR products were Sanger sequenced from duplicate reactions . The sequence chromatograms were analyzed by comparing peak heights of the two alleles in heterozygous individuals in the genomic DNA versus cDNA . For measuring expression of the VTRNA2-1 small RNA we used a custom TaqMan assay ( Applied Biosystems ) according to the recommended protocol . Peripheral blood mononuclear cells ( PBMCs ) were isolated from whole blood following the manufacturer's instructions for FicollPaque Plus reagent ( GE ) . Chromatin Immunoprecipitation was performed on the PBMCs using the Magna-ChIP G Kit from Millipore following the suggested protocol with some slight modifications . Cell fixation was performed for six minutes . Sonication was performed using a Fisher Sonic Dismembrator at a power setting of 3 . 2 for 2 minutes total sonication time with 2 seconds sonication followed by 8 seconds recovery . Four micrograms of anti-CTCF antibody ( Millipore ) was used in an overnight incubation at 4°C with sheared chromatin and protein G Magna ChIP beads . PCR for the STEAP3-C2orf76 DMR was then performed using primers: Forward , GACAGACTCTGCTGCCACCT and Reverse , AGCAGCTTCTTCTCGGTATG . Two microliters of purified input , negative control IgG and CTCF immunoprecipitated samples were used for PCR with a touchdown from 60°C to 51°C for 38 cycles . Previously published ASM site lists from the whole genome bis-seq of H1 hESC cells were downloaded [15] , [16] . The dataset included 514 , 181 CpG sites tested for ASM by Chen et al . [15] . Of these , 79 , 365 were CpG sites classified as ASM . For this analysis , in addition to ASM sites , we grouped CpG sites without evidence of ASM into 3 categories: full methylation ( 200 , 318 CpG sites with >90% of reads with methylation at the CpG site ) , partial methylation ( 232 , 724 CpG sites with 10% to 90% of reads with methylation at the CpG site ) or complete lack of methylation ( 1 , 774 CpG sites with <10% of reads with methylation at the CpG site ) . To assess the presence of CTCF binding peaks in a 500 base window centered on each CG tested for ASM , we used ChIP-seq data generated in H1 hESC cells by the Broad/MGH ( UCSC accession number: wgEncodeEH000085 ) , Hudson alpha ( UCSC accession number: wgEncodeEH001649 ) and UT-Austin ( UCSC accession number: wgEncodeEH000560 ) ENCODE groups . We searched for CTCF binding sites present in at least one of these datasets and found 13 , 485 CTCF binding sites in the vicinity of the CpG sites included into this analysis . We performed sub-analyses in order to look specifically at CpG sites in intergenic regions or intragenic region , as well as in gene promoter regions ( 1 kb upstream and downstream of transcription starting sites of genes ) and/or close to CGIs ( within 250 bp ) , using data available from the UCSC genome browser . All p-values for these analyses were calculated using one-sided Fischer exact tests .
|
Allele-specific DNA methylation ( ASM ) is a central mechanism of gene regulation in humans , which can influence inter-individual differences in physical and mental traits and disease susceptibility . ASM is mediated either by parental imprinting , in which the repressed copy ( allele ) of the gene is determined by which type of parent ( mother or father ) transmitted it or , for a larger number of genes , by the local DNA sequence , independent of which parent transmitted it . Chromosomal regions with imprinted ASM have been well studied , and certain mechanistic principles , including the role of discrete differentially methylated regions ( DMRs ) and involvement of the insulator protein CTCF , have emerged . However , the molecular mechanisms underlying non-imprinted sequence-dependent ASM are not yet understood . Here we describe our detailed mapping of ASM across 5 gene regions , including two novel examples of imprinted ASM and three gene regions with non-imprinted , sequence-dependent ASM . Our data uncover shared molecular features – small discrete DMRs , and the binding of CTCF to these DMRs , in examples of both types of ASM . Combining ASM mapping with genetic association data suggests that sequence-dependent ASM at CTCF binding sites influences diverse human traits .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"developmental",
"biology",
"genetics",
"biology",
"genomics",
"evolutionary",
"biology",
"clinical",
"genetics"
] |
2013
|
Comparative Anatomy of Chromosomal Domains with Imprinted and Non-Imprinted Allele-Specific DNA Methylation
|
The unfolded protein response ( UPR ) in the endoplasmic reticulum ( ER ) and the cytoplasmic heat stress response are two major stress response systems necessary for maintaining proteostasis for cellular health . Failure of either of these systems , such as in sustained UPR activation or in insufficient heat shock response activation , can lead to the development of neurodegeneration . Alleviation of ER stress and enhancement of heat shock response through heat shock factor 1 ( HSF1 ) activation have previously been considered as attractive potential therapeutic targets for Alzheimer’s disease ( AD ) —a prevalent and devastating tauopathy . Understanding the interplay of the two aforementioned systems and their cooperative role in AD remain elusive . Here we report studies in human brain and tau pathogenic mouse models ( rTg4510 , PS19 , and rTg21221 ) , identifying HSF1 degradation and UPR activation as precursors of aberrant tau pathogenesis . We demonstrate that chemical ER stress inducers caused autophagy-lysosomal HSF1 degradation , resulting in tau hyperphosphorylation in rat primary neurons . In addition , permanent HSF1 loss reversely causes chronic UPR activation , leading to aberrant tau phosphorylation and aggregation in the hippocampus of aged HSF1 heterozygous knock-out mice . The deleterious interplay of UPR activation and HSF1 loss is exacerbated in N2a cells stably overexpressing a pro-aggregation mutant TauRD ΔK280 ( N2a-TauRD ΔK280 ) . We provide evidence of how these two stress response systems are intrinsically interweaved by showing that the gene encoding C/EBP-homologous protein ( CHOP ) activation in the UPR apoptotic pathway facilitates HSF1 degradation , which likely further contributes to prolonged UPR via ER chaperone HSP70 a5 ( BiP/GRP78 ) suppression . Upregulating HSF1 relieves the tau toxicity in N2a-TauRD ΔK280 by reducing CHOP and increasing HSP70 a5 ( BiP/GRP78 ) . Our work reveals how the bidirectional crosstalk between the two stress response systems promotes early tau pathology and identifies HSF1 being one likely key player in both systems .
Neurofibrillary tangles ( NFTs ) of phosphorylated tau aggregates and senile plaques of amyloid beta ( Aβ ) are the pathological hallmarks of Alzheimer’s disease ( AD ) patients . Aβ toxicity has been known to contribute to synaptic loss and cognitive impairment , the mechanism of which appears to be tau-dependent [1 , 2 , 3] . However , NFTs , not amyloid plaques , have been identified to correlate best with the severity of dementia [4 , 5] . NFTs exist primarily inside the cell , and has been associated with cellular stress responses [6 , 7 , 8] . Such response systems include unfolded protein response ( UPR ) initiated in the endoplasmic reticulum ( ER ) and cytoplasmic heat shock response initiated by heat shock factor 1 ( HSF1 ) activation . Both pathways involve transcriptional activation of the stress-response genes . A key feature of heat shock response is to induce a set of molecular chaperone proteins such as Hsp70 whose function is to correct protein folding in response to numerous cellular stresses . Likewise , UPR is initially triggered as an adaptive response to disturbances in ER homeostasis . In this sense , HSP70 a5 ( BiP/GRP78 ) , a major ER chaperone Hsp70 , has been found to attenuate ER stress by activating sensors of transmembrane ER stress , such as protein kinase RNA-like endoplasmic reticulum kinase ( PERK ) , via direct binding [9] . However , sustained chronic UPR activation as a result of unresolved ER stress can eventually trigger cell death by inducing pro-apoptotic proteins such as C/EBP homologous protein ( CHOP ) primarily through the PERK/eIF2α/ATF4 pathway [10 , 11 , 12] . A close correlation between ER stress markers and NFTs has been consistently reported in human tauopathies such as AD and frontotemporal dementia [7 , 13 , 14] . Ho et al . , 2012 has shown that ER stress can cause tau hyperphosphorylation in primary cultured neurons [15] . However , it is unknown whether and how ER stress causes tau phosphorylation , and vice versa . While these two stress systems ( i . e . heat shock response , UPR ) have been originally considered to be individually triggered by distinct stressors , recent studies have begun to highlight the importance of heat shock responses in relieving ER stress in non-neuronal cells [11 , 16 , 17 , 18] . Nevertheless , how ER stress affects HSF1-mediated stress response is poorly understood , particularly in the contexts of neurons and tauopathy . We recently identified aberrant HSF1 degradation via ubiquitin proteasome system as an important mechanism underlying synucleinopathy [19] . Synucleinopathy has been suggested to be pathogenetically linked with tauopathy as reflected by their frequent co-occurrence in neurodegenerative diseases and the synergistic interaction of tau and synuclein [20 , 21 , 22] . Tauopathy is characterized by a build-up of tau aggregates , and thus the notion that HSF1 could also be degraded by tau aggregation stands as an intriguing possibility . In an attempt to understand human AD tau pathology , experimental murine tauopathy has been generated by introducing mutations in the human tau gene causing frontotemporal dementia and parkinsonism linked to chromosome 17 ( FTDP-17 ) such as P301L , P301S , and ΔK280 [23 , 24 , 25] . Presence of UPR activation in the tau transgenic mouse models has not been clearly defined . Spatara et al . , 2010 reported that PS19 mice harboring human tau P301S variant did not show any signs of UPR activation , whereas Abisambra et al . , 2013 provided some evidence of UPR activation in rTg4510 mice overexpressing human tau carrying the P301L mutation [3 , 26] . There is a study showing that Tau ΔK280 mutation aggregates faster than any other single missense mutation [27] . Tau ΔK280 mutation has never been discussed in terms of its involvement of UPR activation . Here , we examined the impact of ER stress-induced UPR activation on HSF1 protein and vice versa in promoting aberrant tau pathology in AD . We investigated how UPR effector proteins such as CHOP and HSP70 a5 ( BiP/GRP78 ) regulate HSF1 to potentiate a vicious cycle active in the cellular tauopathy model ( N2a-TauRDΔK280 ) . Our work highlights that HSF1 loss may constitute a mechanistic connection between ER stress and tau hyperphosphorylation in tau pathogenesis .
To determine whether expression of HSF1 protein and UPR marker proteins was altered in the mouse tauopathy models , we looked at the brains of PS19 ( tau P301S ) mouse and rTg ( tauP301L ) 4510 mouse overexpressing P301S and P301L mutant tau , respectively . It was previously reported that synaptic function was impaired in 3 month-old PS19 mouse before NFTs that consist of insoluble hyperphosphorylated tau developed at 6 months of age [25] . In the brain of 4 month-old PS19 mouse , we found that overexpressed mutant tau was not hyperphosphorylated at Ser202/Thr205 ( p-Tau , detected using the AT8 antibody ) ( Fig 1A P < 0 . 01 , n = 5 ) . Upon activation of the UPR signal , PERK is phosphorylated at Thr980 ( p-PERK ) . While PERK was not activated , about 30% of HSF1 protein was lost in the brain of PS19 mouse before tau hyperphosphorylation at a presumably later stage ( Fig 1A , P < 0 . 05 , n = 5 ) . Insoluble tau filaments were previously detected in 4 month-old rTg4510 mouse [24] . We observed that overexpressed mutant tau was highly phosphorylated in the brain of rTg4510 mouse , about 2 . 7-fold increase in the levels of p-Tau normalized to total tau protein ( Tau46 ) , at 4 months of the same age with PS19 mouse above ( Fig 1B , P < 0 . 001 , n = 5 ) . Activated PERK causes phosphorylation on Ser51 of the α subunit of eukaryotic translation initiation factor 2 ( p-eIF2α ) . Together with p-PERK and p-eIF2α elevation , increased expression of pro-apoptotic protein CHOP suggested later apoptotic stage of UPR chronically activated in the brain of rTg4510 mouse ( Fig 1B , p-PERK , P < 0 . 001 , n = 5 ) [11 , 12] . We detected a dramatic loss of HSF1 expression levels , about 60% reduction in rTg4510 mouse brain in which PERK ( or UPR ) was activated ( Fig 1B , P < 0 . 01 , n = 5 ) . Furthermore , we looked at the brain of rTg21221 transgenic mouse that overexpresses non-aggregating wild-type human tau since no tau mutations have been identified in AD [28] . In the brain of rTg21221 , while CHOP activation was induced at 4- and 8 months of age , about 30% of HSF1 protein was significantly lost at 8 months of age when compared to control group of mice ( Fig 1C , P < 0 . 001 , n = 7 ( control ) , n = 5 ( rTg21221 ) ) . The Braak NFT staging system is used to classify the anatomical distribution of AD-type NFTs according to stages ( I/II , entorhinal-; III/IV , limbic-; and V/VI , neocortical-stage ) [29] . To identify the sequential relationship of molecular events in AD brain , we determined the expression levels of HSF1 protein and UPR marker proteins in the frontal lobes of 19 human postmortem brain specimen , spanning different Braak NFT stages . Distribution of HSF1 protein mostly in the cytoplasm suggested its inactivation in aged human brain since HSF1 is a nuclear transcriptional factor likely to be located in the nucleus when it is activated ( Fig 1D ) . More importantly , we did observe gradually decreased expression of HSF1 protein that occurred even before formation of NFTs composed of hyperphosphorylated tau ( Fig 1D–1F ) . While PERK activates eIF2α during UPR activation upon ER stress , other serine kinases such as general control nonderepressible 2 ( GCN2 ) can also phosphorylate eIF2α in response to other stresses including amino acid deprivation . In incipient AD stages III/IV ( before the frontal cortex starts to form NFTs ) , we observed about 58% reduction in HSF1 protein and slightly increased p-eIF2α without PERK activation ( Fig 1E and 1F ) . UPR was activated in later stage V patients diagnosed with early onset AD ( EAD ) and AD patients in stage VI , as reflected by marked hyperactivation of PERK and eIF2α ( Fig 1E and 1F ) . About 59% of total HSF1 protein in Stages III/IV was further lost in stages V/VI when UPR was activated . It should be noted that stage V patients diagnosed with EAD demonstrated significantly greater p-PERK and less HSF1 protein expression when compared to AD patients in stage V ( Fig 1E ) . Taken together , these results identify that a dramatic loss of HSF1 protein is an early and progressive event that may precede PERK activation and NFTs formation in murine tauopathy and human AD . Next , we investigated if HSF1 protein loss altered the expression levels of UPR marker proteins and tau phosphorylation in the brain of HSF1 haploinsufficient mouse ( HSF1+/- ) . Phosphorylation at Thr212/Ser214 ( p-Tau , detected using the AT100 antibody ) was not detected in the hippocampus of 2 month-old HSF1+/- and WT ( Fig 2A ) . We did not observe any significant change in the expression levels of p-Tau and CHOP in the hippocampus of HSF1+/- at 2- and 6-months of age ( Fig 2A and 2E and S1A Fig ) . However , PERK and CHOP protein were highly activated in the hippocampus of 9 month-old HSF1+/- , although p-eIF2α was not elevated ( Fig 2B , p-PERK ( Thr980 ) , P < 0 . 001; CHOP , P < 0 . 01 , comparing to 9 month-old WT hippocampus , n = 6 ) . By immunohistochemistry , we found that activated PERK ( p-PERK , green ) largely co-localized with of increased p-Tau ( Ser202/Thr205 , red ) in the hippocampus of HSF1+/- at 13 months of age ( Fig 2C ) . Of note , a dramatic upregulation of p-Tau ( detected by both AT8 and AT100 antibodies ) was strongly found in the hippocampus of 9 month-old HSF1+/- by western blot ( Fig 2D and 2E , p-Tau ( Ser202/Thr205 ) , P < 0 . 01; p-Tau ( Thr212/Ser214 ) , P < 0 . 05 , n = 6 ) . These changes were not detectable in the whole brain lysates of 9 month-old HSF1 +/- ( S1B Fig ) . Tau isolated from PHFs in human AD brain has been reported to contain ~60- , ~64- , and ~68- kDa tau isoforms [30] . In our study , the strongest immunoreactive band for phosphorylated tau protein in HSF1+/- was detected at ~68 kDa ( Fig 2D ) . In the aged HSF1+/- hippocampus , there were two major p-Tau bands at ~64- and ~68- kDa with a minor band at ~72 kDa , similar to the pattern observed in human tauopathy brain [31] ( Fig 2D ) . These are considered to be low molecular weight tau , ranging in 50–68 kDa on SDS-PAGE [32 , 33] . Sarkosyl insolubility assay is used to isolate tau paired helical filaments found in NFTs from AD brain [34] . In the sarkosyl-insoluble fraction from the hippocampus of aged HSF1+/- , we found high molecular weight tau isoform , approximately 110 kDa tau , in addition to a major band of ~68 kDa [35] ( Fig 2D and S1C Fig ) . Paired helical filaments from AD brain bind to Thioflavin S [36] . By double staining using thioflavin S and Tau antibody ( Tau46 ) , aggregates of hyperphosphorylated tau were found in the hippocampus of 13 month-old HSF1+/- , not in WT control ( Fig 2F and S2 Fig ) . Diffuse plaques that showed strong positive signals for thioflavin S consisted of tau proteins ( Fig 2F ) . Thioflavin S staining also revealed globose-type NFTs and diffuse plaques in the cortical areas adjacent to the hippocampus ( Fig 2F ) . Although tau protein is most abundant in axons , abnormal modifications of tau such as hyperphosphorylation can lead to redistribution of tau from neuronal processes to the soma where it likely forms toxic oligomers or aggregates [37] . Hyperphosphorylation of tau has been identified to mislocalize tau to dendritic spines in neurons in AD models [28] . Mouse hippocampal tissues were processed by crude cytoplasmic extraction and synaptosomal fractionation that enriches both presynaptic and postsynaptic compartments . Both cytoplasmic- and synaptic- tau were found to increase in the hippocampus of 13 month-old HSF1+/- compared to WT control hippocampus ( Fig 3A ) . We found increase in the average of the relative expression of p-Tau ( Ser202/Thr205 ) in the cytoplasm and synaptosomal membrane of the hippocampus of 6 month-old HSF1+/- compared to that of 6 month-old WT ( Fig 3A , cytoplasmic p-Tau ( Ser202/Thr205 ) , P < 0 . 05 ) . Clearly , expression levels of p-Tau ( Ser202/Thr205 ) were significantly upregulated in the both cytoplasm and synaptosomal membrane of the hippocampus of 14- and 25- month-old HSF1+/- compared to that of 14 month-old WT ( Fig 3B ) . We observed that loss of HSF1 protein was more severe in the brains of rTg4510 and Braak stages V/VI subjects where PERK was activated when compared to that of age-matched PS19 and stages III/IV subjects that did not show PERK activation , respectively ( Fig 1 ) . Thus , we asked if UPR activation was involved in HSF1 protein loss in rat primary cultured neurons . Thapsigargin triggers ER stress by inhibiting Ca2+-transporting , ATPase mediated uptake of calcium ions into the sarcoplasmic reticulum . It is well documented that rapamycin , an inhibitor of mTOR ( mammalian target of rapamycin ) , represses ER stress in various cell types [38] . Rapamycin seemed to act as a reliable inhibitor of CHOP expression both in the presence and absence of thapsigargin in primary cortical neurons ( Fig 4A–4C , ↓ , P < 0 . 001 , n = 3–4 ) . We demonstrated that thapsigargin triggered pro-apoptotic UPR activation that was attenuated by co-treatment of rapamycin ( Fig 4A–4C ) . Of note , thapsigargin treatment led to ~54% reduction in the HSF1 protein levels in primary neurons ( P < 0 . 05 , n = 4 ) , which was nearly blocked by rapamycin ( Fig 4B and 4C ) . CHOP activation could be further enhanced by salubrinal treatment which inhibits activity of protein phosphatase 1 and prolongs eIF2α phosphorylation . In contrast to rapamycin , salubrinal treatment further aggravated HSF1 loss from thapsigargin-induced ER stress in primary neurons ( Fig 4B , ↓↓ ) . Autophagy activity , as determined by LC3 II accumulation , was increased in thapsigargin-treated neurons ( Fig 4B , ↓ ) . Treatment of autophagy-lysosomal blocker , NH4Cl , inhibited thapsigargin-induced HSF1 loss , revealing HSF1 protein degradation through autophagy-lysosomal system ( Fig 4B , a solid line indicated ) . Although direct HSF1 activator has not been identified , we previously reported that resveratrol could prevent HSF1 degradation from proteotoxic stress [19] . In the current study , resveratrol attenuated thapsigargin-induced HSF1 protein degradation in primary neurons ( Fig 4B , a dashed line indicated ) . Tunicamycin is another chemical UPR inducer that causes protein unfolding by blocking the glycoprotein synthesis pathway . Tunicamycin also reduced HSF1 protein expression in primary cultured cortical and hippocampal neurons ( Fig 4D ) . We then sought in vivo evidence of thapsigargin-induced HSF1 degradation . Five month-old C57BL/6 mice were injected with thapsigargin ( i . p . 2 mg/kg ) or PBS containing 10% DMSO . Remarkably , 6 hrs later , drastic HSF1 degradation , about 50% reduction in HSF1 protein expression , was detected in the mouse brain injected with thapsigargin ( Fig 4E , P < 0 . 01 , n = 5 ) . Loss of HSF1 protein began to occur even in the absence of PERK activation in Braak stage III/IV and PS19 mouse ( Fig 1 ) . There is a possibility that proteotoxic stress from pathogenic tau affected HSF1 expression [19] . To determine if overexpressed mutant tau caused HSF1 protein loss and eIF2α-CHOP activation in vitro , we transiently transfected N2a neuroblastoma cells with wild-type- or P301L- or ΔK280-TauRD . Tau protein typically does not form amyloid fibrils in vitro because of its intrinsic hydrophilic feature . The lack of amyloidogenic propensity can be overcome by using the tau repeated domain ( TauRD ) , the most commonly used form of truncated tau . The four conserved sequence motifs in this domain are essential for tau aggregation . Both eIF2α-CHOP activation and decreased HSF1 protein expression were manifested in ΔK280 TauRD transfected N2a cells ( Fig 5A ) . Thus , we generated N2a cells stably overexpressing TauRD ΔK280 ( N2a-TauRD ΔK280 ) as a cellular model to study the relationship of CHOP activation and HSF1 loss in the following studies . In N2a-TauRD ΔK280 , about 40% of HSF1 protein was lost whereas 2 . 3-fold increase in CHOP protein was found in N2a-TauRD ΔK280 , which was statistically significant ( Fig 5B , HSF1 , P < 0 . 01; CHOP , P < 0 . 05 , n = 4 ) . We observed a remarkable HSF1 increase after treatment with each of three agents to inhibit autophagy-lysosome pathway and MG132 , a proteasomal blocker , suggesting that HSF1 protein was degraded by both UPS and autophagy-lysosome in N2a-TauRD ΔK280 ( Fig 5C ) . Since reduced HSF1 expression caused CHOP activation in the mouse brain ( Fig 2 ) , we asked if HSF1 degradation was related with CHOP activation in N2a-TauRD ΔK280 . Overexpressed HSF1 WT reduced about 60% of CHOP protein expression in N2a-TauRD ΔK280 ( Fig 5D , P < 0 . 001 , n = 4 ) . Conversely , eIF2α-CHOP activation seemed to further promote HSF1 degradation in N2a-TauRD ΔK280 ( Fig 5E , P < 0 . 01 , n = 4 ) , as seen in primary neurons in Fig 4 . Suppression of CHOP activation by rapamycin attenuated thapsigargin-induced autophagy-lysosomal HSF1 degradation in N2a-TauRD ΔK280 ( Fig 5F ) . During thapsigargin and tunicamycin treatments for various time periods , reduction of CHOP expression via siRNA upregulated HSF protein expression in N2a-TauRD ΔK280 ( Fig 5G ) , which was statistically significant ( Fig 5H , P < 0 . 05 , n = 4 ) . However , in the absence of thapsigargin , CHOP was revealed not to be a primary component in HSF1 degradation present in N2a-TauRD ΔK280 , as confirmed by the lack of statistically significant HSF1 change following CHOP silencing ( Fig 5H ) . HSP70 a5 ( BiP/GRP78 ) , a major ER chaperone protein , acts as a negative regulator of UPR signaling [9 , 39] . The promoter of HSP70a5 possesses DNA sequences called heat shock elements where HSF1 can bind for transcriptional activation [40] . In contrast to CHOP elevation , HSP70 a5 ( BiP/GRP78 ) expression normalized to β-actin was slightly reduced in the human AD brains when compared to non-AD brains ( Fig 6A , non-significant ) . In addition , the steady state level of HSP70 a5 ( BiP/GRP78 ) was 70% less and 60% less in N2a-TauRD ΔK280 and rTg ( tauP301L ) 4510 than their control , respectively ( Fig 6B ) . Overexpressed HSF1 protein in N2a-TauRD ΔK280 not only significantly enhanced HSP70 a5 ( BiP/GRP78 ) expression but also remarkably reduced TauRDΔK280 ( Fig 6C , HSP70 a5 ( BiP/GRP78 ) , P < 0 . 05; GFP-TauRD , P < 0 . 01 , n = 4 ) . Overexpressed HSP70 a5 ( BiP/GRP78 ) reduced TauRD ΔK280 accumulation without affecting HSF1 protein expression in N2a-TauRD ΔK280 ( Fig 6D ) . These enhanced expression of HSF1 and HSP70 a5 ( BiP/GRP78 ) led to increased cell survival in N2a-TauRD ΔK280 ( Fig 6E ) . Overexpressed HSF1 mutant ( i . e . HSF1Δ156–226 ) deficient in trimerization of a prerequisite step for transcriptional activation still demonstrated increased cell viability , however to a lesser degree than observed in HSF1 WT ( Fig 6E and S4 Fig ) . In primary neurons and N2a cells ( Fig 7A and 7B ) , in response to tunicamycin treatment , protective UPR was activated that increased expression of HSP70 a5 ( BiP/GRP78 ) ( Fig 7A and 7B , P < 0 . 05 , n = 4 ) . In contrast , this protective stress response was not elicited in our cellular tauopathy model of N2a-TauRD ΔK280 , supporting attenuated HSP70 a5 ( BiP/GRP78 ) expression in tauopathy ( Fig 7A and 7B , n = 4 ) .
Here we provide both in vitro and in vivo evidence that strongly suggests an auto-propagating interplay of UPR activation and HSF1 degradation being a common pathogenic feature in both human AD and tau transgenic mouse AD models ( Fig 7 ) . Characterizations of ER stress on early-stage AD / MCI patients have been neglected in many studies . The underlying mechanisms leading to chronically sustained ER stress in human AD brain are not entirely understood . This experiment looks to identify persistent and striking HSF1 degradation as an integral component in the chronic UPR activation pathway that ultimately causes tau hyperphosphorylation in early AD pathogenesis . It should be noted that proteotoxic stress from tau aggregation could promote de-stabilization of HSF1 protein in several ways that are not clearly understood ( Figs 1 and 5 ) . We implicate CHOP activation in the UPR pathway as one of key players that partially contributes to HSF1 degradation in tauopathy ( Figs 4 and 5 ) . CHOP activation can be regulated in the ER not only by PERK-eIF2α but also by IRE1 or other unknown ER-independent mechanism [38] . Consequently , reduced steady state level of HSF1 is likely to trigger permanent PERK-CHOP activation ( Fig 2 ) that reversely facilitates HSF1 loss , leading to tau hyperphosphorylation in tauopathy ( Figs 2 and 7D ) . It can therefore be reasonably concluded that suppression of pro-apoptotic protein CHOP could be an efficient means of protecting HSF1 from ER stress-related tauopathy . However , it was also revealed that in the young HSF1+/- mouse brain , HSF1 depletion alone was insufficient to cause either ER stress or tau phosphorylation ( Fig 2 ) , but another unidentified aging-related pathway with which HSF1 degradation associates with may actually cause disturbances in ER homeostasis . The relevant kinase ( s ) affecting tau phosphorylation in CHOP-HSF1 axis should also be determined in further studies . In addition to the role of autophagy-lysosomal activity on the HSF1 protein turnover , given highly poly-ubiquitinated HSF1 protein in N2a-TauRD ΔK280 ( S5B Fig ) , it is necessary to investigate the involvement of ubiquitin-proteasome system in tauopathy as we did in synucleinopathy [19] . Tau toxicity can impair another essential response in ER called Endoplasmic-reticulum-associated protein degradation ( ERAD ) [26] . Hrd1 is an ERAD-associated E3 ubiquitin protein ligase previously found to interact with tau [26] . We observed a tremendous decrease in Hrd1 expression in rTg4510 ( S6 Fig ) . The potential interactions between ERAD and HSF1 should be considered in future studies . Considering HSF1 loss observed in Braak III/IV stages , it also leaves open the possibility of the involvement of amyloid pathology in HSF1 degradation . We found that the extent of HSF1 loss was more severe in tau transgenic mice ( Tg4510 ) than in APP transgenic mice ( Tg2576 ) that produced amyloid beta plaques [41] ( Fig 1D ) . Tg2576 mice were reported not to show any signs of UPR activation by others [42] . However , the more recent in vitro data suggests possible involvement of UPR activation in Aβ toxicity [43] . It is thus necessary to identify whether Aβ can cause ER stress in vivo , and if so , whether or not this mechanism is tau-dependent . Another corollary of HSF1 loss represented in the study is that it can be deterministic of the balance between pro-apoptotic ( CHOP ) and pro-survival responses ( BiP ) to ER stress . Inhibited HSP70 a5 ( BiP/GRP78 ) protein expression is consistently observed in all our tauopathy models ( N2a-TauRDΔK280 , rTg ( taupP301L ) 4510 , and human AD , Fig 6 ) . It may lead to an inability to attenuate UPR [9 , 39] , contributing to tau aggregation and its toxicity in tauopathy ( Fig 6 ) . In vivo studies on Tau ΔK280 transgenic mice indicated that tau toxicity was closely related to its ability to form aggregates [23] . We could garner some evidence to speculate that the highly aggregated fibrillary form of tau is the major causative species to induce HSF1 degradation . Overexpressed full-length Tau WT , Tau P301L , and Tau R406W in N2a cells did not significantly alter HSF1 protein levels ( S5A Fig ) . Each tau repeated domain only ( TauRD ) and the mutant tau transgene lacking N-terminal insert ( expressed in rTg4510 ) are all considered to aggregate faster than full length tau and the mutant tau transgene including one N-terminal insert ( expressed in PS19 ) , respectively [26 , 44 , 45 , 46] . The extent of HSF1 degradation may explain the discrepancies that exist in the presence of UPR activation in different tau transgenic mouse models [3 , 26] ( Fig 1 ) . Whereas the amount of the mutant human tau in PS19 mice is ~5 times the level of endogenous murine tau , expression of the mutant tau in rTg4510 is approximately 13 fold higher than that of the endogenous tau protein [24 , 25] . This higher levels of mutant tau expression in rTg4510 may account for a dramatic decrease of HSF1 protein associated with UPR activation in their brains as compared to PS19 . The consequence of eIF2α phosphorylation appears to be biphasic [47 , 48] . Though salubrinal is known to inhibit ER stress via protein synthesis attenuation ( to prevent ER protein overload ) [49] , elevated p-eIF2α by salubrinal rather facilitated UPR-induced HSF1 degradation in N2a-TauRD ΔK280 of our study ( Fig 5 ) . Therefore , furthermore comprehensive consideration of both pathways and their functional outcomes in the cellular context is required . The role of tau aggregates is increasingly recognized in Huntington’s disease ( HD ) . It was previously reported that strong AT8-immunoreactive phosphorylated tau and sarkosyl-insoluble tau were present in human HD patients , exhibiting rod-like tau deposits along neuronal nuclei [50 , 51] . Aberrant HSF1 protein degradation in the brains of both HD mice and human HD patients has been recently demonstrated [52] . Therefore , not only in AD but also in HD pathogenesis , tau aggregates might play a critical role in nuclear HSF1 protein degradation , possibly in conjunction with UPR activation as revealed in the current study . Our results are informative of the caution that should be taken in future designs of therapeutic approaches seeking to treat neurodegenerative diseases amenable to UPR inhibition .
The data on human subjects were analyzed anonymously . All procedures on mice were approved by the Institutional Animal Care and Use Committee ( IACUC ) at University of Tennessee Health Science Center ( UTHSC ) . Mice were kept in accordance with the institutional guidelines regarding the care and use of laboratory animals . rTg4510 and rTg21221 mice ( maintained in Dr . Hui Zheng’s laboratory , Baylor college of medicine and Dr . Karen H . Ashe’s laboratory , University of Minnesota , respectively ) were produced by crossing Tau responder line with CaMKIIa-tTA transactivator line and the mouse brain tissue lysates were used in this study . PS19 mice were obtained on a C57BL/6J genetic background from Jackson laboratory ( JAX ) . HSF1 +/- mice ( JAX ) were crossed once to C57BL/6J and maintained via intercrossing the F1 HSF1+/- mice . To induce UPR activation , five month-old C57BL6 mice were intraperitoneally administered with thapsigargin ( 2 mg/kg ) dissolved in PBS with 10% DMSO . Tissue specimens of Non-AD and AD patients were obtained from the Human Brain and Spinal Fluid Resource Center , which is sponsored by NIHDS/NIMH , National Multiple Sclerosis Society , Department of Veterans Affairs . Cases derived from short-postmortem interval ( PMI ) autopsies from the University of Kentucky AD Center ( UK-ADC ) cohort [53] . Premortem clinical evaluations and pathological assessments were as described previously [54] . Frontal cortical sections correspond to Brodmann Area 9 . Tissue used for the biochemical analyses were snap-frozen during the autopsy in liquid nitrogen and then stored at -80 oC . The inclusion/exclusion criteria that were applied: PMI <4 hrs; no evidence of frontotemporal dementia ( clinically ) or frontotemporal lobar degeneration ( pathologically ) ; no cancer in the brain parenchyma; and no large infarctions in the brain , or microinfarcts found within 3 cm of the brain tissue samples . The neuropathological features were assessed using standard neuropathological procedures as described in detail elsewhere [54 , 55] . Mouse hippocampus was homogenated into sucrose-HEPES buffer ( 4 mM HEPES pH 7 . 4 , 320 mM Sucrose , 2 mM EGTA , containing protease and phosphatase inhibitors ) . Centrifuge the homogenate at 1 , 000 g for 10 min at 4 oC to remove the pelleted nuclear fraction . The resulting supernatant from further centrifugation at 15 , 000 g for 15 min at 4 oC yielded the crude synaptosomal pellet . Resuspend the resulting pellet into sucrose-HEPES buffer . The resulting supernatant from further centrifugation at 15 , 000 g for 15 min at 4 oC was discarded to remove contaminants . The pellet was then lysed in 4 mM pH 7 . 4 HEPES and incubated at 4 oC for 30 min . The resulting pellet from further centrifugation at 25 , 000 g for 30 min contained synaptosomal membrane fraction . Sarkosyl-insoluble tau protein from mouse hippocampus was biochemically isolated as described in Calignon et al [56] . Mouse brain tissues and human brain were fixed in 4% paraformaldehyde , followed by sectioning and then blocking and incubation of primary antibody overnight and then AlexaFluor-conjugated secondary antibody . A double-staining protocol was used to compare Tau46-immunoreactivity to thioflavin S staining of plaques in the mouse brain . Mouse brain tissues were immunostained with Tau46 antibody , followed by Thioflavin S staining . Thioflavin S staining procedure was performed as described in Ly et al [57] . Images were captured under a confocal microscope ( Olympus America , Center Valley , PA , USA ) and Olympus fluorescent microscope . N2a cells were cultured in a 1:1 mixture of DMEM and Opti-MEM containing 10% fetal bovine serum ( FBS ) . Cells were transfected with plasmids using Lipofectamine 2000 ( Invitrogen ) . For generation of stably transfected cell lines ( N2a-TauRDΔK280 ) , G418 at 1mg/ml was included in the culture medium . Primary cortical and hippocampal neurons were prepared from E17 rat embryos and maintained in neurobasal medium supplemented with 0 . 8 mm l-glutamine and B27 . Primary neurons cultured for 10–18 days in vitro ( DIV ) were used for the study . Plasmid constructs used in transient transfection include full length tau ( pcDNA-WT , R406W ) and repeated domain tau constructs ( pcDNA-TauRD WT , TauRD P301L , TauRD ΔK280 ) ; pcDNA-HSF1 WT , S303A and HSF1 Δ156–226; pcDNA-HSP70a5 ( BiP/GRP78 ) . HSF1 WT and HSF1 S303A were generously given by Dr . Dennis Thiele at Duke University . HSF1 Δ156–226 ( trimerization mutant ) was cloned by using a method of site-directed mutagenesis ( Agilent Technologies ) . Tunicamycin , thapsigargin , salubrinal , resveratrol , piceid , celastrol , rapamycin , and riluzole were all purchased from Sigma . siRNA oligomers for CHOP and BiP were purchased from Sigma . Tissues and cells were lysed using RIPA buffer ( 10 mM Tris-Cl ( pH 8 . 0 ) , 1 mM EDTA , 1% Triton X-100 , 0 . 1% sodium deoxycholate , 0 . 1% SDS , 140 mM NaCl ) containing protease inhibitor ( leupeptin , pepstatin A , phenylmethylsulfonyl fluoride ( PMSF ) , aproptinin ) . The resulting lysates were subjected to western blots . Western blots were performed as previously described [19] . Both monoclonal antibody ( Santa Cruz , E-4 ) and polyclonal antibody ( Cell signaling , 4356 ) were used to study HSF1 protein expression . Both antibodies detected ~ 82 kDa HSF1 but showed different banding patterns . p-PERK/p-eIF2α/ eIF2α/BiP/CHOP/Tau ( Tau46 ) ( these six from Cell signaling ) were also used in the study . Tau ( Tau46 ) mouse monoclonal antibody detects all six isoforms of tau based on the amino acid sequence . For phosphorylated tau , two specific antibodies of AT8 ( MN1020 ) and AT100 ( MN1060 ) were purchased from Thermofisherscientific . Western quantification was based on the intensity of interested signal using densitometry and ImageJ software program . Cells in a 24 well-plate were incubated with 10 μl of the 12 mM MTT stock solution in a total 100 μl fresh culture medium for 2~3 hrs at 37°C . Reaction was ceased by incubation with 100 μl of the SDS-HCl solution for 2 . 5 hrs . Absorbance at 570 nm was detected by Beckman Coulter DTX 880 multimode detector . All statistical analysis was performed by Student’s t-test for two groups’ comparison and one-way ANOVA with a Tukey test for multiple comparisons . Quantitative data in all graphs were presented by means ± SEM ( standard error of mean ) .
|
Tauopathy including Alzheimer’s disease ( AD ) is characterized by a build-up of tau aggregates in the brain , highly associated with failure of cellular protein homeostasis . Proteostasis can be achieved by protein quality control system to cope with numerous stresses such as proteotoxic stress from misfolded proteins . This cellular protective system includes heat shock response regulated by heat shock factor 1 ( HSF1 ) activation and unfolded protein response in ER . Despite the importance of stress response in maintaining proteostasis , their role in neurodegenerative diseases like tauopathy is not clearly understood . The current study reports how the interplay between the two stress response systems , unfolded protein response and HSF1 promotes early tau pathology and identifies HSF1 protein degradation being one likely key player in both human AD and tau transgenic mouse AD models . We identify aging-associated AD-like neuropathological changes in the hippocampus of HSF1 heterozygous knock-out mice . We speculate that that HSF1 loss may constitute a mechanistic connection between ER stress and tau hyperphosphorylation in tau pathology . This study demonstrates the potential therapeutic significance of stabilizing HSF1 protein in treating AD .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"cellular",
"stress",
"responses",
"neurodegenerative",
"diseases",
"cell",
"processes",
"endoplasmic",
"reticulum",
"brain",
"neuroscience",
"animal",
"models",
"protein",
"expression",
"model",
"organisms",
"experimental",
"organism",
"systems",
"molecular",
"biology",
"techniques",
"alzheimer",
"disease",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"heat",
"shock",
"response",
"animal",
"cells",
"proteins",
"mouse",
"models",
"dementia",
"mental",
"health",
"and",
"psychiatry",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"biochemistry",
"cellular",
"neuroscience",
"hippocampus",
"anatomy",
"cell",
"biology",
"post-translational",
"modification",
"secretory",
"pathway",
"neurology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types"
] |
2017
|
Bidirectional interplay of HSF1 degradation and UPR activation promotes tau hyperphosphorylation
|
The tick-borne flavivirus , Kyasanur Forest disease virus ( KFDV ) causes seasonal infections and periodic outbreaks in south-west India . The current vaccine offers poor protection with reported issues of coverage and immunogenicity . Since there are no approved prophylactic therapeutics for KFDV , type I IFN-α/β subtypes were assessed for antiviral potency against KFDV in cell culture . The continued passage of KFDV-infected cells with re-administered IFN-α2a treatment did not eliminate KFDV and had little effect on infectious particle production whereas the IFN-sensitive , green fluorescent protein-expressing vesicular stomatitis virus ( VSV-GFP ) infection was controlled . Further evaluation of the other IFN-α/β subtypes versus KFDV infection indicated that single treatments of either IFN-αWA and IFN-αΚ appeared to be more effective than IFN-α2a at reducing KFDV titres . Concentration-dependent analysis of these IFN-α/β subtypes revealed that regardless of subtype , low concentrations of IFN were able to limit cytopathic effects ( CPE ) , while significantly higher concentrations were needed for inhibition of virion release . Furthermore , expression of the KFDV NS5 in cell culture before IFN addition enabled VSV-GFP to overcome the effects of IFN-α/β signalling , producing a robust infection . Treatment of cell culture with IFN does not appear to be suitable for KFDV eradication and the assay used for such studies should be carefully considered . Further , it appears that the NS5 protein is sufficient to permit KFDV to bypass the antiviral properties of IFN . We suggest that other prophylactic therapeutics should be evaluated in place of IFN for treatment of individuals with KFDV disease .
Kyasanur Forest disease virus ( KFDV ) is a tick-borne flavivirus that was identified in 1957 following a monkey epizootic and a coinciding human outbreak in south-west India [1] . KFDV cases previously were localized within the Shimoga district of Karnataka; however KFDV has been recently discovered in the neighboring states of Kerala , Tamil Nadu , Goa and Maharashtra [2–5] and , possibly China in 1989 [6] increasing the potential public health risk associated with this pathogen . A vaccine for KFDV is available for those living in affected areas and those living within a 5 kilometer radius of a positive case from either humans , monkeys or tick pools [7] , but there has been issues with implementation and efficacy . The most troubling aspect of vaccine use is that less than half of the target population actually receive the full three-dose regimen that is required for protection [8 , 9] . With the annual number of cases ranging from 400–500 and an associated fatality rate of 3–5% [10] , there is a need for alternative therapeutic options , besides the current vaccine and tick bite prevention measures . KFDV is a member of the tick-borne encephalitis serocomplex which includes: tick-borne encephalitis , the former Russian spring-summer encephalitis , Omsk hemorrhagic fever , Powassan , Langat and Louping-Ill viruses [11] . A variant of KFDV , Alkhumra hemorrhagic fever virus located in Saudi Arabia [12] and in Egypt [13–15] , also is part of this complex [16] . The single-stranded positive-polarity RNA genome of KFDV is 10 , 774 bases in length and encodes a single polyprotein: C-prM-E-NS1-NS2A-NS2B-NS3-NS4A-NS4B-NS5 [17] . KFDV , Alkhumra hemorrhagic fever virus and Omsk hemorrhagic fever virus are unique to this complex as they primarily cause hemorrhagic fever manifestations with neurological involvement [18] . Interferon ( IFN ) was first described for its ability to interfere with virus infection in 1957 by Isaacs and Lindenmann [19 , 20] . In response to viral infection , IFN is released from infected cells to surrounding uninfected cells . Upon binding to its receptor and subsequent activation of the Jak/STAT signaling cascade , the cellular “antiviral state” is established . The antiviral state allows naïve cells to become resistant to viral infection via expression of IFN-stimulated genes ( ISG ) that have many roles in protecting the host from infection [21–23] . KFDV , like many other flaviviruses including dengue , yellow fever , Langat , West Nile , tick-borne encephalitis and Japanese encephalitis viruses [24–27] employs the NS5 protein to prevent antiviral state establishment and compromised IFN signalling [28] , whereas NS4B-2k has been also implicated for dengue , West Nile and yellow fever viruses [26 , 29 , 30] . The ramifications of a compromised Jak/STAT pathway were highlighted in mice lacking the IFN-α/β receptor . Infection of these mice with West Nile [31] , dengue [32] and Langat viruses [33] led to increased virus burden , earlier onset of clinical signs and higher mortality when compared to control mock-infected type mice . Of the type 1 IFNs ( 14 subtypes of α and one of β ) , only IFN-α2a and α2b are FDA-approved for clinical treatment in humans [22] . Case reports indicate a relatively narrow window for post-exposure treatment against other flaviviruses such as West Nile , Saint Louis encephalitis and Japanese encephalitis viruses with either IFN-α2a or 2b , especially in the immune-compromised or during the severe stages of disease [34–37] . In tissue culture , both IFN-α2a and α2b are potent at reducing the replication of dengue virus 2 [38 , 39] , West Nile [31 , 40 , 41] , Langat [25] and other flaviviruses [42] . Interestingly , strain-specific responses were observed between dengue virus 1 , 3 and 4 , yellow fever virus ( vaccine strains FNV and 17D ) [42] and Japanese encephalitis virus ( strains Vip , KE-093 , KE-094 and KE-095 ) [43] . Similarly , the antiviral activities of other IFN-α/β subtypes differ dramatically for vesicular stomatitis virus ( VSV ) , human immunodeficiency virus [44] and human rhinoviruses [45] . To our knowledge , with the exception of IFN-α2a , the other IFN-α/β subtypes have not been evaluated against KFDV . Current treatment options for KFDV are limited to supportive care , because no therapeutic drugs have been clinically approved . Using VSV as an IFN-sensitive virus control for comparison , the antiviral efficacy of the IFN-α/β subtypes was assessed against KFDV infection in cell culture . We conclude that IFN-α2a was unable to eradicate KFDV from cell culture , but appeared to offer protection form cytopathic effects ( CPE ) . Although some IFN subtypes at higher doses did restrict KFDV replication , albeit marginally , further scrutiny could not demonstrate a strong dose-dependent response in terms of virion production Thus , it is not appropriate to use protection from CPE as the sole determinant of potency . In conclusion , KFDV is not sensitive to the antiviral effects of IFN and despite the lack of a concise mechanism for IFN resistance , it is clear that the NS5 protein is a major contributor .
A549 and BHK cells were seeded in 6-well plates for 80–90% confluency at time of infection . The cells were infected with 11 TCID50 units ( equivalent to a multiplicity of infection ( MOI ) of 0 . 00001 ) of KFDV or VSV-GFP for 1 hour , then virus maintenance medium supplemented with a final concentration of 2 , 000 U/ml of IFN-α2a was added and incubated for 96 hours ( KFDV ) and 48 hours ( VSV-GFP ) ; this was designated as passage 0 . Controls wells were treated similarly without the addition of IFN . Supernatants were harvested and stored at -80°C for titration . The infected-cells were washed with PBS , passaged by use of trypsin ( HyClone ) and the cells were split ( 1:3 ) into two new 6 well plate wells . These passage 1 samples were treated again with 2 , 000 U/mL IFN-α2a or were left untreated and re-incubated for 72 hours ( KFDV ) and 48 hours ( VSV-GFP ) . The procedure was repeated from passage 1 to passage 2 . Pictures from virus-infected cells were captured using an EVOS microscope for KFDV samples under CL-4 conditions and with an Axiovert 40 CFL microscope ( Carl Zeiss , Toronto , Ontario , Canada ) for VSV-GFP samples under CL-2 conditions . A549 cells were prepared in 24-well plates with DMEM growth medium for 80–90% confluency at time of infection . In the pre-treatment scenario , the cells were treated with or without 1 , 000 U/ml of the selected IFN-α/β subtypes was added and incubated 24 hours before virus infection . Then the cells were infected with KFDV at a MOI of 1 for 1 hour , inoculum was removed , cell monolayers were washed and virus maintenance medium was added . In the post-treatment experiments , KFDV at a MOI of 1 was adsorbed for 1 hour , inoculum was removed , monolayers were washed , and then fresh virus maintenance medium supplemented with or without 1 , 000 U/ml of the selected IFN α/β subtypes . Supernatants from three-independent biological replicates were harvested and stored at -80°C for virus quantification after an incubation period of 72 hours . Statistical significance was determined using One-way ANOVA analysis followed by Tukey’s post-test .
Using virus clearance assays that involved successive passaging of infected cell cultures along with the re-application of antivirals [47 , 48] , we investigated the efficacy of IFN-α2a against KFDV and for control purposes , against VSV-GFP . To address this , both A549 and BHK-21 cells were infected with either KFDV or VSV-GFP at 11 TCID50 units ( MOI of 0 . 0001 ) . After infection the cells were treated with 2 , 000 U/mL of IFN-α2a at 1 hpi ( passage 0 ) . Virus supernatants were harvested for titration , and passage 0 infected-cells were subcultured into two wells and IFN was either added or withheld ( passage 1 ) and this procedure was repeated for passage 2 . Initially the passage 0 , IFN-treated cells showed a reduction in KFDV titre from the mock-treated controls of 100 . 8 TCID50/mL ( p< 0 . 1 ) and 100 . 4 TCID50/mL ( p< 0 . 1 ) in both A549 and BHK-21 cells respectively . However , decreases were more striking for the IFN-sensitive VSV-GFP as reductions of 105 . 2 TCID50/mL ( p< 0 . 01 ) and 106 . 5 TCID50/mL ( p< 0 . 01 ) were observed in A549 and BHK-21 cells respectively ( Fig 1A and Fig 2A ) . When comparing the IFN-treated cells from passage 0 to passage 1 and passage 1 to passage 2 , only VSV-GFP showed marked differences in titres . There were no apparent decreases in KFDV titres produced from IFN-treated A549 cells ( passage 1: decrease of 100 . 1 TCID50/mL ( not significant ) from passage 0 , and for passage 2: increase of 100 . 7 TCID50/mL ( p< 0 . 1 ) from passage 1 ) ( Fig 1A ) and from IFN-treated BHK-21 cells ( passage 1: increase of 100 . 6 TCID50/mL ( not significant ) from passage 0 , and passage 2: decrease of 100 . 3 TCID50/mL ( not significant ) from passage 2 ) ( Fig 2A ) . The results from KFDV contrast that of VSV-GFP in which sharp declines in titres were observed in A549 cells ( passage 1: decrease of 101 . 6 TCID50/mL ( p< 0 . 1 ) from passage 0 , and a reduction of 101 . 7 TCID50/mL ( p< 0 . 01 ) from passage 1 ) ( Fig 1A ) and BHK-21 cells ( passage 1: a decrease of 102 . 9 TCID50/mL ( p< 0 . 01 ) , and passage 2: no virus detected ) cells ( Fig 2A ) . As anticipated , when IFN-treated cells from either passage 0 or passage 1 were subcultured and not further treated with IFN , virus titres and CPE for both KFDV and VSV-GFP were characteristic of the mock-treated passage 0 controls . Despite the apparent inability of IFN to significantly impact KFDV propagation , the appearance of CPE was notably reduced compared to mock-treated in both cell types tested ( Fig 1B and Fig 2B ) . Even with the lack of visible CPE in the IFN-treated cells , KFDV could not be eliminated from infected cell culture by treatment with IFN-α2a . Previous studies evaluating IFN treatment against flaviviruses in cell culture have revealed that IFN is more effective when added to cells before virus infection and becomes less effective over time once infection has been established [25 , 31 , 39 , 41] . To further expand our understanding of the efficacy of IFN treatment against KFDV infection , we evaluated the antiviral activity of the IFN-α/β subtypes in both pre-infection and post-infection time frames . A549 cells were infected with KFDV ( MOI of 1 ) and treated with 1 , 000 U/mL of each IFN-α/β , which was added either 24 hours before or 1 hour after infection . When CPE reached 90–100% in mock-treated controls , supernatants were harvested and titrated . Regardless of time of addition , both IFN-αK and IFN-αWA were more effective than IFN-α2a at reducing KFDV titres ( Fig 3 ) . When compared to the mock-treated controls , fold reductions in pre-infection and post-infection time frames were 16-fold ( IFN-αK ) , 14-fold ( IFN-αWA ) and 3-fold ( IFN-α2a ) ( Fig 3 , grey bars ) and , 132-fold ( IFN-αK p<0 . 01 ) , 37-fold ( IFN-αWA p<0 . 01 ) and 6-fold ( IFN-α2a p<0 . 1 ) ( Fig 3 , black bars ) , respectively . These data suggest that IFN-αK and IFN-αWA may induce responses that are more effective than those produced by IFN-α2a for restricting KFDV infection regardless of the time of treatment relative to infection . The two candidate IFN subtypes , IFN-αK and IFN-αWA were observed to be more effective than IFN-α2a at repressing KFDV replication at a concentration of 1 , 000 U/mL . To further explore this finding , we sought to determine if the potency differences of IFN-αWA compared to IFN-α2a can be maintained over a range of concentrations . This was tested by analyzing the ability of each IFN subtypes to protect A549 and BHK-21 cells from the cytopathology of KFDV ( 11 TCID50 units equal to MOI of 0 . 0003 ) and a IFN-sensitive control virus ( VSV-GFP ) ( 11 TCID50 units equal to MOI of 0 . 0003 ) following a two-fold dilution scheme for each IFN . The traditional approach to defining IFN potency was utilized; these assays involve the use of multiple cell types including A549 cells , treated with IFN and challenged with a strong cytolytic virus like VSV , then analyzed empirically with monolayer staining/spectroscopy [21 , 49–51] . As an additional measure , IFN strength was assessed by infectious virion production determined by TCID50 titration from select IFN concentrations . For the cytopathology assessment , upper and lower parameters were defined as either full CPE ( 0% cell protection ) from mock-treated samples and absence of CPE ( 100% cell protection ) from un-infected samples , thereby allowing IC50 and IC90 to be determined for each IFN . Initially both IFN subtypes appeared equally as potent , as the IC50 values for IFN-αWA and IFN-α2a were comparable against KFDV ( 5 . 2 and 7 . 4 U/mL ) and VSV-GFP ( 5 . 1 and 5 . 6 U/mL ) in A549 cells ( Table 1 ) . However , the same trend was not observed in BHK-21 cells with IC50 values of 23 . 3 ( IFN-αWA ) and 6 . 9 U/mL ( IFN-α2a ) for KFDV . An IC50 of 988 . 1 U/mL was determined for IFN-α2a against VSV-GFP and a value could not be determined ( ND ) for IFN-αWA . This may have been because the amount of IFN required to reach the IC50 was greater than the highest concentration tested and therefore could not be determined using the curve fitting software ( Table 1 ) . IC90 values obtained from KFDV-infected A549 cells that were treated with IFN-αWA ( 406 . 8 U/mL ) and IFN-α2a ( 125 . 4 U/mL ) were significantly higher than VSV-GFP-infected A549 cells that were treated with IFN-αWA ( 48 . 5 U/mL ) and IFN-α2a ( 50 . 0 U/mL ) ( Table 1 ) . A similar trend was observed for KFDV-infected BHK-21 cells . The 90% protection from IFN-αWA ( 23 , 711 . 0 U/mL ) was extrapolated by curve fitting software as it was much greater than the highest amount tested ( 16 , 000 . 0 U/mL ) . This predicted value was significantly higher than that of IFN-α2a ( 2 , 048 . 0 U/mL ) . With respect to VSV-GFP-infected BHK-21 cells that were treated with IFN , the IC90 was determined to be 3 , 808 . 0 U/mL for IFN-αWA and 56 . 2 U/mL for IFN-α2a ( Table 1 ) . IFN-induced cytotoxicity ( CC50 ) was not determined as a concentration of 16 , 000 . 0 U/mL , an amount that was three times larger than the highest concentration used for experiments led to cell death in A549 ( 25% cytotoxicity ) and BHK-21 ( 11% cytotoxicity ) cells . For control parameters , un-treated ( 0% cytotoxicity ) and 10% Triton X-100-treated ( 100% cytotoxicity ) samples defined the upper and lower limits of CC50 ( Table 1 ) . While it was apparent that high concentrations of both IFN-αWA and IFN-α2a are required to prevent KFDV-induced CPE , the experiments utilizing BHK-21 cells highlight the problems of obtaining qualitative data for IFN-protection via CPE , especially against VSV-GFP . To further define IFN potency with respect to the variation of IC values and dose-dependence , virus titres from a range of IFN concentrations ( 2 , 000 . 0 , 500 . 0 , 62 . 5 and 7 . 8 U/mL ) selected from the two-fold IFN dilution experiments in both A549 and BHK-21 cells were compared to un-treated controls . This was compared as titre reductions for both KFDV and VSV-GFP ( control virus ) . For the highest ( 2 , 000 . 0 U/mL ) and lowest ( 7 . 8 U/mL ) concentrations assessed for KFDV-infected A549 cells , titre reductions of 100 . 7 and 100 . 5 TCID50/mL ( IFN-αWA ) and , 100 . 5 and 100 . 4 TCID50/mL ( IFN-α2a ) , respectively , were measured . In BHK-21 cells infected with KFDV , titre reductions were not observed for 2 , 000 . 0 and 7 . 8 U/mL of IFN-αWA; however , reductions of 100 . 9 and 100 . 1 TCID50/mL were observed for IFN-α2a . By contrast in VSV-GFP , titre reductions were more pronounced in the A549 cells treated with both 2 , 000 . 0 U/mL ( 107 . 9 TCID50/mL for IFN-αWA and 106 . 9 TCID50/mL for IFN-α2a ) and 7 . 8 U/mL ( 103 . 6 TCID50/mL for IFN-αWA and 103 . 5 TCID50/mL for IFN-α2a ) . And for BHK-21 cells treated with 2 , 000 . 0 U/mL ( 101 . 8 TCID50/mL for IFN-αWA and 104 . 4 TCID50/mL for IFN-α2a ) and 7 . 8 U/mL ( 100 . 3 TCID50/mL for IFN-αWA and 101 . 5 TCID50/mL for IFN-α2a ) ( Table 2 ) . When comparing VSV-GFP to KFDV , it is clear that KFDV is not sensitive to the effects of IFN . While the traditional monolayer staining procedure demonstrated a dose-dependent relationship between CPE and IFN concentration , the assay showed variability from experiment to experiment , especially in BHK-21 cells . The measurement of virus titre proved more reliable as the true IFN-insensitive nature of KFDV was revealed . Similar to other flaviviruses , our previous work had revealed that the NS5 protein and not the NS4B-2k protein of KFDV could inhibit IFN signaling by interrupting the Jak/STAT pathway [28] . We characterized this further in hopes of determining if the failure of the IFN-α/β subtypes to restrain KFDV replication was due to the effects of NS5 . VeroE6 cells were transfected with plasmids expressing the KFDV: NS5 protein and NS4B-2k proteins which previously have been shown to have anti-IFN and lacking anti-IFN activity , respectively [28] and , for control purposes the Ebola virus VP24 protein which is an IFN-signaling repressor [52] . The cells were then treated with 1 , 000 U/mL of Universal IFN with exception of the mock-treated control , before being infected with VSV-GFP ( MOI of 2 ) . The mock-treated controls demonstrated extensive CPE and GFP expression indicating an un-inhibited infection ( Fig 4A ) . As anticipated IFN prevented VSV-GFP propagation in NS4B-2k and empty-expression vector controls as indicated by CPE and GFP expression and these were comparable to those seen in uninfected cells . VSV-GFP was only able to overcome the effects of IFN when NS5 was present and this was similar to the VP24 control ( Fig 4A ) . With the KFDV NS5 established as a primary antagonist against Universal IFN activity , the assay was extended to compare the ability of NS5 to circumvent the other IFN-α/β subtypes measured by VSV-GFP infectious particle production . VeroE6 cells were either transfected with plasmids expressing NS5 , VP24 , or mock-transfected , then treated with 1 , 000 U/mL of each IFN-α/β subtypes , with exception of a mock-treated control and , finally infected with VSV-GFP ( MOI of 2 ) . When mock-treated controls reached 90–100% CPE , virus-containing supernatants were harvested and titrated . The mock-treated cells were susceptible to VSV-GFP and NS5 expression did not negatively-impact infection as 109 . 5 TCID50/mL was detected under both circumstances . All of the IFN-α/β subtypes limited VSV-GFP replication ( dark grey bars ) ; however , the expression of NS5 reversed the antiviral effects of all IFN subtypes ( light grey bars ) . Therefore , NS5 was able to recover VSV-GFP and permitted virus titres to increase by ( reported from lowest to highest ) : 101 . 1 for IFN-αG to 105 . 1 for IFN-αWA and , 105 . 6 for IFN- β ( Fig 4B ) . Furthermore , NS5 appeared to be a stronger repressor of Universal IFN than VP24 , as NS5 allowed for virus titre to increase by 102 . 1 TCID50/mL ( p< 0 . 01 ) ( Fig 4B ) . Thus , the NS5 protein of KFDV appears to be a potent inhibitor of the antiviral effects afforded by IFN .
As a treatment in clinical settings against flavivirus diseases , IFN-α2a and IFN-α2b have had variable success . When compared against the IFN-sensitive VSV-GFP virus , we discovered that IFN-α2a was unable to reduce KFDV replication in cell culture . Although some IFN subtypes appeared to be more potent than IFN-α2a at reducing KFDV replication , further analysis indicated that regardless of IFN subtype , IFN was inadequate against KFDV replication . It also appears that KFDV-induced CPE prevention is an unreliable criterion for the inhibitory effects of IFN , as the extent of CPE production did not coincide with the impact of IFN on viral titres . Furthermore , the minimized impact of IFN on KFDV to be an inherent ability of the virus correlated with activity of the NS5 protein . For several flavivirus cell culture models , the antiviral potency of IFN decreases significantly once flavivirus infection has been established , especially after 4–6 hpi [25 , 31 , 40 , 43 , 53] . The finding that treatment at 1 hpi was more potent at reducing both VSV-GFP and KFDV than treatment at 24 hours pre-infection may be an A549 cellular-dependent observation . It is well known that the induction of the cellular antiviral program and duration of response ( IFN and ISG expression ) can vary with cell type [23] . With this in mind , another unrelated cell type BHK-21 , which like A549 can respond to IFN-signalling , but is a weak IFN producer [48 , 54] , was added into the virus clearance assay . In both cell types it was clear that IFN failed to quell KFDV infection . However since the infected cells were passaged , it may be possible that the passaged cells used for seeding new wells have been irreversibly compromised from the anti-IFN actions of KFDV . Thus , despite the cells’ ability to replicate and become confluent , they could have remained in a state that permitted KFDV infection , rendering IFN ineffective . Although the latter experiments indicate that NS5 can interfere with IFN signalling , its duration has not been investigated . Another interesting observation is the lack of CPE despite high titres of KFDV . While it is not clear why this is occurring , other flaviviruses promote autophagy induction to prolong cell survival and prevent apoptosis for long-term replication within autophagy compartments [55–57] . Additionally , the inhibition of apoptosis using a steroid ( dehydroepiandrosterone ) could prevent CPE during infection with Japanese encephalitis virus without impacting virus replication [58] . Perhaps KFDV is inhibiting apoptosis or utilizing autophagy to prevent apoptosis in the presence IFN and allowing apoptosis in the absence of IFN , as depicted in the differences in CPE development . Further studies should be conducted to confirm either of these hypotheses . Initial screening of IFN-α/β subtypes suggested that IFN-αWA and IFN-αK were more potent than IFN-α2a at preventing KFDV virus production . While it is still unknown why these α and β subtypes display such a range in antiviral activity , it may be due to the strength of receptor binding or IFN-subtypes specific antiviral ISG expression [59 , 60] . It has been reported that virus concentration can impact distinct induction of not only ISG expression but also IFN-α subtype induction [61] . Perhaps , the potency of IFN-αWA and IFN-αK against the high MOI of KFDV challenge virus caused the expression of unique ISG profiles and should be further investigated for tick-borne flavivirus-specific ISGs . TRIM79α is a perfect example of a potent ISG that prevents infections by tick-borne but not mosquito-borne flaviviruses [62] . More in depth analysis of the IFN subtypes revealed that regardless of the concentration or subtypes applied , IFN may provide some protection from CPE but has very limited activity against KFDV replication in cell culture . This was evident when the more potent IFN subtypes ( IFN-αWA and IFN-αK ) were comparable at restricting KFDV replication to IFN-α2a at the lower MOI . This contrasting result is likely due to the fact that the MOI used influences the cellular ISG expression profiles [61] . In this case , the lower MOI may have reflected this difference in the ability of IFN-αWA and IFN-αK to induce anti-KFDV ISGs as observed from the higher MOI utilized in the screening assay . Despite its traditional use for measuring IFN potency , the dye-uptake assay demonstrated that CPE prevention is not a valid measure for IFN potency against KFDV . Thus , any sort of CPE or cell viability assay for IFN potency against KFDV and perhaps other flaviviruses must be carefully considered , as the most accurate measurement proved to be determination of the number/concentration of infectious virus particles . An example of such a discrepancy was seen in Alkhumra hemorrhagic fever studies treated with IFN-α2a , as the concentrations required to reduced infection by 50% differed from 684 ( +/- 499 ) U/mL for CPE ( metabolic-cell viability assay ) and 12 ( +/- 6 ) U/mL for virus titres ( qRT-PCR ) [63] . Although genetic-quantification is not the same as infectious particle counts , this still demonstrates the inconsistency seen with CPE-based assays . Moreover , the same metabolic assay revealed that KFDV infection could be restricted by 50% infection with 863 ( +/- 450 ) U/mL IFN-α2a . Unfortunately this was not confirmed by virus titre [63] . In another study , IC50 was defined by the visual-determination CPE reduction and IC90 was determined by reductions in virus titres . Viruses such as: dengue virus 1 had IC50 and IC90 values of 15 . 8 and 1 U/mL respectively and Zika virus had similar IC50 and IC90 values of 34 and 30 U/mL respectively [42] . In summation , it would appear that the most accurate measurement of IFN potency would infectious particle release . Our data indicate that IFN does not significantly reduce KFDV replication and virion production . It appears that the KFDV NS5 protein plays an important role in the failure of IFN-α/β subtypes to control KFDV propagation . Further experimentation with knockdown assays was not feasible because the NS5 protein is critical for virus replication and our previous mutational analysis indicated that the anti-Jak/STAT pathway activity of NS5 was within the RNA-dependent RNA polymerase domain [28] . Aside from the Jak/STAT pathway it remains unclear which other IFN-induced pathways are targeted by the NS5 protein , allowing for recovery of the IFN-sensitive VSV-GFP virus in the presence of IFN . It would be interesting to examine further if KFDV much like other flaviviruses , can influence ISG production from not only the Jak/STAT pathway [28] , but the phosphatidylinositol 3-kinase ( PI3K ) , cyclic GMP-AMP synthase ( cGAS ) [64] and IRF-1 [65] pathways . Flaviviruses respond to IFN very differently within the mosquito-borne and tick-borne groups and , even within strains or serotypes . IFN concentrations as low as 1 U/mL and as high as 200 U/mL have been successful at reducing in vitro infection from many flaviviruses by 90% [42] . Unlike these flaviviruses , KFDV infection was not limited by 90% using any of the IFN-α/β subtypes . Moreover , case reports indicate that IFN treatment can be marginal against flavivirus diseases and this is confounded by individuals who are immuno-compromised or those who delay or cannot readily access medical care [34–37] . While IFN does stimulate the adaptive immune response as summarised in [66] , it is difficult to say how well KFDV infection would be controlled in vivo and should be investigated for confirmation . Nonetheless , the in vitro data presented here suggests that IFN may be an inadequate prophylaxis for KFDV infections and perhaps other treatment options should be explored .
|
Since 1957 Kyasanur Forest disease virus ( KFDV ) has caused seasonal infections and periodic outbreaks in south-west India . It is estimated that nearly 500 people acquire KFDV annually and 3–5% of those infected succumb to the disease . The vaccine strategy is complicated by the lack of coverage , compliance and efficacy , highlighted by the fact that less than half of the target population received the recommended three dose-regimen . Besides the prevention of tick bites and vaccination , there are no approved antivirals for KFDV infection . Based on these observations , the commonly-used-IFN-α2a was assessed and was not capable of limiting KFDV virus titres . Further characterization of the other IFN-α/β subtypes used at different concentrations revealed that KFDV replication was insensitive to all subtypes , even though signs of cellular damage were reduced . Thus , infectious titre , rather than monolayer staining or cytopathic effect ( CPE ) monitoring , is more reliable for IFN analyses . The capability of KFDV to overcome the antiviral properties of IFN was attributed to the NS5 protein . Thus , other treatment options need to be evaluated for patients suffering with Kyasanur Forest disease .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"japanese",
"encephalitis",
"virus",
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"vesicular",
"stomatitis",
"virus",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"biological",
"cultures",
"microbiology",
"toxicology",
"viruses",
"rna",
"viruses",
"cell",
"cultures",
"research",
"and",
"analysis",
"methods",
"proteins",
"medical",
"microbiology",
"microbial",
"pathogens",
"viral",
"replication",
"cytotoxicity",
"biochemistry",
"rhabdoviruses",
"flaviviruses",
"virology",
"viral",
"pathogens",
"interferons",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2016
|
Limited Effects of Type I Interferons on Kyasanur Forest Disease Virus in Cell Culture
|
Recent comprehensive sequence analysis of the maize genome now permits detailed discovery and description of all transposable elements ( TEs ) in this complex nuclear environment . Reiteratively optimized structural and homology criteria were used in the computer-assisted search for retroelements , TEs that transpose by reverse transcription of an RNA intermediate , with the final results verified by manual inspection . Retroelements were found to occupy the majority ( >75% ) of the nuclear genome in maize inbred B73 . Unprecedented genetic diversity was discovered in the long terminal repeat ( LTR ) retrotransposon class of retroelements , with >400 families ( >350 newly discovered ) contributing >31 , 000 intact elements . The two other classes of retroelements , SINEs ( four families ) and LINEs ( at least 30 families ) , were observed to contribute 1 , 991 and ∼35 , 000 copies , respectively , or a combined ∼1% of the B73 nuclear genome . With regard to fully intact elements , median copy numbers for all retroelement families in maize was 2 because >250 LTR retrotransposon families contained only one or two intact members that could be detected in the B73 draft sequence . The majority , perhaps all , of the investigated retroelement families exhibited non-random dispersal across the maize genome , with LINEs , SINEs , and many low-copy-number LTR retrotransposons exhibiting a bias for accumulation in gene-rich regions . In contrast , most ( but not all ) medium- and high-copy-number LTR retrotransposons were found to preferentially accumulate in gene-poor regions like pericentromeric heterochromatin , while a few high-copy-number families exhibited the opposite bias . Regions of the genome with the highest LTR retrotransposon density contained the lowest LTR retrotransposon diversity . These results indicate that the maize genome provides a great number of different niches for the survival and procreation of a great variety of retroelements that have evolved to differentially occupy and exploit this genomic diversity .
Transposable elements ( TEs ) were first discovered in maize ( Zea mays ) [1] , but have subsequently been found in almost every organism investigated , from archaea and eubacteria to animals , plants , fungi and protists [2] . TEs are dynamic , abundant and diverse components of higher eukaryotic genomes , where they play key roles in the evolution of genes and genomes . The class I TEs transpose through reverse transcription of a transcribed RNA intermediate , while most class II TEs transpose through a cut-and-paste mechanism that mobilizes the DNA directly . However , there are some class II TEs , for instance IS91 of bacteria and Helitrons in eukaryotes , that are believed to transpose through a rolling-circle DNA replication process that does not involve element excision [3] , [4] . In most plant species , a particular type of class I element , the long terminal repeat ( LTR ) retrotransposons , has been observed to be the major TE , accounting for >80% of the nuclear DNA in many angiosperms [5] . The other two types of class I elements , LINEs and SINEs , have also been observed in all carefully annotated flowering plant genomes , but their copy numbers and overall contributions to genome composition have not usually been large . However , in lily ( Lilium speciosum ) and grapevine ( Vitis vinifera ) , LINEs appear to be more numerous and/or active than in most plant species investigated [6] , [7] . A wealth of recent studies has indicated that the class I elements , especially LTR retrotransposons , are primary contributors to the dynamics of genome structure , function and evolution in higher plants . Even within species , the LTR retrotransposon arrangement and copy number can vary dramatically in different haplotypes [8]–[11] . Some LTR retrotransposons acquire and amplify gene fragments [12] , [13] , and sometimes fuse their coding potential with those of other genes [14] , to create “exon shuffled” candidate genes that have the potential to evolve novel genetic functions [15] . Retroelements of all types may also serve as sites for the ectopic recombination events that can cause chromosomal rearrangements: duplications , deletions , inversions and translocations . Retroelement insertions can donate their transcriptional regulatory functions to any adjacent gene , and the prevalence of this process over evolutionary time is indicated by the many fragments of retroelements and other TEs that are found in current plant gene promoters [16] . In angiosperms , polyploidy and retroelement amplification are the major factors responsible for the greater than 1000-fold variation in genome size [5] . In some lineages , amplification of only one or a few LTR retrotransposon families has been observed to more than double genome size in just a few million years [17] , [18] . In other organisms , like maize , many different LTR retrotransposon families have amplified in recent times to create a large and complex genome [19] . Despite the abundance , ubiquity and genetic contributions of TEs in plants , no previous investigation has made comprehensive efforts to fully discover or characterize all of the TEs in any angiosperm genome . Even the best annotated plant genomes , those of Arabidopsis thaliana and rice ( Oryza sativa ) , were initially examined only at a cursory level to find highly repetitive elements and those with homology to previously known TEs . Hence , subsequent studies on these genomes continue to yield new families of TEs of various types . The first exception to this rule has been the draft sequence analysis of the ∼2300 Mb maize genome , where a consortium of TE researchers has used several independent approaches in an attempt to discover and describe as many TEs as possible [20] . Even before its nearly full genome analysis , maize was the source of the best-studied TE populations in plants , including the LTR retrotransposons , where detailed analysis of small segments of the genome uncovered a great diversity of elements in different families that are mostly arranged as nested insertions [21] . The maize LTR retrotransposons were classified into 47 families [22] , and comparisons between families indicated differences in their times of transposition [23] , their preferential associations with different chromosomal regions [23]–[25] , and their levels of expression [26] . In order to fully describe the contributions of TEs to genome structure and function , one needs to first find and describe all of the TEs in a genome . Given that that average flowering plant genome is ∼6500 Mb [27] , they are expected to be composed of complex intermixtures and highly variable structures of TEs , so identification and analysis of the complete TE set will be a daunting task . Hence , we know very little about TE abundances and arrangements in anything but unusually tiny plant genomes , like those of Arabidopsis , rice and sorghum . Here , a comprehensive identification and characterization of retroelements is reported for the maize genome from inbred line B73 [20] . Hundreds of new retroelement families were discovered , and dramatic preferences in their distributions , associations and activities were uncovered . These first comprehensive studies open a window onto the true complexity of genome structure and evolution in a moderate-sized angiosperm genome .
In order to find all elements , LTR retrotransposons were sought by a combination of approaches that relied on both structure and homology , as described in Materials and Methods . The structure of an integrated LTR retrotransposon can be simply described as a terminal 5′ repeat that starts/ends in TC/GA ) , followed by a primer binding site that is used for the initiation of reverse transcription ( i . e . , replication ) , followed by polycistronic ( and sometimes frame-shifted ) genes that encode for several proteins necessary for element replication and integration , followed by a polypurine tract that is involved in the switch to second strand DNA synthesis , followed by the 3′ LTR . Searching for these canonical structures employed LTR_STRUC [28] , combined with custom Perl scripts . All intact LTR retrotransposons were identified in a set of 16 , 960 sequenced maize BACs ( bacterial artificial chromosomes ) [20] . In addition , LTR retrotransposons homologous to known TEs in the maize LTR retrotransposon exemplar database ( http://maizetedb . org/ ) were found by running the RepeatMasker program ( vers 3 . 19 ) [29] on the assembled B73 genome using default parameters . The element discovery process yielded 406 unambiguously distinct families of LTR retrotransposons that contained at least one intact member ( Table 1 ) , with intact being defined as the presence of two LTRs flanked by target site duplications ( TSDs ) . Families were defined by established sequence relatedness criteria [30] , and most families were named using the sequence-based criteria developed by San Miguel and coworkers [31] . Of these families , the great majority ( 363 ) were found by this structure-based screen and had not been previously described . A few ( 90 ) additional full-length LTR retrotransposons were identified that lacked sufficient structural or internal sequence information to allow one to determine their family status , and these are currently given the generic family name “unknown” ( see Materials and Methods ) . LINEs were detected by their TSDs flanking a block of sequence of appropriate length ( 5–10 kb for L1-like superfamily member searches and 3–5 kb for RTE-like superfamily member searches ) , terminated on one end with a simple sequence repeat , usually poly A . Further , these candidates were required to encode at least one LINE-specific protein motif . SINEs are non-autonomous retroelements that use the enzymatic machinery of autonomous LINEs to retropose ( for a review see [32] ) . SINE discovery was mainly based on the detection of the characteristic internal RNA polymerase III promoter , as described in Materials and Methods . Prior to this search , only the ZmAU SINE family had been identified in maize [33] . Using a structure-based approach , an additional three SINE families were discovered , and are now named ZmSINE1 , ZmSINE2 and ZmSINE3 ( Figure 1A ) . All four maize SINE consensus sequences possess an internal RNA polymerase III promoter composed of conserved A and B boxes , suggesting an ancestral relationship to tRNAs . As for the pSINE family in rice and the TS SINE family in tobacco [34] , [35] , ZmAU , ZmSINE1 and ZmSINE2 members ends with a poly ( T ) stretch of 4 to more than 20 bases , a feature found only in these five plant SINE families [32] . In contrast , ZmSINE3 members end with a poly ( A ) stretch , a feature found for Brassicaceae SINEs [36] as well as for all other eukaryotic tRNA-related SINEs [32] . Despite this structural difference , ZmSINE2 and ZmSINE3 likely have the same LINE partner as they show strong 3′-end sequence homologies with the maize LINE1-1 consensus sequence ( Figure 1B ) . This implies that , in the target-primed reverse transcription process leading to SINE integration by the LINE machinery , the same LINE reverse transcriptase can prime reverse transcription on a poly ( A ) as well as a poly ( U ) -ending RNA template . Because TEs in maize and other organisms tend to insert into each other , it was possible that other TE sequences inside a retroelement might be misidentified as an intrinsic part of the retroelement . Hence , all of the retroelements identified in maize were carefully compared to the comprehensive databases for other ( i . e . , class I ) TEs in maize [20] to produce a filtered set of retroelement sequences . The filtered LTR retrotransposon sequences for all 406 families were used with a RepeatMasker approach [29] to find all of the significant homologies in the B73 draft sequence [20] . At the default settings employed , similarity as small as a contiguous perfect match of 24 bp was identified as a valid homology . With this approach , over 1 . 1 million LTR retrotransposon fragments were identified in the B73 maize genome , contributing ∼1 . 5 Gb , or about 75% of the ∼2 . 05 Gb of the genome that has been sequenced ( Table 1; [20] ) . As expected , the most abundant families were those that had been previously known , like Huck , with the four most numerous families each contributing 7–12% of the nuclear DNA . The 20 most numerous LTR retrotransposon families generate ∼70% of the sequenced B73 genome ( Table 2 ) , while the remaining 386 families mostly consist of low-copy-number families with a high diversity but lesser genomic abundance ( Figure 2 and Table S1 ) . Many cases were observed of gene fragments inside LTR retrotransposons ( Table S2 ) . A total of 425 intact LTR retrotransposons were observed to contain gene fragments , from a minimum of 189 independent gene fragment captures . No case was identified , under the conditions employed , where a single LTR retrotransposon contained inserted fragments from more than one standard nuclear gene . Other classes of TEs in maize are even more active in gene fragment acquisition , including 1194 gene fragment captures by Helitrons and 462 by other DNA transposons , including Pack-MULEs [20] . It is not known whether these gene fragments play any role in maize genetic function , for instance in the creation of a new gene or in epigenetic regulation of their donor loci . Thirty different families ( with family members defined as those with >80% sequence identity [30] ) of LINEs were detected in the maize genome , with 13 of these not having been previously found and/or identified as separate families ( Table 1 ) . Approximately 35 , 000 LINEs ( many as fragments of intact elements ) were found in the B73 sequence , but this number is certain to be an overestimate caused by the many gaps and incorrect assemblies that are expected in the current maize genome draft sequence [20] . These LINEs contribute 20 Mb of DNA to the draft genome sequence , or about 1% of the total ( Table 1 ) . Overall , SINEs represent around 0 . 5 Mb and 0 . 02% of the sequenced portion of the B73 maize genome [20] . The copy numbers are 49 , 134 and 23 for the ZmAU , ZmSINE1 and ZmSINE3 families , respectively . ZmSINE2 is the major SINE family , with 1382 members . Based on phylogenetic criteria ( Figure S1 ) , the ZmSINE2 family can be further divided into three distinct subfamilies . A phylogenetic approach was used to study the amplification dynamics of SINEs in maize . The ZmSINE1 , ZmSINE2 and ZmSINE3 families contain very young members ( Figure S1 ) , close to the family consensus , suggesting very recent transposition activity . Tree topologies for these families are also typical of the “gene founder” model wherein a very small number of “master” elements are active while the vast majority of derived copies have no significant amplification potential [37] . The ZmAu family is mainly composed of more diverged members , suggesting little or no activity in the recent past . In order to look at the behaviors ( e . g . , insertion specificities or amplification level ) of the TEs across a genome , it is essential to determine their relatedness and then use this information to generate families of close relatives . Once families are generated , then family-specific behaviors can be investigated . Transposable elements of all classes tend to vary in relatedness across a spectrum , such that two TEs recently derived by transposition from the same parent element may be 100% identical in sequence , while others with a more ancient relationship can show any degree of further divergence . However , the very rapid removal of DNA from higher plant genomes [38] , [39] , especially from maize [40] , by the progressive accumulation of small deletions indicates that TEs that last shared a common ancestor more than a few million years ago ( mya ) are usually largely or fully deleted from the genome . Hence , TE families can be defined by an arbitrary but consistent criterion of nucleotide sequence divergence , and a value of 80% identity has been selected by a consortium of researchers in this field [30] . In the maize genome , the classification of LTR retrotransposons into families was a major challenge because of the exceptional complexity that was observed . Nonetheless , similar to the case in the much simpler rice genome [41] , all-by-all BLAST analysis of LTRs was sufficient to unambiguously define families by the 80% identity rule . Not all families could be classified in their appropriate superfamily ( i . e . , copia or gypsy ) , usually because of an absence of the genes needed for the definitive gene order criterion or for phylogenetic analysis , and these were dubbed RLX . The individual family identifications were clear , however , and each family was given a unique name . Some of these family designations conflict with previous names [42] , but these earlier names were not applied with any specific rule , and thus were certain to be both misleading and temporary . For instance , the LTR retrotransposon collection called CRM [20] was actually found to represent four related , but clearly separate , LTR retrotransposon families that we have now named CRM1 , CRM2 , CRM3/CentA , and CRM4 . Our consistent analysis using agreed-upon criteria [30] caused other such splittings of previously lumped families , and also lumped some different named families into single families that fit the 80% identity criterion ( e . g . , Cinful and Zeon are actually a single family that has now been named Cinful-zeon ) . The new names , and the names that had previously been applied by unspecified and/or inconsistent homology criteria , are now shown in Table S1 . The assembled physical and genetic map of maize inbred B73 [20] allows placement of any class of sequence along that portion of the genome that was sequenced . Overall , LTR retrotransposons are found to be most abundant in pericentromeric heterochromatin and least abundant in the more gene-rich arms on all chromosomes ( Figure 3 ) . However , different LTR retrotransposons are found to be differentially clustered in such analyses , with the general observation that the gypsy superfamily of LTR retrotransposons is concentrated in the pericentromeric heterochromatin while the copia superfamily shows a preferential accumulation in the more euchromatic regions of the chromosome arms [20] . Despite this general pattern , individual families show deviations from the rule . For instance , the gypsy family Huck was found to exhibit a more ‘copia-like’ distribution on chromosome 1 ( Figure S2 ) . Another gypsy family , Grande , shows a relatively even distribution across 10 Mb bins of this same chromosome . Hence , there are families that accumulate in a pattern that contrasts with the general behavior of their superfamilies in maize . A more dramatic correlation between LTR retrotransposon family property and insertion/accumulation pattern was observed by comparing the copy numbers of intact elements in a LTR retrotransposon family with the nature of the sequences within 500 bp ( on each side ) of the insertion site . Low-copy-number families were found to be most often inserted into the regions in or near genes ( or gene fragments ) , while high-copy-number families were observed to primarily accumulate inside other LTR retrotransposons ( Figure 4 ) . LINEs of both RIT and RIL ( L1-like ) families were found to be fairly evenly distributed across all chromosomes , with a higher abundance in distal regions of the chromosomes ( Figure S3 ) . Although maize LINEs have been observed to show a preferential association with genic regions , especially introns [43] , their common occurrence in pericentromeric DNA suggests that many insertions are not in or near genes . Of the 1991 SINEs discovered , 1174 were found in the introns or UTRs ( untranslated regions ) of genes and 21 in putative coding exons ( data not shown ) . Only 796 were found in the intergenic space that makes up more than 85% of the sequenced B73 genome [20] . Hence , like SINEs in other species , these small TEs show a very strong preference for association with genes in the maize nuclear genome . In this regard , the general distribution of SINEs across the maize chromosomes ( Figure S4 ) was found to exhibit a pattern quite similar to the gene distribution [20] . As previously observed in other organisms by numerous scientists studying many different genomes , maize TEs were found to make up a greater quantity of the total DNA in the gene-poor pericentromeric regions than in other parts of the genome ( Figure 3 ) . However , as mentioned above and observed previously ( reviewed in [44] ) , LINEs , SINEs and some LTR retrotransposon families accumulate preferentially in areas that are near genes . Figure 5 shows the relationship between LTR retrotransposon abundance and LTR retrotransposon family richness across chromosome 1 of maize inbred B73 , and this general pattern was found to be the same across all other chromosomes ( data not shown; Table S3 ) . Hence , on all maize chromosomes , those regions that have the most total LTR retrotransposons also have the fewest kinds of LTR retrotransposons . This observation echoes the relationship between the number of species and the abundance of individual species in both terrestrial and aquatic environments , but has no precedent that we are aware of in TE studies . The insertion dates of intact LTR retrotransposons was observed to vary according to the distance from the centromere . Younger elements are enriched in the euchromatic regions whereas older elements are most abundant in the pericentromeric regions ( Figure 6 ) . An analysis of variance showed that the average insertion date per 1 Mb bin varied according to distance from the centromere ( F = 2 . 08; P<0 . 0001 ) , and this relationship held across most of the chromosomes ( Table 3 ) . The average date of LTR retrotransposon insertion for a given family was also observed to correlate with the current perceived copy numbers of the LTR retrotransposon families . As a general pattern , the lower-copy-number elements were more ancient insertions ( averaging about 1 . 2 mya ) compared to the highest-copy-number elements ( averaging about 0 . 7 mya ) ( Figure 7 ) . Because most of the higher-copy-number LTR retrotransposons are of the gypsy superfamily ( Table 2 ) , and show an overall pericentromeric accumulation bias [20] , one expected the opposite result because of the slower rate of LTR retrotransposon removal in gene-poor ( and thus recombination-poor ) regions like the pericentromeres [45] .
The landmark sequencing of the very complex and fairly large maize genome was accomplished at a small fraction of the cost of previous clone-by-clone sequencing projects because of the expertise of the researchers involved , a low redundancy of initial shotgun sequencing , and because of a decision to not finish any regions of the genome that appeared to lack gene candidates [20] . Hence , a very comprehensive TE discovery and masking process was necessary to facilitate finishing that was efficiently targeted on genes . One disadvantage of this approach , however , is that most sequenced regions are composed of many tiny contiguous sequences ( contigs ) . Our analysis of the current B73 assemblies ( data not shown ) indicates a median contig size of ∼7 kb with ∼60% of the assembly occurring in contigs larger then 30 kb . Thus , a structure-based search approach that requires intact elements , like the one employed in this project , will miss any families where the only intact members are fractured by sequence gaps or inaccurate scaffolding of contigs . This is expected to be most problematic for large TEs ( like LTR retrotransposons ) and for those that only have a few intact members . Hence , our prediction that ∼75% of the B73 maize genome is composed of LTR retrotransposons is a minimum estimate . Also because of the many tiny sequence gaps in the assembly , there will be many occasions when an intact retroelement was identified by RepeatMasking as several fragments of an element . Hence , calculation of the ratio of intact to fragmented LTR retrotransposons is not valid with this dataset . In contrast , this same analysis with the random sampling of fully sequenced and annotated clones known as the GeneTrek approach does allow accurate quantification of the relative abundance of different TE structures . In such a GeneTrek analysis , the ratio of intact to truncated LTR retrotransposons in maize was found to be ∼2∶1 [40] , [46] , quite different from the ratio of ∼1∶27 that was calculated ( Baucom and Bennetzen , data not shown ) as an artifact of this same analysis on the currently fractured B73 assembly [20] . There are also many large sequence gaps , and numerous sequenced BACs with no home in the assembly , for the B73 draft sequence [20] . It is likely that about 90% of the maize nuclear genome is present in the current assembly ( ∼2005 Mb out of ∼2300 Mb ) . From all previous full genome sequences in multicellular eukaryotes that have centromeres , the standard observation has been that the majority of the unsequenced regions are in the gene-poor areas around the centromeres and in other heterochromatic blocks . Because these gene-poor chromosome segments also tend to be LTR retrotransposon-rich , these results provide a further reason to believe that the B73 maize genome contains more than 75% LTR retrotransposons , with an upper limit of ∼85% . Importantly , however , the overall quantitation of retroelement contributions to the B73 genome is not dramatically biased by the gaps and other intrinsic errors in the current assembly . As shown in Figure S5 , most LTR retrotransposons exhibit the same relative abundance when used to mask the current B73 draft assembly as they do when used to mask a shotgun dataset from the same B73 line ( R2 = 0 . 99 , p<0 . 0001 ) . The few exceptions to this observation ( e . g . , Ipiki ) are likely to be LTR retrotransposons that are preferentially abundant in that ∼10% ( e . g . , near centromeres ? ) of the maize genome that is not present in the assembly [20] . Previous maize studies had uncovered primarily the high-copy-number retroelements [21] , [23] , with some exceptions of low-copy-number TE discovery associated with particular mutations [47] , [48] or carefully sequenced and annotated small segments of the maize genome [46] . All of the LTR retrotransposons found in these earlier studies were also found in this analysis , at the approximate predicted frequencies . The major difference , however , was the large dataset available in the current study , and thus the discovery of hundreds of additional LTR retrotransposon families . Only by this comprehensive analysis on the majority of the maize genome was it possible to determine the exceptional complexity of retroelements in maize , and their different properties of dispersal and divergence . Rice , with an ∼400 Mb nuclear genome , has 172 identified LTR retrotransposon families that contribute ∼97 Mb , distributed across 48% with only a single intact element , 20% with 2 intact elements and 32% with 3 or more intact elements [41] . Maize , in contrast , has 406 identified LTR retrotransposon families , just over twice as many , but they contribute ∼1700 Mb of DNA to the maize nuclear genome . These maize elements are distributed across 42% singleton intact elements , 21% with 2 intact elements and 37% of families with 3 or more intact elements . Hence , the >17X greater amount of LTR retrotransposons in maize compared to rice is not primarily caused by a greater number of element families in maize but instead by a much higher copy number of a very small number of superabundant families . Two of the many misconceptions about TE properties in higher eukaryotes are that they are highly repetitive and are randomly scattered about the genome . In fact , many TE families are present in very low copy numbers . The median family copy number of intact LTR retrotransposon with TSDs in B73 maize was measured to be 2 ( mean ∼77 ) , with a total of 256 families that contained only one or two intact LTR retrotransposons that were detected . Most LTR retrotransposon families are distributed quite unevenly across the genome , probably an outcome of both differences in insertion preferences and different rates of LTR retrotransposon removal in different chromosomal locations [44]–[46] , [49] . The previous observation that LTR retrotransposons show a dramatic bias in whether they insert into LTRs or the internal regions of other LTR retrotransposons [21] was not observed , however , and it now seems likely that the previous conclusion was an artifact of a small sample size . Studies in rice and other organisms suggest that LTR retrotransposons are more rapidly removed ( sometimes by unequal homologous recombination to generate solo LTRs ) in regions with high recombination rates , like areas around genes and in the cores of centromeres [45] , [46] . One example of this analysis was that the ratio of solo LTRs to intact elements was found to be much higher in gene-rich and recombination-rich euchromatic regions than in gene-poor and recombination-poor pericentromeric regions [44] . Although natural selection should also more rapidly remove individuals from a population that contain retroelements or other TEs detrimentally inserted into coding and gene regulatory regions , this process alone cannot explain the differential retroelement accumulation properties that we observe . For instance , why would LINEs , SINEs and low-copy-number LTR retrotransposons not be depleted in genic regions , while high-copy-number LTR retrotransposons are ? A simpler explanation is that different retroelements are directed to preferentially insert in different parts of the genome by the biases of their integrases for association with specific chromatin proteins , as observed with Ty elements in yeast [50] . We have no idea how many types of DNA::protein configurations are actually present in plants , of course , but it is very clear that chromatin consists of more than just hetero- and eu- varieties [51] , so sufficient variability should be present to allow a great wealth of different TE insertion specificities , as has been recently reported in Arabidopsis [52] . Particularly fascinating are the high-copy-number LTR retrotransposons like Ji and Opie that preferentially avoid insertion into genes , but primarily insert into heterochromatin near genes , while other high-copy-number elements like Gyma avoid inserting into genes or heterochromatin near genes , preferring instead an accumulation into large gene-free heterochromatic blocks [46] . Unlike low-copy-number LTR retrotransposons , which are associated with de novo mutations in many plant species , neither class of high-copy-number LTR retrotransposons is associated with a mutation caused by insertion into a gene . Perhaps TE insertion profiles will be a uniquely useful route to uncover and map a broad spectrum of novel chromatin structures . Genomic complexity is not just a matter of the number of different sequences , but also of the variability in their arrangement and stability . The factors that determine differences in these arrangements , such as differential insertion specificities and differences in retention , are only beginning to be understood . It is already clear , though , that TE insertion and retention biases are the major forces that determine local genome structure in maize and other complex plant genomes . The mechanisms responsible for these biases , and their outcome vis-à-vis gene/genome function and evolution , are only now beginning to be understood . Viewed from the standpoint of the TE , much of the diversity in TE populations and their arrangement takes on a new and informative light . A previous model proposed that low-copy-number TEs must insert near or into genes so that they have a reasonable chance of expression and activity in subsequent generations , while highly repetitive TEs need to avoid insertions that disrupt genes in most cases because 1000 or 10 , 000 such insertions would lead to a dead host [44] . Hence , abundant TEs rely on their abundance per se to guarantee transmission and the opportunity for activity in future generations . The data for LTR retrotransposon abundance versus copy number shown here agrees with this model , as does the fact that ( to date ) none of the high-copy-number LTR retrotransposons have been shown to cause a de novo mutation , while low-copy-number LTR retrotransposons ( e . g . , Bs1 , Tnt1 , Tos17 ) that make up a relatively small part of their genomes have caused many new mutations [47]–[49] , [53] . The analysis of the maize genome suggests that the copy number for this transition is fairly low , 10–100 intact copies per genome ( Figure 4 ) , for this change in lifestyle . LTR retrotransposon families with copy numbers less than ten were usually found to preferentially accumulate in genic regions , while most LTR retrotransposon families with copy numbers higher than 100 were found to be enriched in gene-poor regions like pericentromeric heterochromatin . The insertion preferences of LTR retrotransposons can contribute to their potential for more than just transcriptional activity . Elements that land in recombination-rich regions have a greater chance of inter-element unequal events that can create novel LTR retrotransposons with possible new properties [38] . Insertion into an LTR provides the opportunity to acquire the gene regulatory properties of the target LTR retrotransposon . Moreover , insertion of an LTR retrotransposon into an LTR retrotransposon would usually eliminate the target element as a potential competitor for future amplification . The observed relationship between LTR retrotransposon family richness and LTR retrotransposon abundance across the maize chromosomes is the most compelling indicator , in this study , of the validity of the conceptualization of TEs as competitor organisms whose world is the nuclear genome . When an environment is highly suitable for proliferation of a category of life , a few highly adapted types of individuals ( e . g . , species or , in this case , LTR retrotransposons ) crowd out all other competitors to create a dense but diversity-poor ecosystem . Other species , here proposed to be the lower-copy-number LTR retrotransposons , disseminate themselves at lower abundances across less productive environments that thus become diversity-rich . Of course , it is not at all clear what aspect ( s ) of these TE-enriched regions might make them “productive” from a TE perspective . Perhaps it is something as simple as a lower rate of TE removal by ectopic recombination [45] . This view of genomic life provides another angle to investigate TEs , as highly adapted commensals , but in no way suggests that they cannot be utilized when the opportunity arises for a process that benefits the plant host . The occasional creation of new genes by TE capture and shuffling of gene fragments or through fusion of TE genes ( or regulatory regions ) with nearby genes falls into this category . What remains constant in these considerations is the long-term evolutionary value of the instability and diversity generated by retroelements and other TEs .
New families of maize LTR retrotransposons were discovered by several iterations of masking and re-investigation . First , 5 , 075 maize BACs were downloaded on February 22 , 2007 from the Washington University maize sequencing project [20] and masked using the RepeatMasker program [29] with a database of previously known maize LTR retrotransposons . Masked regions were removed from the sequence , and LTR_STRUC [28] was used to find new elements . This program identifies LTR retrotransposons based on the presence of LTRs , matching target site duplications ( TSDs ) , and the presence of the canonical TG/CA motif found at the 5′ and 3′ end of each LTR ( although deviations are permitted ) , and thus is a structure-based screen rather than one that requires sequence homology to a known TE . This process was designed to uncover old and fragmented families of LTR retrotransposons after masking out the younger and previously discovered families [21] , [22] . Next , 15 , 708 maize BAC sequence data sets were downloaded March 1 , 2008 from the Washington University sequencing project and were first masked at a quality score of ‘40 , ’ then screened with LTR_STUC . 13 , 362 LTR retrotransposons were found and , along with the sequences uncovered in the initial screen , placed into families using the RepMiner classification tools ( http://repminer . sourceforge . net/ ) [54] . This process generated ∼600 maize LTR retrotransposon exemplar sequences that best describe each of 412 identified families . Each exemplar was annotated for LTR position , the primer-binding site sequence and the genes involved in the transposition process . Exemplars were identified as members of either the copia or gypsy superfamilies based on the position of the reverse transcriptase gene in relation to the integrase gene , and by using a maximum-likelihood gene tree of reverse transcriptase . Both methods of superfamily designation were 100% congruent . Exemplar sequences that did not contain internal coding regions with an identifiable homology to LTR retrotransposon genes were given the ‘unknown’ superfamily designation . Each exemplar was hand-curated to ensure that exemplars where not chimeric annotations that contained insertions of other LTR retrotransposon sequences . DNA transposons inserted within the LTR retrotransposon exemplars were identified by homology-based searches against the maize TE database ( http://maizetedb . org/ ) and were excluded from the exemplar sequence by masking . Family nomenclature follows established methodology [30] in which the TE classification can be deduced from the full family name . In this system , family names are given a three character prefix that represents the class , order and superfamily of the individual family . For example , families with the RLG prefix represent LTR retrotransposons that are members of the gypsy superfamily while the RLC prefix represents families that are members of the copia superfamily . LTR retrotransposons that could not be assigned to the gypsy or copia superfamilies were assigned the RLX prefix . The B73 maize genome represented as an Accessioned Golden Path ( AGP ) assembly [20] was downloaded from the Arizona Genomics Institute ( http://www2 . genome . arizona . edu/genomes/maize ) . This dataset was investigated for LTR retrotransposon content using the default settings in RepeatMasker [29] with the curated exemplar library of maize LTR retrotransposons ( http://maizetedb . org/ ) . The RepeatMasker annotation of the maize AGP assembly was uploaded to a custom MySQL relational database to facilitate manipulation and querying of sequence features mapped onto the maize genome assembly . The RepeatMasker output files derived from masking the AGP with the exemplar database were translated to General Feature Format ( GFF ) style coordinates using the cnv_repmask2gff . pl program [55] . These coordinates were uploaded to a MySQL database using custom Perl scripts . The database served as the query engine to trim overlapping features resulting from the RepeatMasker annotation and provided the framework to query distribution related information . The MySQL database schema and custom Perl scripts used to generate the non-redundant distribution information are available from the authors upon request . Each of the AGP chromosomes was spatially binned into 10 Mb non-overlapping units and the percent LTR retrotransposon composition within each bin was determined , as was the number of distinct families present within each bin . The strength and direction of the correlation between percent LTR retrotransposon composition and family richness was determined using the Resample program [56] separately for each chromosome . The sequence files for the 16 , 007 BAC assemblies incorporated in the maize AGP were downloaded from GenBank . Full-length LTR retrotransposons were identified by LTR_STRUC and mapped onto these BACs through the use of batch annotation scripts available in the DAWGPAWS annotation package [55] . This process resulted in a database of 35 , 229 full-length LTR retrotransposons . The 5′ LTR sequences of this dataset of full-length LTR retrotransposons were used to classify the elements into families using at least 80% identity in a BLASTn analysis employing the exemplar database . LTR retrotransposons that were not homologous to families present within the exemplar database ( 1 , 979 ) were removed from analysis , with the exception of the gene capture analysis , explained below . Further , sequences that were 2 standard deviations greater in length than the assigned family's mean length ( 2 , 135 ) were also removed from analysis . These sequences were found to harbor full-length insertions of other LTR retrotransposons and thus do not provide an accurate characterization of the most recently intact elements . The resultant database of full-length LTR retrotransposons consisted of 31 , 115 individual sequences distributed among 406 distinct families . Six families initially identified on the maize BACs used to create the exemplar database were not found in the current assembly of the AGP , potentially due to the fact that 981 BAC sequences released from the Washington University sequencing effort were not used to assemble the AGP . The location of full-length LTR retrotransposons on the AGP was determined using the data conversion table provided by the Arizona Genomics Institute . The insertion date of each full-length LTR retrotransposon was determined by estimating the amount of divergence between the 5′ and 3′ LTRs [23] . Perl programs were used to automate this process; the two LTRs of each mined LTR retrotransposon were first aligned using ClustalW [57] , and the genetic divergence between the two was estimated using the baseml module of PAML ( [58] , vers . 4 ) . The time since insertion of each LTR retrotransposon element was estimated using the substitution rate of 1 . 3×10−8 per site per year [11] . To determine if distance to the centromere explained variation in insertion dates , the GLM procedure of the SAS statistical package ( vers . 9 . 2 ) was used to perform an analysis of variance with the square-root transformed average insertion date per bin as the dependent variable and the distance of each bin to the centromere as the independent variable . This analysis was performed separately for each chromosome . Investigation into the insertion-site specificity of each full-length LTR retrotransposon was conducted by a performing a BLASTn search to four separate databases , namely those containing maize genes [20] and those containing DNA transposable elements , Helitrons , and LTR retrotransposons ( http://maizetedb . org/ ) . 500 bp of maize sequence flanking the 3′ and 5′ sides of each element was used as the query in separate nucleotide BLAST analyses , and the results were parsed for at least 80% identity . No annotations >5 bp away from the query sequence were included , because the objective was to determine what type of sequence the LTR retrotransposons inserted into , rather than those sequences that were simply nearby . A set of curated genes from the rice genome ( RAPDB , vers . 4 ) was used to search the full-length maize LTR retrotransposons for instances of host gene capture . The full-length LTR retrotransposon dataset was screened for homology to rice genes at an Expect value of e−5 . Significant BLAST hits were screened for TE genes , and genes were also removed if annotated as ‘rice gene family candidate’ and present in high copy number ( >20 ) , as they are likely to be undiscovered TE genes . The full-length LTR retrotransposons that were not placed into families based on the 80% identity rule were retained in this analysis as they represented ∼20% of the total gene capture events . The annotations of these particular LTR retrotransposons indicated that they exhibit general LTR retrotransposon features , such as target site duplications and a TG/CA motif at the end of the LTRs , and as such represent LTR retrotransposons of ‘unknown’ family classification . Trace files of whole genome shotgun ( WGS ) DNA sequence reads for maize inbred B73 were obtained from those deposited by the Joint Genome Institute ( JGI ) to the NCBI Trace Archive ( http://www . ncbi . nlm . nih . gov/Traces/trace . cgi ? ) . These sequence files were trimmed of low quality bases and vector sequence using Lucy [59] . Organellar sequences were identified by BLAT [60] . Alignments to maize chloroplast and mitochondrial DNA and were removed from further analysis . This filtering resulted in a dataset of 1 , 028 , 203 high quality sequence reads totaling 79 , 6326 , 632 bp of genomic DNA . These data represent an approximately one-third sample sequence coverage of the B73 genome . The JGI shotgun data were annotated for LTR retrotransposons using RepeatMasker ( [29] , vers . 3 . 19 ) with the same database and parameter set used to annotate the AGP . Overlapping features from the RepeatMasker output were identified using the same methodology described for LTR retrotransposon annotation of the maize AGP assembly . Significant outliers between the ratios found in the AGP and the ratios found in the JGI shotgun data were identified by performing an outlier analysis in the SAS statistical package ( vers . 9 . 2 ) . The approach to identify potential SINE families was divided into several steps . The first step was the search for anchors , which were defined as small regions containing SINE features ( see below ) . Following that , a 500 bp region flanking the anchor on each side was extracted . These sequences were used to perform a non-stringent search for direct repeats ( likely to be TSDs ) that were less than 350 bp apart . The sequences that passed the filter were aligned using ClustalW [57] , alignments were refined using muscle [61] and corrected by hand using Seaview [60] . A first approach for SINE identification consisted in developing an hmm model using hmmer ( http://hmmer . wustl . edu ) for the region harboring the main anchor , which is the internal ( tRNA-related ) promoter for RNA polymerase III , defined for SINEs as an “A” box ( RRYNNRRYGG ) around position +14 of the start of the repeat and a B box ( GGTTCGANNCC ) around position +54 of the start of the repeat . This anchor was designed using known plant SINE elements . This model was then used to search the whole pseudomolecule representing the draft sequence of the B73 maize genome [20] . A second approach consisted in identifying tRNAs using tRNAscan-SE and using those sites described as “Pseudo tRNAs” as anchors . A third approach consisted in using the last 30 bases of maize LINE consensus sequences to screen for homology by BLASTn against the B73 draft genome , and to then use these homologies as anchors . In this case , to make sure that SINEs were distinguished from severely truncated LINEs , these homologies were searched for the presence of internal A and B boxes typical of tRNA-derived SINEs . A search for 5S RNA-derived SINEs was also performed , using as anchor the A/IE/C conserved boxes of the 5S RNA internal polymerase III promoter , without success . SINEs that did not share significant sequence identity ( <50% ) outside of the common SINE features ( internal polymerase III promoter and 3′-terminal end ) were classified in distinct families . For SINEs that do have significant homologies ( >50% ) outside of the common SINE features ( >50% ) , further subfamily classifications were proposed using phylogenetic criteria . The SINE sequences were aligned using the ClustalW multiple-alignment program [57] with some manual refinements ( i . e . , elimination of unnecessary gaps at the beginning and end of the ClustalW alignment ) . Evolutionary distances were calculated using the Jin-Nei distance method of the Dnadist program ( PHYLIP package version 3 . 573c [62] . The coefficient of variation of the Gamma distribution ( to incorporate rate heterogeneity ) and the expected transition to transversion ratio ( t ) were obtained by pre-analyzing the data with the Tree-Puzzle program [63] . Phylogenetic trees were inferred using the Neighbor-Joining ( NJ ) method ( PHYLIP package version 3 . 573c [62] ) . Consensus trees were inferred using the Consense program ( PHYLIP package ) . The significance of the various phylogenetic lineages was assessed by bootstrap analyses [64] .
|
Although TEs are a major component of all studied plant genomes , and are the most significant contributors to genome structure and evolution in almost all eukaryotes that have been investigated , their properties and reasons for existence are not well understood in any eukaryotic genome . In order to begin a comprehensive study of TE contributions to the structure , function , and evolution of both genes and genomes , we first identified all of the TEs in maize and then investigated whether there were non-random patterns in their dispersal . We used homology and TE structure criteria in an effort to discover all of the retroelements in the recently sequenced genome from maize inbred B73 . We found that the retroelements are incredibly diverse in maize , with many hundreds of families that show different insertion and/or retention specificities across the maize chromosomes . Most of these element families are present in low copy numbers and had been missed by previous searches that relied on a high-copy-number criterion . Different element families exhibited very different biases for accumulation across the chromosomes , indicating that they can detect and utilize many different chromatin environments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/genome",
"projects",
"genetics",
"and",
"genomics",
"genetics",
"and",
"genomics/plant",
"genomes",
"and",
"evolution",
"genetics",
"and",
"genomics/genomics"
] |
2009
|
Exceptional Diversity, Non-Random Distribution, and Rapid Evolution of Retroelements in the B73 Maize Genome
|
Human cytomegalovirus ( HCMV ) infections are life-threating to people with a compromised or immature immune system . Upon adhesion , fusion of the virus envelope with the host cell is initiated . In this step , the viral glycoprotein gB is considered to represent the major fusogen . Here , we present for the first time structural data on the binding of an anti-herpes virus antibody and describe the atomic interactions between the antigenic domain Dom-II of HCMV gB and the Fab fragment of the human antibody SM5-1 . The crystal structure shows that SM5-1 binds Dom-II almost exclusively via only two CDRs , namely light chain CDR L1 and a 22-residue-long heavy chain CDR H3 . Two contiguous segments of Dom-II are targeted by SM5-1 , and the combining site includes a hydrophobic pocket on the Dom-II surface that is only partially filled by CDR H3 residues . SM5-1 belongs to a series of sequence-homologous anti-HCMV gB monoclonal antibodies that were isolated from the same donor at a single time point and that represent different maturation states . Analysis of amino acid substitutions in these antibodies in combination with molecular dynamics simulations show that key contributors to the picomolar affinity of SM5-1 do not directly interact with the antigen but significantly reduce the flexibility of CDR H3 in the bound and unbound state of SM5-1 through intramolecular side chain interactions . Thus , these residues most likely alleviate unfavorable binding entropies associated with extra-long CDR H3s , and this might represent a common strategy during antibody maturation . Models of entire HCMV gB in different conformational states hint that SM5-1 neutralizes HCMV either by blocking the pre- to postfusion transition of gB or by precluding the interaction with additional effectors such as the gH/gL complex .
Human cytomegalovirus ( HCMV ) belongs to the family of β-herpes viruses and is a clinically important pathogen . While infection in hosts with a functional immune system is usually clinically asymptomatic , the virus can cause significant morbidity and mortality in individuals with an immature or suppressed immune system . As such , the virus still represents a potentially severe clinical complication in transplant recipients [1] . Congenital HCMV infection is also the most common infectious cause of neurological disorders in children [2] . Hence , the development of vaccines against HCMV is considered a top priority [3] . Herpes viruses enter cells via a cascade of molecular interactions , which ultimately results in the fusion of the viral envelope with target cell membranes . In an initial step the virus attaches to the target cell surface via a non-specific , low-avidity binding to heparan sulfate proteoglycans and in subsequent steps interacts with more specific , higher avidity receptors ( for review see [4] ) . While for some herpes viruses cellular receptors and their viral ligands have been well characterized , the situation is less clear for HCMV . On the host side , different molecules such as integrins [5] , EGFR [6] or PDGF-α receptor [7] have been postulated as specific receptors . The viral ligand that was described in these studies for HCMV was in all cases glycoprotein B ( gB ) . However , some of these findings were also challenged [8] , [9] . Receptor binding initiates a cascade of events that enables fusion of viral and cellular membranes . The core fusion complex for herpes viruses in general consists of gB and gH/gL ( reviewed in [10] ) . In the case of HCMV , gH/gL associated proteins such as gO or the UL128-131 complex determine cell tropism and/or mode of entry [11] , [12] . Fusion takes place at the plasma membrane in the case of fibroblasts [13] whereas endo-/epithelial cells are infected by fusion in an endocytic compartment [14] , [15] , indicating that the fusion complex is functional in different pH-environments . Events similar to macropinocytosis may also be involved in HCMV infection of certain cell types , highlighting that the overall situation is relatively complex [8] , [16] . It is current consensus that within the core fusion complex , gB represents the actual fusogen while gH/gL function as accessory proteins , which activate gB . Crystal structures of gB from herpes simplex virus type 1 ( HSV-1 ) and Epstein-Barr virus ( EBV ) revealed that gB shares structural similarities with other type III fusion proteins from unrelated viruses such as VSV-G and baculovirus gp64 [17]–[21] . These data suggested that currently available gB structures represent the postfusion form rather than prefusion form of the proteins , although this is not proven . In analogy to VSV-G , where structures of the post- and prefusion forms are available , gB is expected to undergo a conformational transition from its original prefusion form as present in the viral envelope to the final postfusion form occurring after fusion of the viral and host cell membranes [19] , [20] . The central role of gB during the fusion event makes gB a prime target for host defense mechanisms . Indeed , virtually all HCMV infected individuals develop antibodies against gB , and the neutralizing capacity of sera from HCMV convalescent individuals correlates with the anti-gB antibody titer [22] , [23] . Due to its high immunogenicity , HCMV gB was selected as vaccine antigen , and phase II clinical trials using a recombinantly produced gB protein have been conducted . Partial protection was seen both in seronegative women or kidney transplant recipients [24] , [25] . We recently isolated a panel of human monoclonal antibodies ( mAb ) against gB from healthy HCMV seropositive donors that target two novel antigenic domains ( AD ) in gB , termed AD-4 and AD-5 [26] . Several antibodies directed either against AD-4 or AD-5 showed neutralizing activity in nanomolar concentrations in in vitro assays [26] . These antibodies were shown to be effective in a postadsorption step of infection , to neutralize different virus strains and , importantly , to block infection of different target cells types such as fibroblasts , epithelial cells and dendritic cells with similar efficiency [26] . Thus , these antibodies can be considered broadly neutralizing . SM5-1 , directed against AD-4 , was identified as the most potent neutralizing mAB in this panel . AD-4 structurally corresponds to domain II ( Dom-II ) of the crystal structure of the homologous gB protein from HSV and EBV [17] , [18] . The fold of Dom-II resembles that of PH domains , and in all crystal structures of type III fusion proteins , Dom-II is formed by a discontinuous protein sequence and comprises in strain HCMV AD169 amino acids 121–132 and 344–438 of gB [26] , [27] . Interestingly , in case of HSV it was shown that antibody binding to Dom-II not only neutralizes HSV but at the same time blocks interaction with gH/gL [28] . Since the sequence of Dom-II is highly conserved between different HCMV isolates and since interaction of gB with gH/gL is a prerequisite for activation of the fusion machinery in all herpes viruses blocking this interaction by molecular effectors appears to be a promising strategy for preventing infections . Here , we report on the first crystal structure of a complex between a neutralizing antibody and a herpes virus envelope protein domain , namely the Fab fragment of the AD-4 -specific monoclonal antibody SM5-1 in complex with an isolated HCMV gB Dom-II . The structure shows that the molecular determinants for picomolar binding cluster within light chain complementarity determining region 1 ( CDR L1 ) and predominantly within an unusually long heavy chain CDR H3 of the antibody . An alignment of additional antibody clones from the same germline lineage in combination with structural analysis and molecular dynamics ( MD ) calculations highlighted important structural aspects of the somatic maturation process and the concomitant emergence of high affinity antibodies . We expect the results of our study to inform the structure-based design of anti-HCMV vaccines .
The crystal structure of an engineered variant of Dom-II from HCMV gB was solved in complex with the antigen-binding fragment ( Fab ) of antibody SM5-1 at 2 . 1 Å resolution ( Table 1 , Fig . 1 ) . In addition the structures of engineered Dom-II and of the SM5-1 Fab fragment were solved individually at resolutions of 1 . 8 and 1 . 9 Å , respectively ( Table 1 , Fig . 2 ) . In the engineered Dom-II fragment , residues 112–132 and 344–438 of HCMV gB were linked by an artificial 5-residue-long linker segment in order to produce a sequence-contiguous Dom-II domain ( Fig . 2 ) [26] . Although no electron density is visible for the linker residues and a few N- and C-terminal residues in the structure of unbound Dom-II , the overall structure compares well with the structures of discontinuous Dom-IIs from full-length gB proteins [18] . When considering all common main chain atoms , the structure of engineered unbound HCMV Dom-II differs from that of Dom-II from HSV-1 gB ( PDB entry: 2gum [18] ) by an r . m . s . deviation of 1 . 7 Å and from Dom-II from EBV gB ( PDB entry: 3fvc [17] ) by a deviation of 1 . 5 Å . These deviations closely match expectations from sequence differences . Whereas the overall sequence identity between HCMV and HSV-1 gB is 28% ( UniProt sequences P06437 vs P06473 [29] ) , the identity between the corresponding Dom-II segments is 34% ( HCMV gB residues 119–132 and 344–438 versus HSV-1 gB residues 142–153 and 364–459 ) . HCMV gB Dom-II aligns significantly worse with analogous VSV-G Dom-III in either the pre- or postfusion form ( PDB entries 2j6j and 2cmz [19] , [20] ) . Here , the r . m . s . deviations range from 3 . 1 to 3 . 3 Å . HCMV gB Dom-II folds into an all-antiparallel β-sheet sandwich and consists of a 4-stranded and a 5-stranded β-sheet that are oriented at almost a right angle to each other . A five-turn-long α-helix interconnects two of the β-strands in the 5-stranded β-sheet . As has been noted before the domain topology of Dom-II resembles that of pleckstrin homology ( PH ) domains with the caveat that the mentioned α-helix is absent in PH domains [18] , [30] . Moreover , PH domains display an alternative α-helix , namely after the last β-strand . This helix is missing in EBV gB and in a previous HSV-1 gB structure ( PDB entry 2gum ) [17] , [18] . However , such a PH domain-like helical segment extends from Dom-II in a more recent crystal structure of HSV-1 gB ( residues 463–475 , PDB entry: 3nwf ) [31] . It should be noted that this helical extension is not present in the HCMV gB Dom-II structure reported here , since the corresponding residues ( 442–454 ) were not included in engineered Dom-II [32] . Also , HCMV gB , in contrast to HSV-1 gB , is posttranslationally cleaved after position 459 and this might influence the local structure [33] , [34] . The closest structurally related PH domain that can be identified by a DALI search is that of the Rho GTPase-activating protein 27 ( ArhGAP27 , PDB entry 3pp2 ) . The r . m . s . deviation is as high as 3 . 0 Å when calculated for 75 out of 115 possible equivalent Cα positions [35] . Until now , no indications exist that Dom-IIs in herpes viruses use any of the canonical ligand interactions sites observed in PH domains for binding either phosphotyrosines , polyproline helices or inositol phosphate head groups [30] . To our knowledge , this also extends to protein-protein interaction modes [30] . Interestingly , with respect to the possible involvement of Dom-II in protein ligand interactions , we observe extensive residual positive difference electron density along a face of Dom-II of HCMV gB , namely close to Arg131 and Glu422 ( Fig . 2E ) . Although the identity of the ligand cannot be inferred from the composition of the crystallization solution , the elongated shape of the electron density suggests that the ligand might correspond to a peptide . The ligand straddles a hydrophobic pocket in Dom-II and surprisingly the same pocket is also targeted by CDR H3 of the antibody SM5-1 in the structure of the antibody antigen complex ( see below ) . The structure of unbound SM5-1 resembles that of other Fab fragments . No significant conformational changes occur in either SM5-1 or Dom-II upon complex formation with exception of the conformation of CDR H3 of SM5-1 and changes in the elbow angle ( Fig . 2D ) . In the complex , CDR H3 participates in extensive contacts with Dom-II , and all residues within CDR H3 become clearly defined in the electron density map ( Fig . 1D ) . Overall , the r . m . s . deviations between the free versus bound structures of Dom-II and SM5-1 are very low . They range from 0 . 27 Å for the VH-VL domain pair of SM5-1 to 0 . 33 Å for Dom-II and 0 . 72 Å for the CH1-CL domain pair of SM5-1 . Upon formation of the complex , the elbow angle between the variable and constant domains of SM5-1 changes from 170° in the free state to 195° in the bound state ( Fig . 2D ) [36] . Inspection of the antigen-antibody-combining site shows that HCMV gB Dom-II recognition is achieved predominantly via an unusually long heavy chain CDR H3 with additional contributions from CDR L1 ( Fig . 1E , Fig . 3 ) . Of the total solvent accessible surface area of 775 Å2 buried by SM5-1 in the complex interface , H3 contributes 515 Å2 ( 66 . 5% ) and L1 210 Å2 ( 27% ) ( Fig . S1 ) . Although changes in surface accessibility are observed for residues in additional CDRs upon binding , no residue in CDR L3 , H1 and H2 makes direct contacts with Dom-II as judged by an inter-atom distance cutoff >3 . 5 Å , and CDR L2 contributes only a single hydrogen bond to the ( Fig . 3 , 4A and C ) . In antibody-antigen structures CDR H3 donates almost always the most extensive surface area contribution to the antigen-combining site , however , a scenario where antigen-binding is achieved almost exclusively by only two out of six CDRs remains unusual [37] . Antigen recognition in SM5-1 is reminiscent of the recognition of the class I fusion protein haemagglutinin from influenza A by the C05-Fab fragment [38] . Here also , a 24-residue-long H3 loop extends from the surface of the antibody and , in combination with H1 , suffices for antigen recognition . In case of SM5-1 , antigen recognition also includes a contribution from framework residue 3A , namely Lys67 of the light chain ( Fig . 3 ) . The residues in Dom-II that are recognized by SM5-1 cluster preferentially within two contiguous peptide segments , namely 359-EAED-362 and 379-KQEVN-383 ( Fig . 3 and 4 ) . Whereas 359-EAED-362 interacts mainly with light chain CDR L1 , 379-KQEVN-383 participates in a broad and extended hydrogen-bonding network with CDR L1 and preferentially with CDR H3 ( Fig . 4A–D ) . SM5-1 recognizes in addition a highly hydrophobic pocket on the surface of Dom-II . This pocket is located at the same position where we observed residual positive difference electron density in the structure of unbound Dom-II ( Fig . 2E ) . The CDR H3 segment 105-SNSGLSLL-112 binds across the opening of this pocket , and the side chain of Leu109 points into the pocket ( Fig . 4E ) . However , Leu109 appears not to be able to completely fill out this pocket ( Fig . S2A ) . The Leu109 contact is sealed-off from the surrounding solvent by a number of hydrogen bond interactions and these mainly involve the backbone atoms from the adjacent serine and leucine residues in CDR H3 . These hydrophilic contacts also include a bidental interaction between Dom-II residue Arg131 and CDR H3 residue Ser107 that appears to stabilize the conformation of CDR H3 ( Fig . 4D ) . Overall , the atomic architecture of the interaction patch that contains CDR H3 residue Leu109 resembles that typically observed for protein interaction hot spots [39] . Residue Tyr364 and the segment 378-KKQE-381 were identified as key determinants for binding and neutralizing activity of a series of SM5-1 related antibodies in a previous mutagenesis study [32] . The two residues Tyr364 and Lys379 were especially critical for SM5-1 binding ( YK-motif , [32] ) . In the structure of the complex , Lys379 adopts a central role , since , together with Glu359 , it is the only Dom-II residue that contacts both CDRs L1 and H3 ( Fig . 3 , Fig . S2B ) . In contrast , Tyr364 does not directly bind to SM5-1 . It lays in immediate spatial proximity to Dom-II residues Glu359 and Lys379 and shields the side chain of Lys379 from the surrounding water molecule . Its role might be to help orienting key Dom-II residues that directly interact with SM5-1 ( Fig . S2B ) . Overall , the structural data presented here are in good agreement with the previously reported mutational analysis [32] . SM5-1 belongs to a series of anti-HCMV gB mAbs that were isolated from an individual donor at a single time point [26] . All clones are related , i . e . are derived from one B-cell that was clonally diversified by somatic mutation and selection during germinal center reaction . Since all residues that interact with the antigen in the structure of the complex are largely conserved in these clones it seems reasonable to assume that the structural mode of Dom-II-recognition is conserved among these antibodies ( Fig . 3 ) . The binding affinities of the individual antibodies against HCMV gB differ by more than two orders of magnitude , and the highest affinity is observed for SM5-1 ( Table S1 ) [26] , [32] . The latter observation is consistent with the fact that SM5-1 is the most mature antibody of this family since it displays the largest number of mutations compared to the germline sequence ( Fig . 3 ) . When comparing the dissociation constants to the neutralization activities of the related mAbs , a clear correlation can be observed . However , the neutralization activities do not increase by the same extent as the binding affinities suggesting that affinity maturation as it occurs in the germinal center reaction does not translate directly into a better neutralization activity against HCMV . Interestingly , the off-rates of the antibodies ( i . e . the dissociation rate ) correlate best with their neutralization activities . In SM5-1 the CDRs L1 and H3 but also CDRs L3 , H1 and H2 clearly underwent affinity maturation from the germline sequence ( Fig . 3 ) . However in the crystal structure the latter three CDRs do not interact with Dom-II . This raises the possibility that these CDRs contribute to the enhanced neutralization activity of SM5-1 and interact with parts of gB or even additional viral components that are not present in recombinant Dom-II . To test this hypothesis SM5-1germ was generated , a SM5-1 derivative in which the CDRs H1 , H2 and L3 were reverted to the respective germline sequence ( Fig . S3 ) . In neutralization assays we could not detect any difference between SM5-1 and SM5-1germ , indicating that affinity maturation within those CDRs which are not in contact with recombinant Dom-II has no measurable impact on the neutralization activity of SM5-1 ( Fig . 5 ) . These data further strengthen the notion that the neutralization capacity of SM5-1 rests entirely on the interaction mode between DomII and SM5-1 that has been mapped in the crystal . The data above also suggests that affinity-maturated residues of SM5-1 , which do not contact Dom-II , might rather play a role for stabilizing the structure of the antibody itself . Inspection of the SM5-1 structure reveals that indeed several of the respective residues form tight intramolecular interactions including Lys30 , Asp31 , and His32 of CDR H1 . As an example , His32 from CDR H1 contacts Asp99 located at the N-terminus of the long CDR H3 . This region is further stabilized by additional polar interactions of residues His115 , Asn116 , Arg117 of CDR H3 . The latter residues have emerged during affinity maturation of SM5-1 and also do not form direct interactions with the antigen . To investigate the role of these polar residues in more detail , molecular dynamics ( MD ) simulations were performed for SM5-1 and for a 6-fold in silico substituted variant SM5-1* , in which the respective residues of CDR H1 and H3 were replaced to match the sequence of the less mature SM1-6 that binds Dom-II with lower affinity ( heavy chain K30T , D31G , H32T , H115Y , N116D and R117V ) . Comparison of the dynamics of SM5-1 and SM5-1* reveals that the substitutions mainly enhance the flexibility of the long CDR H3 loop , and this is most pronounced for residues 104–113 ( Fig . 6 ) . The overlay of the structures collected over the simulation time also showed a more prominent deviation of CDR H3 of SM5-1* from the Dom-II-bound conformation , which served as starting structure for the simulation ( Fig . 6B , C ) . The higher flexibility of CDR H3 in SM5-1* can be attributed to the loss of important stabilizing side chain interactions in the anchor region of CDR H3 . For example , in SM5-1 Asp99 is fixed by multiple H-bonds with the side chain of His32 , Tyr113 and His115 as well as with the backbone of Tyr33 and Asn116 ( Fig . S4 ) . Due to the replacement of His32 and His115 with tyrosine , this compact network of interactions involving the N- and C-terminus of CDR H3 is lost leading to a destabilization of the long connecting loop . The same effect is also detected in a control MD simulation starting from the structure of unbound SM5-1 ( Fig . S5 ) . Thus , it appears that the specific role of a number of affinity matured residues in SM5-1 is not to directly interact with Dom-II but to stabilize the binding-competent conformation of CDR H3 .
Analysis of the structure of SM5-1 in complex with gB Dom-II revealed that a significant portion of the mutated residues do not directly interact with Dom-II . MD simulations provide a functional explanation for this puzzling finding and suggest that these residues play a role in the stabilization of the SM5-1 CDR H3 loop itself . Intramolecular CDR H3 stabilization of antibodies has been reported previously and can be achieved by the formation of hydrogen-bonded β-hairpins , disulfide bridges , or the enrichment of prolines [40] , [41] . All these structural features are expected to enhance rigidity , which will reduce the entropy changes upon antigen recognition and , therefore , result in higher binding affinity . Long CDR H3s frequently exhibit extended β-hairpin structures as exemplified by the HIV-1 gp120-binding antibodies 10-1074 or PGT122 [42] , [43] . The CDR H3 of SM5-1 adopts an extended conformation , but does not display a regular β-hairpin structure . Consequently , the ability to gain intramolecular stabilization by backbone hydrogen bonds is rather limited; however , this is compensated by an additional stabilization through side chain hydrogen bonds . A similar combination of backbone and side chain hydrogen bonds is also present in the long CDR H3 of the haemagglutinin-binding human mAB C05 , which resembles a hammerhead topology [38] . As is the case for SM5-1 , the CDR H3 conformation of C05 does not significantly change upon antigen binding suggesting that intramolecular stabilization of long CDR H3 loops favors antigen binding by reducing the entropic loss caused by the interaction [38] . In contrast to backbone hydrogen bonds , which can be formed by any amino acid , the formation of side chain hydrogen bonds requires the presence of distinct amino acids with functional side chain groups . Thus , the emergence of polar residues during antibody maturation may reflect an enhanced intramolecular stabilization and does not necessarily imply that the respective residue forms interactions with the antigen . Consequently , this observation has also implications for other areas of research: Firstly , it underscores the difficulty of predicting correct antibody-antigen structures by docking approaches . These approaches frequently consider all residues , which have emerged during affinity maturation , as part of the interface . Such an approach will consequently fail in cases like SM5-1 by producing wrong complex geometries . Secondly , the structural principles deduced from SM5-1 may also affect the choice of residues to be considered in antibody optimization procedures ( e . g . phage display ) . As for any viral fusion protein , HCMV gB is assumed to adopt two different conformational states , namely a pre- and postfusion conformation . Moreover , the transition from the pre- to the postfusion state is expected to facilitate the fusion of the viral envelope with the host cell membrane . However , in case of gB proteins , the exact extent of the conformational rearrangement during the transition is far from clear . Currently , crystal structures of gB proteins are only available from HSV-1 and EBV [17] , [18] . These crystal structures are considered to display the postfusion conformation . Only in case of a more distantly related class III fusion protein , namely protein G from VSV , both the prefusion and postfusion conformations have been experimentally characterized [19] , [20] , [27] . In order to visualize , how SM5-1 could possibly recognize entire HCMV gB , we considered two gB models , namely one derived from the crystal structure of HSV-1 gB ( postfusion state ) and one modelled according to the prefusion conformation of VSV-G ( [26] and data not shown ) ( Fig . 7 ) . In the model of HCMV gB derived from the crystal structure of HSV-1 gB , the antibody-binding epitope is fully accessible from the solvent [26] . The binding mode of SM5-1 to Dom-II can be readily transferred onto entire gB without notable steric hindrances ( Fig . 7D , E and F ) . This also holds true if Dom-II is extended by an additional α-helix that became visible in a recent crystal structure of HSV-1 gB and that was previously absent either because of high flexibility and/or proteolytic cleavage [18] , [31] . It is not clear whether this helix would actually form in HCMV since HCMV gB is physiologically cleaved in vivo after this segment ( see above ) . The corresponding segment also lacks multiple hydrophobic residues that could anchor this segment to the Dom-II core structure . However , if such a helix is added to the gB-bound SM5-1 model then additional antibody antigen interactions possibly arise , namely between CDR L3 of SM5-1 and gB ( Fig . S6 ) . By contrast , L3 does not participate in Dom-II binding in the present crystal structure . Overall , the antigen-binding mode observed in SM5-1 in complex with Dom-II is fully compatible with an HCMV gB structure modelled upon the putative postfusion conformation observed in HSV-1 gB . This binding mode is further supported by recent electron microscopy studies that showed that addition of a HSV-2 Dom-II specific monoclonal antibody C226 to HSV-2 gB yields electron microscopy images that fully agree with the postfusion model of HCMV gB in complex with SM5-1 presented here [44] . Furthermore the dissociation constants of SM5-1 against isolated recombinant Dom-II and entire HCMV gB are very similar , arguing strongly against the fact that additional epitopes present on gB are missing in the Dom-II preparation [26] . This agreement also holds true if one includes potential glycosylation sites of HCMV gB in these considerations . As a mature protein , gB consists of a proteolytically processed two-chain disulfide-linked protein ( residues 1–459 and 460–906 ) , and in SDS-PAGE experiments the N-terminal part represents a diffusely migrating band indicating significant modification by glycosylation [33] , [34] . Computer-based analysis for potential N-linked glycosylation sites reveals 8 sites in the N-terminal part , while the C-terminal part contains no predicted site , which is consistent with its migration as a defined band in SDS-PAGE . Adding complex N-linked glycans to the gB model still allows for unrestricted access of the antibody-binding epitope of Dom-II by SM5-1 ( Fig . S7 ) . An HCMV gB model derived from the crystal structure of VSV-G is less reliable than an HSV-1 gB-derived model because of significant lower sequence identity between HCMV gB and VSV-G . Nevertheless , the overall structural similarity between the Dom-II segments of VSV-G and HCMV gB can be readily identified even if the respective domains superimpose only poorly ( see above ) . When modelling HCMV gB based on the prefusion conformation of VSV-G , the Dom-II antibody-binding epitope remains accessible , and the crystal structure of Dom-II in complex with SM5-1 can again be superimposed on the VSV-G-derived HCMV gB model without any severe clashes ( Fig . 7A , B and C ) . Thus , from these models it can be concluded that SM5-1 should be able to recognize entire gB irrespectively of whether gB adopts a conformation similar to the postfusion conformation observed in HSV-1 gB or the prefusion conformation of VSV-G . SM5-1 and related antibodies were isolated using recombinant gB , and it has been suggested that during eukaryotic production the truncated ectodomain of gB preferentially adopts the postfusion conformation which is also remarkably stable [45] , [46] . Nevertheless , the fact that SM5-1 can bind to extracellular virions as presented in conventional neutralization assays shows that SM5-1 binds also to the prefusion conformation of gB in vivo in agreement with the above described gB-binding models . SM5-1 neutralizes HCMV at a post-attachment state [26] . Therefore it can be excluded that SM5-1 merely blocks the attachment of the virus to the host cell . Also , the SM5-1 Fab fragment shows similar virus-neutralization capacities as complete IgG , indicating that crosslinking of gB on the viral surface is not required for neutralization ( N . Spindler , unpublished results ) . It remains unclear , which step of the viral entry pathway is blocked upon SM5-1 binding . One possibility is , that SM5-1 prevents formation of the active fusion complex by impeding the interaction of gB with gH/gL . Congruently to our previous publications on HCMV gB , Dom-II was identified as crucial target for neutralizing antibodies in HSV gB [28] , [32] . Atanasiu et al . reported for HSV-1 that anti-Dom-II antibodies were able to inhibit cell-cell fusion and block the interaction of gB with the gH/gL complex [28] . Occupation of the gH/gL-binding site in gB by SM5-1 may thus interfere with gB-gH/gL interaction . Whether a similar mode of action results in neutralization of free virions remains to be shown . Of note , in the case of EBV gB domains III , IV , and V are postulated as gH/gL interaction domains [47] . As an alternative to blocking the gH/gL-binding site , SM5-1 could also neutralize HCMV by inhibiting the transition from the prefusion to the postfusion state of gB and thereby impede membrane fusion . Although the affinity of SM5-1 for the prefusion state of gB is currently not known , circumstantial evidence as well as the above derived models suggest that SM5-1 is able to form thermodynamically stable complexes with both the prefusion and postfusion conformation of gB . However , this does not rule out that binding of SM5-1 to gB is able to kinetically inhibit the conformational transition . In the absence of data on how the fusion complex is activated on the virus envelope , the neutralizing mechanism of SM5-1 remains speculative and must await further studies . Vaccination using recombinantly produced gB incompletely protects against HCMV infection/disease [24] , [25] . The reasons for this are unknown . While the vaccine induced gB-binding and virus neutralizing antibodies comparable to natural infection when sera were analyzed in conventional fibroblast-based assays , the specific neutralizing activities were on average 15-fold lower than those observed following natural infection when epithelial cells were used as target cell type [48] . Also , the specificity of antibodies with respect to recognition of individual antigenic domains on gB has been found to differ between naturally HCMV infected and gB-vaccinated individuals [49] . This might again be directly linked to the fact that gB exists in a pre- and postfusion conformation which may differ considerably . Vaccination with gB might give rise to antibodies that recognize the postfusion state with higher affinity than the prefusion state , whereas a reversed antibody-binding preference could arise during a natural HCMV infection . Current technology allows for the comparative analysis of the individual antibody repertoire following infection and vaccination and this may provide valuable information for the future optimization of gB-based vaccines . Obviously , further investigations are needed to determine how exactly SM5-1 blocks the viral entry mechanism . Clearly , the crystal structure presented here will be valuable for further exploring the prefusion and postfusion state of HCMV . However , more importantly , the structural insights provided here open up a window of opportunity for the structure-based design of HCMV vaccines .
For the expression of GST fusion proteins of engineered variants of Dom-II in which residues 112 to 132 and 344 to 438 of the HCMV gB protein were linked by a 5-residue long artificial linker sequence , the respective coding sequences were cloned into pGEX-6P-1 plasmids ( GE Healthcare ) as described and DH10B bacteria were transformed with the expression vectors [32] . The fusion proteins were purified from Escherichia coli lysates following incubation with glutathione Sepharose 4B ( GE Healthcare ) for 2 h at room temperature . Unbound bacterial protein was removed by washing with PBS , and proteins were eluted with 10 mM reduced L-glutathione . To remove the GST part , the bound Dom-II fusion protein was proteolytically cleaved using PreScission protease ( GE Healthcare ) by incubation for 5 h at 4°C and the Dom-II part was eluted . For further purification anion-exchange chromatography in 20 mM Tris , pH 8 . 0 was performed using a HiTrapQ FF 1-ml column ( GE Healthcare ) on an ÄKTApurifier ( GE Healthcare ) followed by size exclusion chromatography in 20 mM Tris , 150 mM NaCl , pH 7 . 4 on a HiLoad 16/60 Superdex75 column ( GE Healthcare ) . The Fab fragment was prepared by papain digestion of the SM5-1 IgG molecule using a Fab preparation Kit from Thermo Scientific Pierce ( Germany ) . Papain digestion was carried out according to the manufacturer's instructions . The Fc-part and undigested IgG molecules were removed by affinity purification using a HiTrap Protein A HP 1-ml column ( GE Healthcare ) . Size exclusion chromatography in 20 mM Tris , 150 mM NaCl , pH 7 . 4 was used as final purification step of the Fab fragment using a HiLoad 16/60 Superdex200 column . For complex formation , purified SM5-1 Fab and Dom-II were co-incubated at a ratio of 1∶2 ( w/w ) followed by size exclusion chromatography in 20 mM Tris , 150 mM NaCl , pH 7 . 4 using a HiLoad 16/60 Superdex200 column to remove non-complexed proteins . Engineered Dom-II was crystallized using the hanging-drop vapor-diffusion method and equilibrating a 1–4 µl mixture containing protein solution ( 11 . 9 mg/ml Dom-II in 20 mM Tris-HCl , pH 7 . 5 ) and reservoir solution ( 1 . 5 M Li2SO4 , 0 . 1 M Na-HEPES salt , pH 7 . 5 ) at ratios 1∶2 . 5 or 1∶3 against 700 µl of reservoir solution . Crystallization of the SM5-1 Fab fragment was achieved with the sitting-drop vapor-diffusion set-up . 200 nl of protein solution ( 9 . 7 mg/ml in 20 mM Tris-HCl , pH 7 . 5 ) were mixed with 200 nl of reservoir solution ( 4 M sodium formate ) and equilibrated against reservoir solution . Crystals of individual Dom-II and SM5-1 Fab grew at 19°C within two days . In both cases , crystals were flash-cooled in liquid nitrogen without addition of cryoprotectants . Dom-II of gB in complex with SM5-1 was crystallized with the sitting-drop vapor-diffusion method using protein concentrations between 10 and 15 mg/ml in 20 mM Tris-HCl , pH 7 . 5 and mixing 200 nl of protein solution with 200 nl of reservoir solutions . To increase the chances for obtaining well-diffracting crystals of the complex , we crystallized SM5-1 in complex with several different mutants of Dom-II that bind SM5-1 with similar affinities than wild-type Dom-II [32] . However , most complexes only produced crystals that were intergrown and unsuitable for X-ray diffraction experiments . In case of the complex formed between double mutant gB Dom-II-I397A/N398A and SM5-1 Fab , single crystals could be isolated after equilibrating the protein-reservoir mixture against 70 µl reservoir solution ( 0 . 1 M HEPES , pH 7 . 5 , 0 . 2 M L-proline , and 24% PEG 1500 ) for one month at 19°C . Crystals of the complex were briefly soaked with 20% ethylene glycol before being transferred into liquid nitrogen . In the crystal structure , and as expected , the residues at positions 397 and 398 in Dom-II do not participate in the interaction with SM5-1 Fab ( Fig . 3 ) . All diffraction data sets were collected using 0 . 5 to 0 . 6° rotation frames at PX beamline BL 14 . 1 at Hemholtz Zentrum Berlin BESSY synchrotron facility and processed with program XDS [50] . All three crystal structures could be solved by molecular replacement using program PHASER within the CCP4 program suite [51] , [52] . The structure of Dom-II comprising residues 118 to 132 and 344 to 438 of the HCMV gB protein could be solved with a mixed search model that was based on the corresponding Dom-II from HSV-1 glycoprotein B ( residues 142 to 152 and 364 to 459 , PDB ID 2gum ) . The mixed model was generated using the Joint Center of Structural Genomics SCRWL server [53] , [54] . The unrelated Fab fragment of the human anti-HIV-1 gp120 reactive antibody E51 ( PDB ID 1rzf ) was used as a search model for the SM5-1 Fab structure . Flexible parts , such as CDR loops were deleted and both variable domains and constant domains were considered separately during the molecular replacement calculations . Ensemble 1: residues 109 to 210 ( CL ) and 114 to 214 ( CH1 ) ; ensemble 2: residues 2–23 , 34–54 , 63–91 , 96–106 ( VL ) and 2–23 , 33–51 , 57–95 , 103–111 ( VH ) . The gB Dom-II-SM5-1 Fab complex structure was solved using the crystal structures of the previously solved individual proteins as search models . Here also , any flexible regions in SM5-1 were removed and two different ensembles consisting of either the variable or constant domains used during the molecular replacement search of the SM5-1 fragment . All structures were refined with the CCP4 program REFMAC5 until convergence and until no further details could be interpreted in the electron density maps [52] , [55] . During the final refinement rounds a TLS refinement step was performed in order to improve the fit between the models and the experimental data . The elbow angle of the SM5-1 Fab fragment was calculated with the AS2TS web server using the following immunoglobulin domain boundaries: VL: residues 1 to 110; VH: 1 to 131 , CL: 111 to 212 and CH1: 132 to 232 [36] , [56] . Changes in the accessible surface areas were calculated with program AREAIMOL from the CCP4 program suite using the default solvent probe radius of 1 . 4 Å [52] . All molecular illustrations were generated with program PyMol ( http://www . pymol . org/citing ) . Potential N-linked glycosylation sites were predicted using the NetNGlyc server [57] . The coding sequences of SM5-1 and a partially germline-reverted mutant derivative ( SM5-1gern ) were constructed according to Fig . S3 . The respective nucleotide sequences coding for the heavy and light chain variable regions were chemically synthesized by Life Technologies ( Germany ) and cloned into eukaryotic IgG expression vectors [58] . To produce IgG , 293T cells were co-transfected with respective IgG heavy and light chain expression vector plasmids . The cell culture supernatant was harvested 8 days post transfection and IgG was purified using a HiTrap Protein A HP 1-ml column ( GE Healthcare ) on an ÄKTAprime chromatography system ( GE Healthcare ) . Virus infection of fibroblasts and neutralization was exactly as described in Pötzsch et al . 2011 [26] . Molecular dynamics ( MD ) simulations were performed for SM5-1 and for a 6-fold in silico variant ( SM5-1* ) , in which the respective residues were exchanged to match the sequence of the less mature SM1-6 ( K30T , D31G , H32Y , H115Y , N116D and R117V ) . For both SM5-1 and SM5-1* two independent simulations were performed starting either from the conformation present in complex with Dom-II or from the crystal structure of unbound SM5-1 . This resulted in a total of four simulations . To reduce computational cost , only the VH-VL domain ( aa 1–131 and 1–110 of chain H and L , respectively ) of the Fab-fragment were included in the simulation . The protonation state of titratable amino acids was calculated by PROPKA [59] and the Swiss PDB Viewer was applied to add unresolved residues and side chains , as well as for the amino acid exchanges [60] . All MD simulations were performed using the AMBER 11 program package and the ff99SB force field parameters [61] , [62] . Proteins were placed in periodic truncated octahedral water boxes with at least 15 Å of solvent between any atom of the solute and the periodic box edges . The systems were neutralized by adding the respective number of Cl− ions . Initially , all four systems were minimized in a three-step procedure and then gradually heated to 310 K following a previously established protocol [63] , [64] . Subsequently , 100 ns MD simulations were performed with a time step of 2 fs and periodic boundaries . The Settle-algorithm [65] was applied to constrain covalent bonds involving hydrogens . The obtained data of the simulation was analyzed and visualized using AMBER Tools and VMD [66] . Coordinates and structure factors of Dom-II , SM5-1 Fab , and the Dom-II-SM5-1 Fab complex have been deposited with the Protein Data Bank ( PDB ) under accession codes 4OSN , 4OSU , and 4OT1 , respectively .
|
Human cytomegalovirus ( HCMV ) belongs to the family of β-herpes viruses . HCMV infections are not only life threatening to people with a compromised immune system but also the most common viral cause of congenital defects in newborns . Hence , the development of HCMV vaccines was ranked top priority by the US Institute of Medicine in 1999 . Virtually all infected individuals develop antibodies against the envelope protein gB , which plays a crucial role in the infection process . Here , we describe the crystal structure of a fragment of the virus neutralizing antibody SM5-1 in complex with an antigenic determinant of gB , namely Dom-II . The structure shows that antigen antibody interactions are concentrated within two CDRs of SM5-1 . Computational methods and an analysis of additional antibody sequences from the same lineage reveal that additional key contributions to high affinity binding are provided by residues that stiffen the extra-long CDR H3 loop without directly contacting the antigen . We suggest that the optimization of such indirect contributions represents a common and yet undervalued principle of the antibody maturation process . Furthermore our data suggest that the neutralizing effect of SM5-1 either originates from blocking membrane fusion or from preventing interaction of gB with other envelope proteins .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases",
"structural",
"characterization",
"medicine",
"and",
"health",
"sciences",
"cell",
"biology",
"macromolecular",
"crystallography",
"genetics",
"x-ray",
"crystallography",
"biology",
"and",
"life",
"sciences",
"immunology",
"congenital",
"disorders",
"molecular",
"biology",
"crystallographic",
"techniques",
"research",
"and",
"analysis",
"methods"
] |
2014
|
Structural Basis for the Recognition of Human Cytomegalovirus Glycoprotein B by a Neutralizing Human Antibody
|
Selective IgA deficiency ( IgAD; serum IgA<0 . 07 g/l ) is the most common form of human primary immune deficiency , affecting approximately 1∶600 individuals in populations of Northern European ancestry . The polygenic nature of IgAD is underscored by the recent identification of several new risk genes in a genome-wide association study . Among the characterized susceptibility loci , the association with specific HLA haplotypes represents the major genetic risk factor for IgAD . Despite the robust association , the nature and location of the causal variants in the HLA region remains unknown . To better characterize the association signal in this region , we performed a high-density SNP mapping of the HLA locus and imputed the genotypes of common HLA-B , -DRB1 , and -DQB1 alleles in a combined sample of 772 IgAD patients and 1 , 976 matched controls from 3 independent European populations . We confirmed the complex nature of the association with the HLA locus , which is the result of multiple effects spanning the entire HLA region . The primary association signal mapped to the HLA-DQB1*02 allele in the HLA Class II region ( combined P = 7 . 69×10−57; OR = 2 . 80 ) resulting from the combined independent effects of the HLA-B*0801-DRB1*0301-DQB1*02 and -DRB1*0701-DQB1*02 haplotypes , while additional secondary signals were associated with the DRB1*0102 ( combined P = 5 . 86×10−17; OR = 4 . 28 ) and the DRB1*1501 ( combined P = 2 . 24×10−35; OR = 0 . 13 ) alleles . Despite the strong population-specific frequencies of HLA alleles , we found a remarkable conservation of these effects regardless of the ethnic background , which supports the use of large multi-ethnic populations to characterize shared genetic association signals in the HLA region . We also provide evidence for the location of association signals within the specific extended haplotypes , which will guide future sequencing studies aimed at characterizing the precise functional variants contributing to disease pathogenesis .
The major histocompatibility complex ( MHC ) locus has been one of the most intensively studied regions in the vertebrate genome since it was first discovered in the mouse in 1936 [1] . In humans , gene products from the MHC were initially identified as surface markers on leucocytes , which led to its alternative designation as the human leukocyte antigen ( HLA ) complex . This genomic region spans approximately 3 . 6 megabase pairs ( Mb ) of genomic sequence and encodes over 200 genes , many of which with a defined immune function [2] . More recently , the definition of an extended MHC region ( xMHC ) has been proposed , which encompasses approximately 7 . 6 Mb and 421 annotated gene loci [3] . The HLA locus contains the canonical HLA Class I and Class II gene clusters . Class I HLA molecules are expressed on the surface of most human cells , while constitutive expression of Class II molecules is restricted to antigen presenting cells , including dendritic cells , macrophages and B cells . These molecules function to present peptide antigens to T cells to initiate adaptive immune responses . The relevance of this locus to the pathogenesis of common human diseases is clearly evidenced by the reported association of polymorphisms in the HLA region with over one hundred diseases , particularly autoimmune and inflammatory conditions [4] . Similarly , the genetic association with HLA markers has been well documented in selective IgA deficiency ( IgAD; serum IgA concentration <0 . 07 g/l ) , the most common form of human primary immunodeficiency [5] , [6] . However , despite the robust association with a large number of immune-mediated conditions , the conservation of some ancestral haplotypes [7] , [8] , which can encompass large genomic segments containing many genes , has greatly impaired our ability to unambiguously identify the gene ( s ) contributing to disease susceptibility . The HLA was first described as a risk locus for IgAD through the association with HLA Class I and Class II markers [9]–[11] and ascribed to specific conserved haplotypes . Most notably , the extended HLA-A*01-B*08-DRB1*0301-DQB1*02 ( DR3 ) haplotype has been identified as the single strongest genetic risk factor for IgAD in Northern European populations [12] . A striking 13% of DR3 homozygotes have been estimated to be IgA deficient [13] , although this figure might be inflated due to publication bias [14] . An even larger proportion ( 67% ) appears to suffer from at least one form of Ig deficiency , including selective IgG3 , IgG4 , IgD and IgA deficiency [15] . Conversely , the HLA-DRB1*1501-DQB1*06 ( DR2 ) haplotype , has been shown to confer strong protection against IgAD , with homozygous individuals showing a virtual complete protection from the disease [12] . Positive associations have also been described with 2 other extended haplotypes , namely the HLA-B*14-DRB1*0102-DQB1*05 ( DR1 ) and the HLA-B*44-DRB1*0701-DQB1*02 ( DR7 ) haplotypes [12] . Interestingly , the risk conferred by DR7 and particularly DR1 haplotypes has been shown to be greater than that conferred by DR3 haplotypes in populations of Southern European ancestry [16]–[18] . Despite the strong association at the HLA locus , there has been no consensus as to the precise location of the causal variants within these haplotypes , with some groups suggesting its placement to the telomeric end of the HLA Class III region [19]–[21] and others in the HLA Class II region [22] , [23] . To refine the complex association signals in the extended MHC , we genotyped a panel of 1 , 686 SNPs and imputed with high confidence the genotype of 9 , 905 SNPs spanning the entire HLA locus . We used the genotype information to reconstruct the individual long-range haplotypes and impute common HLA alleles . Here , we report single and multi-marker association analyses of both HLA and non-HLA variants with IgAD in a combined sample of 772 IgAD cases and 1976 matched controls from 3 independent European populations . These data provide the most complete fine-mapping effort of the HLA locus in IgAD to date .
We recently performed a genome-wide association study of IgAD and reported the disease-association of 2 novel non-HLA loci: IFIH1 and CLEC16A [24] . We also confirmed the HLA locus as the strongest genetic risk factor for IgAD [24] . To better characterize the association at the HLA locus , we used the genotyping data from 1 , 686 SNPs located within a 10 Mb region of chromosome 6 ( 25–35 Mb ) containing the extended HLA region , and employed a long-range haplotype phasing approach to infer , with high confidence , individual HLA haplotypes . To further refine the fine mapping of this region , we also imputed the genotypes of an additional 9 , 905 SNPs using the HapMap2 reference set . Allele-based association tests were then performed on the 1 , 686 genotyped and 9 , 905 imputed SNPs and on the imputed HLA alleles that survived stringent quality control steps in 430 IgAD cases and 1090 controls from Sweden and Iceland and in two independent replication cohorts from Spain ( 256 cases and 322 controls ) and Finland ( 86 cases and 564 controls ) . To avoid spurious associations due to specific population differences that are associated with the HLA locus , we used all available genome-wide genotyping data to describe the population substructure using a principal component method [25] , [26] . Exclusion of the genetic outliers in each population ensured the genetic homogeneity of our sample , as evidenced by the modest distributional inflation ( λGC ) observed in each case-control series ( Sweden/Iceland: λGC = 1 . 05; Spain: λGC = 1 . 05; Finland: λGC = 1 . 04 ) . Allele-based association tests were performed in each population independently , and combined P values were calculated using a meta-analysis . In addition , we also phased the genotypes from a set of 49 selected SNPs spanning the classical MHC region to reconstruct the extended haplotypes in each individual . SNP selection was optimized to combine a set of 11 SNPs that best captured the information of the known HLA alleles that were typed directly in our sample , together with a second set of 38 common SNPs , which had low pairwise LD and were distributed uniformly across the classical HLA region . Importantly , this strategy ensured that we recapitulated the full diversity of the haplotype structure of the HLA locus , and allowed us to map the individual recombination events within the extended haplotypes ( see Methods ) . We tested the association of the more common extended haplotypes and used the haplotype background information to map the recombination events more accurately and the location of the putative causal variants within each specific disease-associated haplotype . The primary association peak in the HLA locus mapped to the Class II region ( Figure 1 ) . Consistent with a recent report , that included approximately 270 overlapping IgAD patients and 670 controls from the Swedish cohort used in this study [27] , the strongest association was observed with the imputed HLA-DQB1*02 allele ( combined P = 7 . 69×10−57; OR = 2 . 80 , 95% CI 2 . 46–3 . 20; Figure 1 ) . The association with the DQB1*02 allele was approximately 6 orders of magnitude more significant than the most associated imputed SNP , which mapped to the HLA-DRB1 gene ( rs3891175; combined P = 4 . 31×10−51; OR = 2 . 82 , 95% CI 2 . 44-2 . 27 ) and over 1010-fold more significant than the most associated genotyped marker ( rs204999; combined P = 2 . 02×10−46; OR = 2 . 47 , 95% CI 2 . 18–2 . 81; Figure 1 ) . HLA-DQB1*02 is a common allele in populations of European ancestry , and is linked to both the HLA-DRB1*0301 and -DRB1*0701 alleles that have been previously described as strong IgAD risk factors . In this study , we confirmed the disease risk conferred by the DRB1*0301 ( combined P = 1 . 56×10−34; OR = 2 . 49 , 95% CI 2 . 14–2 . 90; Table 1 ) and DRB1*0701 alleles ( combined P = 8 . 68×10−17; OR = 2 . 03 , 95% CI 1 . 70–2 . 41; Table 1 ) . One important distinction between these 2 haplotypes is that the DRB1*0301 allele is typically linked with the DQB1*0201 allele , while the DRB1*0701 allele is linked to the DQB1*0202 variant [28] . These two DQB1*02 alleles differ only by one amino acid in the third exon , and are therefore difficult to differentiate by traditional typing methods using 2-digit resolution . Despite the sequence similarity , the HLA Class II sequences surrounding these 2 alleles are distinct and contain different HLA-DQA1 alleles . To discriminate between the two HLA-DQB1*02 alleles , we genotyped the HLA-DQB1 gene in 15 Finnish IgAD patients carrying the DQB1*02 allele ( 5 DRB1*0301/X , 5 DRB1*0701/X and 5 DRB1*0301/DRB1*0701 heterozygous individuals ) using 4-digit resolution . We confirmed that all DRB1*0301 alleles were always associated with the DQB1*0201 allele , while 9/10 ( 90% ) of the DRB1*0701 alleles were associated with the DQB1*0202 subtype ( data not shown ) . Analysis of the haplotype structure showed that , as noted previously [29] , DRB1*0301 haplotypes showed very low levels of historical recombination and were generally found as extended conserved haplotypes ( Figure 2 ) . In our sample , of the 907 haplotypes carrying the DRB1*0301 allele , 634 ( 69 . 9% ) encompassed the extended HLA-B*0801-DRB1*0301-DQB1*02 haplotype , and 273 ( 30 . 1% ) were recombinant ( non-B*0801 ) haplotypes . Importantly , recombinant DRB1*0301 haplotypes that lacked B*0801 showed no evidence of association with IgAD ( combined P = 0 . 42; OR = 1 . 10 , 95% CI 0 . 84–1 . 44; Table 1 ) . In contrast , the extended HLA-B*0801-DRB1*0301-DQB1*02 haplotype was found to be the most disease-associated HLA haplotype ( combined P = 3 . 37×10−43; OR = 3 . 33 , 95% CI 2 . 79–3 . 97; Table 1 ) . The distribution of DRB1*03 haplotypes varied across the populations studied , with a higher percentage of extended HLA-B*0801-DRB1*0301-DQB1*02 haplotypes in Northern Europeans , as compared with the Spanish population ( Table S1 ) , likely reflecting the increased frequency of HLA-B*18-DRB1*0301-DQB1*02 haplotypes in Southern Europeans . Nevertheless , in all 3 cohorts , we observed a consistent and significant increase in the frequency of extended B*0801-DRB1*0301-DQB1*02 compared to recombinant ( non-B*0801 ) DRB1*0301 haplotypes in IgAD cases ( combined P = 6 . 3×10−9; OR = 2 . 65 , 95% CI 1 . 90–3 . 71; Table S1 ) . Similarly , recombinant ( non-DRB1*0301 ) B*0801 haplotypes were also not associated with IgAD ( combined P = 0 . 228; OR = 0 . 92 , 95% CI 0 . 64–1 . 33; data not shown ) . Given the strong effect size of the risk allele present on the extended HLA-B*0801-DRB1*0301-DQB1*02 haplotype , we have 100% and 96% power to detect the association with the observed recombinant ( non-B*0801 ) DRB1*0301 and ( non-DRB1*0301 ) B*0801 haplotypes , respectively , at a significance level ( α ) of 5×10−5 , and even 100% and 50% power at a much more stringent genome-wide significant threshold of 5×10−8 ( Table S2 ) . We next performed a conditional logistic regression analysis , using the genotype of the extended HLA-B*0801-DRB1*0301-DQB1*02 as a covariate . The imputed HLA-DRB1*0701 allele was found to be the most significantly associated marker ( combined Pcond = 1 . 77×10−22; Figure 3 ) . In contrast to DRB1*0301-containing haplotypes , haplotypes linked to the DRB1*0701 allele have a high rate of historical recombination ( Figure 2 ) . Although , DRB1*0701 is found frequently within the HLA-B*13-DRB1*0701-DQB1*02 , B*44-DRB1*0701-DQB1*02 and B*57-DRB1*0701-DQB1*0303 extended haplotypes , the DRB1*0701 allele can also be found on other haplotypes , in association with different HLA Class I and Class II alleles ( Figure 2 ) . However , all DRB1*07 haplotypes carrying the DQB1*02 allele were found to be similarly associated with increased risk to IgAD ( combined P = 6 . 35×10−17; OR = 2 . 23 , 95% CI 1 . 83–2 . 72; Table 1 ) . Conversely , we found no strong evidence for the association of recombinant ( non-DQB1*02 ) DRB1*07 haplotypes with IgAD ( combined P = 0 . 027; OR = 1 . 34 , 95% CI 0 . 96–1 . 86; Table 1 ) . Further supporting this observation , we have over 80% power to detect the association of recombinant ( non-DQB1*02 ) DRB1*07 haplotypes at α = 5×10−8 ( Table S2 ) . These data point to an important role of the DQB1*0202 allele , or a region in tight linkage with DQB1*0202 , in the association of the DRB1*0701-DQB1*02 haplotypes with IgAD . In summary , the data suggest that DRB1*0301 and DRB*0701 haplotypes independently contribute to risk of IgAD , and the signal at DQB1*02 is essentially the sum of these independent effects . After further conditioning on both the HLA-B*0801-DRB1*03-DQB1*02 and –DRB1*0701-DQB1*02 haplotypes , the most significant residual HLA association signal corresponded to the HLA-DRB1*0102 allele ( Pcond = 2 . 95×10−19; Figure 3 ) . Association of the DRB1*01 allele with IgAD has been well documented , and contains both the more common DRB1*0101 and the less common DRB1*0102 subtypes . Importantly , we found that the association with IgAD was exclusively contributed by haplotypes containing the DRB1*0102 allele , which showed association with risk in all 3 independent European populations ( combined P = 5 . 86×10−17; OR = 4 . 28 , 95% CI 2 . 92–6 . 26; Table 1 ) . Conversely , we found essentially no evidence for the association of the HLA-DRB1*0101-DQB1*05 haplotype with IgAD ( combined P = 0 . 038; OR = 1 . 17 , 95% CI 0 . 98–1 . 39; Table 1 ) . Haplotype analysis showed that the DRB1*0102 is usually present on the extended B*1402-DRB1*0102-DQB1*0501 haplotype , which is a rare haplotype in Northern Europeans ( 0 . 2% allele frequency ) , but more common in the Spanish population ( 3% allele frequency; Table 1 ) . Nevertheless , the DRB1*0102 allele was a strong risk factor in all 3 cohorts , and contributed the strongest individual effect of all the variants in the HLA region , with a combined OR = 4 . 59 ( 95% CI 2 . 89–7 . 30 , Table 1 ) . The recombination rate was modest across the extended B*1402-DRB1*0102-DQB1*05 haplotype . Of the 140 DRB1*0102-containing haplotypes , 47 ( 33 . 6% ) were recombinant ( non-B*1402; Figure 2 ) . Despite the low number of recombinant haplotypes , the trend towards increased risk was maintained , even in the absence of the B*1402 allele ( combined P = 3 . 62×10−3; OR = 3 . 22 95% CI 1 . 70–6 . 13; Table 1 ) . In addition , in a single-locus association test , the HLA-DRB1*0102 allele was found to be more significantly associated with IgAD than the HLA-B*14 allele alone ( combined P = 1 . 07×10−12; OR = 2 . 68; 95% CI 1 . 99–3 . 61; data not shown ) , suggesting that the causal allele on the extended B*1402-DRB1*0102-DQB1*05 is likely to be closer to the HLA-DRB1 than to the HLA-B region . To characterize additional independent loci , we next conditioned on the HLA-B*0801-DRB1*03-DQB1*02 , -DRB1*0701-DQB1*02 and -DRB1*0102 alleles . The top residual association signal corresponded to the protective HLA-DRB1*15 allele ( Pcond = 3 . 26×10−18; Figure 3 ) . The protective effect of the HLA-DRB1*15 allele was evident in all 3 cohorts ( combined P = 2 . 24×10−35; OR = 0 . 13; 95% CI 0 . 09–0 . 19; Table 1 ) , and was observed for both DRB1*15 subtypes: the common DRB1*1501 allele , which was found to travel mostly on the extended B*0702-DRB1*1501-DQB1*06 haplotype , and the rare DRB1*1502 allele . DRB1*1501 haplotypes showed a relatively high rate of historical recombination , and 51 . 2% excluded B*0702 ( Figure 2 ) . However , the recombinant DRB1*15 haplotypes were equally associated with protection from IgAD ( combined P = 3 . 70×10−18; OR = 0 . 14 , 95% CI 0 . 09–0 . 24; Table 1 ) , suggesting that the causal variant associated with the DRB1*1501 signal is located within the linkage disequilibrium ( LD ) block containing HLA-DRB1 and HLA-DQB1 in the Class II region . We next performed further conditional logistic regression analysis , conditioning on all four association signals noted above: B*0801-DRB1*03-DQB1*02 , -DRB1*0701-DQB1*02 , DRB1*0102 and DRB1*1501 . The association signal was significantly reduced , with only a few additional SNPs reaching the genome-wide association significance threshold ( P<5×10−8; Figure S1 ) . These SNPs mapped specifically to two discrete genomic regions: ( i ) the BTNL2-HLA-DRA locus on the telomeric end of the HLA class II region ( rs743862; Pcond = 8 . 07×10−14; Figure S1 ) ; and ( ii ) the HLA-DQB1 region ( rs9275141; Pcond = 1 . 48×10−11; Figure S1 ) , supporting the contribution of these two additional regions to risk for IgAD .
Markers in the HLA region represent the strongest genetic risk factor associated with IgAD . Nevertheless , the extensive conservation of the disease-associated haplotypes has hindered our ability to confidently map the causal variants . In this study , we performed the largest fine-mapping effort to date of the HLA locus in IgAD , using a high-density SNP panel to characterize the complex association at this locus , and to map the location of the independent susceptibility loci . The primary association signal in the HLA region was found to be the HLA-DQB1*02 allele , including both the DQB1*0201 and the DQB1*0202 alleles , which are linked with the HLA-DRB1*03 and -DRB1*07 haplotypes respectively . Using a haplotype-based analysis , we found that the primary association signal was caused by the additive effect of two independent susceptibility alleles located on the extended HLA-B*08-DRB1*0301-DQB1*02 and DRB1*0701-DQB1*02 haplotypes . Importantly , by using a long-range haplotype phasing approach in a large sample size , we were able to demonstrate that recombinant ( non-B*0801 ) DRB1*0301 haplotypes carrying the DQB1*0201 allele showed no evidence of association with disease . Despite the low historical recombination rate on DRB1*0301 haplotypes , this approach allowed us to compare a sufficiently large number of recombinant haplotypes to increase our confidence in these observations . Interestingly , the frequency of extended DRB1*0301 haplotypes was consistently increased in IgAD patients from all 3 independent populations , suggesting that , despite obvious population specific differences in the distribution of these haplotypes , the underlying association signal is identical in the different populations . Similarly , recombinant ( non-DRB1*0301 ) B*0801 haplotypes , showed no evidence of association with IgAD , suggesting that the causal allele in the extended HLA-B*08-DRB1*0301-DQB1*02 is likely to be located in the telomeric end of the Class II region or in the Class III region . To characterize further independent susceptibility loci , we next performed a stepwise conditional logistic regression analysis . Following conditioning on the primary HLA-B*08-DRB1*0301-DQB1*02 association signal , the most significantly associated HLA allele was found to be the DRB1*0701 allele , further supporting the presence of an independent risk allele on DRB1*07 haplotypes . Despite the high recombination rate , we found that DRB1*07 haplotypes carrying different Class I allele were all similarly associated with IgAD . This was in contrast to non-DQB1*0202 recombinant haplotypes , which showed no evidence for association . The approach used in this study to compare the association of the recombinant haplotypes in a large cohort to the respective extended risk haplotypes was validated by power calculations ( summarized in Table S2 ) showing that the lack of association of the recombinant ( non-B*0801 ) DRB1*0301 , ( non-DRB1*0301 ) B*0801 and ( non-DQB1*02 ) DRB1*07 haplotypes are not likely to be negative results due to a lack of power in the study . Taken together , these data point to a causal role of the DQB1*0202 allele or another variant in tight LD with it in the HLA-DQA1–DQB1 locus in the Class II region . We therefore propose a model whereby the strong association observed with DQB1*02 allele results from the additive effect of two independent risk loci: the DQB1*0202 allele , or another marker in high LD and close proximity , traveling on HLA-DRB1*0701-DQB1*0202 haplotypes and an independent allele , most likely located within the border of the HLA Class III and Class II region , traveling specifically on the extended HLA-B*0801-DRB1*0301-DQB1*02 haplotype . Conditioning on the HLA-B*0801-DRB1*0301-DQB1*02 and -DRB1*0701-DQB1*02 association signals , we characterized two additional secondary association signals in the HLA region . The HLA-DRB1*0102 allele was found to be a robust independent risk allele , while DRB1*1501 was strongly protective . The most significant secondary association signal was the DRB1*0102 allele traveling on the extended HLA-B*1402-DRB1*0102-DQB1*05 haplotype . The prevalence of the B*1402-DRB1*0102 haplotype showed marked population differences , and was more frequent in populations of Southern European ancestry , as evidenced by the 4 . 5% allele frequency in the Spanish controls , compared to only 0 . 3% and 0 . 4% in the Swedish and Finnish controls respectively . To have a better estimate of the DRB1*0102 allele frequency in Northern Europeans , we performed an extensive survey of the Swedish volunteer bone marrow donor registry ( the Tobias registry ) . We identified 296 copies of the B*1402-DRB1*0102 haplotype in 23 , 610 healthy individuals that were surveyed , corresponding to a 0 . 6% allele frequency in the Swedish population . Despite the specific population differences , the DRB1*0102 association was consistent in all 3 independent cohorts and showed the highest OR of any HLA allele . Interestingly , the prevalence of IgAD in the Iranian population has been shown to be similar to that observed in European populations [30] . However , in Iranians the strongest HLA association with IgAD is linked to the B*14 and DRB1*01 alleles [31] . Taken together , these data clearly support the presence of a strong risk variant on the conserved HLA-B*1402-DRB1*0102-DQB1*05 haplotype . In fact , the effect size observed with this allele is consistent with the presence of a rare variant with strong penetrance on the HLA-B*1402-DRB1*0102-DQB1*05 background . Importantly , the lack of association with the more common DRB1*0101 subtype , which shares the association with the DQB1*05 allele , and the strong association with recombinant DRB1*0102 haplotypes , supports the location of the causal variant telomeric to the Class II region , most likely within the HLA Class III region . It is interesting to note that the 2 missense mutations that we have recently characterized in MSH5 , L85F and P786S , located in the HLA Class III region , travel specifically on this extended haplotype [32] . It should be noted , however , that this hypothesis was not supported by a recent study by Pozo et al , suggesting that recombinant ( non-DRB1*0102 ) haplotypes carrying the L85F variant did not show evidence of association with IgAD [33] . Nevertheless , the conservation and low frequency of recombinant HLA-B*1402-DRB1*0102-DQB1*05 haplotypes in European populations , warrant further analyses and a larger sample size to confirm whether these missense mutations are the causal variants for IgAD in this haplotype . Another independent association signal identified was the protective effect conferred by the DRB1*1501 allele . The association of DRB1*1501 with disease protection has been well documented , and further supports the important role of variants in the HLA Class II region to the complex association signal observed on the HLA locus . In fact , although the recombination rate in the DRB1*1501 haplotypes was elevated , all haplotypes sharing the HLA Class II fragment containing the DRB1*1501 allele conferred similar protection from IgAD . These data support the location of a protective allele mapping specifically to the LD block containing the HLA-DRB1 and HLA-DQB1 genes , and are consistent with the previous hypothesis that the protective effect is due to the presence of a negatively charged aspartic acid in position 57 of the HLA-DQβ chain in DRB1*1501 haplotypes . In contrast , in risk haplotypes , the aspartic acid residue is replaced by a neutral alanine [12] . Of interest , the association with the HLA locus in IgAD shares some striking similarities with the association in type 1 diabetes ( T1D ) , where the DRB1*0301 allele is a strong risk factor and the DRB1*1501 allele confers protection against the disease [34] , [35] . Similarly we have recently characterized the association between IgAD and two novel non-HLA loci , IFIH1 and CLEC16A [24] , which are also known to be associated with T1D [36] , suggesting a shared genetic predisposition to both diseases . Despite these similarities , there are some notable disease-specific differences in the association with the HLA between the two conditions: i ) the DRB1*04 allele , one of the strongest risk factor in T1D , is not associated with IgAD; ii ) unlike T1D , the DRB1*0301 association with IgAD is restricted to the extended B*0801-DRB1*0301-DQB1*02 haplotype; iii ) the DRB1*0701 allele , which is associated with increased risk of IgAD is protective in T1D , while the DRB1*0102 allele shows no evidence for association with T1D [34] , [35] . These data suggest that the association with the HLA locus is the result of multiple independent effects , some of which are shared between different diseases . The characterization of the association signal in the HLA region in different diseases may , therefore , contribute important insights into the location of shared genetic effects , and may help in the identification of the causal variants in this genomic region . In addition to the association of the extended HLA haplotypes , we found further evidence for association of a few additional independent markers in the HLA locus . Most notable were the associations with SNPs in the BTNL2-HLA-DRA locus on the telomeric end of the Class II region and in the HLA-DQB1 region . Taken together , these data confirm that IgAD susceptibility is the result of multilocus effects that span the entire HLA region . In summary , we have mapped the primary susceptibility locus to the HLA Class II region and found evidence for the association of additional independent loci in Class III and Class I regions . Importantly , the fine-mapping strategy provided a better resolution of the individual haplotype background and their specific contribution to disease susceptibility . A summary of the putative causal alleles and their respective location on the specific haplotype background is depicted in Figure 4 . These data are consistent with a previous report from de la Concha et al . that used a similar haplotype-based analysis to determine the location of the IgAD causal variants in the HLA locus in the Spanish population [37] . Here , we build on these results and extend the findings to other European populations using a high-density SNP mapping approach to reconstruct the individual HLA haplotypes with high confidence . Taken together , these studies indicate that although the HLA locus shows strong population differences , the haplotype-specific genetic association signals are similar in the different cohorts , suggesting that conserved causal variants , present in ancestral haplotypes confer similar genetic predisposition to disease independently of the ethnic background . This information will also help generate better models to test the putative interaction of specific HLA variants with the non-HLA risk loci identified through the genome-wide association effort . In addition , the characterization of the role of each specific haplotype to disease susceptibility may provide important information about the precise location of the functional variant ( s ) within the haplotype . These data can provide a rationale to prioritize the specific conserved haplotype segments that should be targeted for future sequencing efforts . With the advent of more cost-effective sequencing technologies , the complete re-sequencing of large haplotype segments in a large targeted sample will become feasible , and may be the only definitive approach to identify the functional HLA variants present on these extremely conserved extended haplotypes .
A total of 861 IgAD patients were enrolled from 4 different centers across Europe: 418 were recruited at the Karolinska Institutet , Karolinska University Hospital Huddinge; 34 were recruited at the Landspitali – University Hospital , Reykjavik; 280 were recruited at the Hospital Universitario La Paz in Madrid; and 129 were recruited from the Finnish Red Cross Blood Service in Helsinki . The diagnosis of IgAD was obtained according to accepted guidelines , with serum IgA levels below the detection threshold ( IgA<0 . 07 g/l ) , as measured by nephelometry in multiple independent blood samples [5] , [38] . 2 , 184 geographically matched control samples were obtained from 3 independent sources: 1 , 115 control samples from the Epidemiological Investigation of Rheumatoid Arthritis ( EIRA ) Swedish inception cohort; 373 samples collected at Hospital Clínico San Carlos in Madrid; and 693 samples from the Nordic Centre of Excellence in Disease Genetics consortium ( NCoEDG; Finnish controls ) . All DNA samples were collected after approval from the relevant research ethics and committees . IgAD patients and the Spanish controls were genotyped at the Feinstein Institute , New York , and the Swedish controls were genotyped at The Genome Institute of Singapore . All genotyping was performed using the Illumina BeadChip technology , on either the HumanHap 300 or Human 610-quad chips . Genotyping data for the Finnish controls was obtained from the NCoEDG , and was generated using the Illumina CNV370 platform . To assure high quality data on the final analysis , we used stringent quality control measures , as described previously [24] . The final number of individuals passing all the quality control steps in the 3 independent case-controls series were as follows: Sweden/Iceland: 430 cases and 1 , 090 controls; Spain: 256 cases and 322 controls; and Finland: 86 cases and 564 controls . We observed minimal inflation of the median χ2 statistic in the 3 populations ( Sweden/Iceland: λGC = 1 . 05; Spain: λGC = 1 . 05; Finland: λGC = 1 . 04 ) , thus ruling out potential population stratification issues on the different case-control series . Fine-mapping of the HLA region was performed by extracting the 1 , 686 SNPs spanning 10 Mb of chromosome 6 ( 25–35 Mb ) that passed all quality control steps . SNP imputation of the HapMap2 ( release# 24 ) reference dataset was performed using IMPUTE [39] , with default settings . Imputation was performed using only the SNPs that passed quality control and were genotyped in all 3 case-control series . To ensure that only SNPs imputed with high confidence were included in the final analysis , we only tested the association of SNPs reaching an imputation score ( proper_info statistic ) >0 . 8 in each case-control series . The IgAD patients from Sweden , Iceland and Finland were genotyped at the HLA-B , -DRB1 and -DQB1 loci using PCR-SSP [40] employing the HLA-B low resolution and the HLA-DQ , -DR SSP Combi Tray kits from Olerup SSP AB , Saltsjöbaden , Sweden . In the Spanish samples , HLA-B was typed using the Low Resolution SSP Typing kit by Biosynthesis ( Lewisville , TX ) . HLA-DRB1 typing ( and subtyping ) and -DQB1 typing were conducted by PCR amplification and hybridization with allele-specific oligonucleotides [41] . The SNP data was used to impute the common HLA-B , -DRB1 and -DQB1 alleles in all the samples studied . The use of SNP data has been previously shown to be useful in inferring HLA alleles by taking advantage of the strong LD structure in the region . de Bakker et al showed that using genotype information from just 1 to 3 neighboring SNPs ( tagging SNPs ) , they were able to accurately infer the genotypes of the most common HLA alleles [42] . Nevertheless , the accuracy of this tagging SNP approach is affected by the variability of most classical HLA alleles , and does not provide complete information about the haplotype background surrounding the HLA alleles and about the occurrence of internal recombination events within the haplotypes . More recently , a different approach has been proposed , using multi-maker SNP data across the entire HLA region to infer the haplotype context surrounding the alleles , and to provide a more accurate estimation of the specific HLA alleles [43] . The SNP selection strategy employed in this study for haplotype phasing included the initial selection of 11 HLA-tagging SNPs that were found to be the most informative for the imputation of the more common HLA-B , -DRB1 and -DQB1 alleles , using the tagger algorithm implemented in PLINK [44] . Then , we supplemented the set of tagging SNPs with a second set of 38 SNPs distributed uniformly across the classical HLA region . Importantly , to capture the full diversity of the haplotype structure and to map the individual recombination events , SNP selection was restricted to common and independent ( low pairwise r2 ) SNPs . Given the lack of accuracy of the tagging approach to infer most HLA-B alleles , we included a higher density of SNPs for the HLA Class I region . The genotype information from the final set of 49 SNPs ( listed in Table S3 ) was used to reconstruct the individual long-range haplotypes , using PHASE v2 . 1 [45] , with the standard pre-defined parameters . Missing genotypes from the final set of 49 SNPs were also imputed during the phasing step . The imputation of the common HLA alleles was then performed by aligning the phased haplotypes and by defining , visually , the longest segment of consecutive SNPs surrounding the tagging SNP that uniquely identifies each common HLA allele . This strategy was possible because we had access to a large number of samples that were HLA typed at the HLA-B , HLA-DRB1 and HLA-DQB1 genes . In fact , 2 or 4-digit HLA-B typing data was available for 410 IgAD patients , while HLA-DRB1 and –DQB1 data was available from 534 and 638 IgAD patients respectively ( Table S4 ) . To test the performance of the method , the sensitivity , specificity and positive predictive value ( PPV ) were calculated on both the training set and on an additional validation set of 79 IgAD patients that were used to determine the allele-specific haplotype segments . Estimation of the sensitivity , specificity and PPV of the imputed alleles in the validation set was very high ( Table S5 ) . The only notable exceptions were the B*0702 allele , which was found to be rare in the validation cohort , and the DQB1*03 allele ( Table S5 ) . Given the computing-intensive nature of the haplotype phasing , the number of SNPs used for this step was limited , and , most likely , insufficient to fully characterize rare HLA alleles . However , it was sufficient to accurately depict the haplotype background of the common HLA alleles that have previously been associated with risk or protection for IgAD .
|
The human leukocyte antigen ( HLA ) locus is robustly associated with many immune-mediated conditions . However , identification of the genetic variants contributing to the disease pathophysiology has been greatly hampered by the extensive chromosomal conservation within this genomic region . To better understand the association of the HLA locus in selective IgA deficiency ( IgAD ) , we used an extensive genotyping database from a recent genome-wide association study ( GWAS ) to generate a high-density SNP map of this region in a combined sample of >2 , 700 individuals from 3 independent European populations . In addition , we took advantage of recent methodological advances to impute the more common HLA-B , -DRB1 , and -DQB1 alleles in all subjects . We confirmed the strong disease-association of the HLA locus and identified several different signals located in specific conserved HLA haplotypes contributing independent risk or protection for IgAD . Further analysis of the chromosomal sequences associated with the associated HLA alleles allowed us to refine the mapping of the susceptibility variants . These findings represent the most comprehensive high-density SNP mapping of the HLA locus in IgAD to date and provide important new information as to the location of the genetic variants contributing to this common immune deficiency .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"immunity",
"genetics",
"population",
"genetics",
"biology",
"immunology",
"genetics",
"and",
"genomics"
] |
2012
|
High-Density SNP Mapping of the HLA Region Identifies Multiple Independent Susceptibility Loci Associated with Selective IgA Deficiency
|
Nucleoporins build the nuclear pore complex ( NPC ) , which , as sole gate for nuclear-cytoplasmic exchange , is of outmost importance for normal cell function . Defects in the process of nucleocytoplasmic transport or in its machinery have been frequently described in human diseases , such as cancer and neurodegenerative disorders , but only in a few cases of developmental disorders . Here we report biallelic mutations in the nucleoporin NUP88 as a novel cause of lethal fetal akinesia deformation sequence ( FADS ) in two families . FADS comprises a spectrum of clinically and genetically heterogeneous disorders with congenital malformations related to impaired fetal movement . We show that genetic disruption of nup88 in zebrafish results in pleiotropic developmental defects reminiscent of those seen in affected human fetuses , including locomotor defects as well as defects at neuromuscular junctions . Phenotypic alterations become visible at distinct developmental stages , both in affected human fetuses and in zebrafish , whereas early stages of development are apparently normal . The zebrafish phenotypes caused by nup88 deficiency are rescued by expressing wild-type Nup88 but not the disease-linked mutant forms of Nup88 . Furthermore , using human and mouse cell lines as well as immunohistochemistry on fetal muscle tissue , we demonstrate that NUP88 depletion affects rapsyn , a key regulator of the muscle nicotinic acetylcholine receptor at the neuromuscular junction . Together , our studies provide the first characterization of NUP88 in vertebrate development , expand our understanding of the molecular events causing FADS , and suggest that variants in NUP88 should be investigated in cases of FADS .
The nucleoporin NUP88 [MIM 602552] is a constituent of the nuclear pore complex ( NPC ) , the gate for all trafficking between the nucleus and the cytoplasm [1] . NUP88 resides on both the cytoplasmic and the nuclear side of NPCs [2] and it is found in distinct sub-complexes: on the cytoplasmic face it associates with NUP214 [MIM 114350] and NUP62 [MIM 605815] as well as NUP98 [MIM 601021] , while on the nuclear side NUP88 binds the intermediate filament protein lamin A [MIM 150330] [2–5] . The NUP88-NUP214 complex plays an important role in the nuclear export of a subset of proteins and pre-ribosomes , which is mediated by the nuclear export receptor CRM1 ( Required for chromosome maintenance , alias exportin 1 , XPO1 [MIM 602559] ) [6–8] . Depletion of NUP88 alters the intracellular localization of NF-κB proteins [9–11] . Moreover , NUP88 is frequently overexpressed in a variety of human cancers and its role therein appears linked to the deregulation of the anaphase promoting complex [12 , 13] and its binding to vimentin [14] . Fetal movement is a prerequisite for normal fetal development and growth . Intrauterine movement restrictions cause a broad spectrum of disorders characterized by one or more of the following features: contractures of the major joints ( arthrogryposis ) , pulmonary hypoplasia , facial abnormalities , hydrops fetalis , pterygia , polyhydramnios and in utero growth restriction [15] . The unifying feature is a reduction or lack of fetal movement , giving rise to the term fetal akinesia deformations sequence ( FADS [OMIM 208150] ) [16] . FADS is a clinically and genetically heterogeneous condition of which the traditionally named Pena-Shokeir subtype is characterized by multiple joint contractures , facial abnormalities , and lung hypoplasia resulting from the decreased in utero movement of the fetuses [15] . Affected fetuses are often lost as spontaneous abortions ( in utero fetal demise ) or stillborn . Many of those born alive are premature and die shortly after birth . In the past , the genetic basis for these disorders was frequently unknown , but due to the recent availability of next generation sequencing , the molecular etiology is becoming increasingly understood . Many cases of FADS result from impairment along the neuromuscular axis and from mutations in genes encoding components of the motor neurons , peripheral nervous system , neuromuscular junction and the skeletal muscle . Genes encoding components critical to the neuromuscular junction and acetylcholine receptor ( AChR ) clustering represent a major class of FADS disease genes , these include RAPSN [MIM 601592] [17 , 18] , DOK7 [MIM 610285] [19] , and MUSK [MIM 601296] [20] , as well as mutations in the subunits of the muscular nicotinic acetylcholine receptor ( AChR ) [17 , 21] . These mutations are expected to affect neuromuscular junctions [22] . Here , we report a Mendelian , lethal developmental human disorder caused by mutations in NUP88 . We demonstrate that biallelic mutations in NUP88 are associated with fetal akinesia of the Pena-Shokeir-like subtype . We confirm in zebrafish that loss of Nup88 impairs locomotion behavior and that the mutant alleles are functionally null . We show that loss of NUP88 affects protein levels and localization of rapsyn in cell lines and subject samples . Consistent with altered rapsyn , AChR clustering in zebrafish is abnormal . We propose that defective NUP88 function in FADS impairs neuromuscular junction formation .
We performed exome sequencing and Sanger sequencing on genomic DNA from individuals affected with FADS from two families ( Fig 1A ) . Clinical and genetic findings are summarized in Table 1 , pedigrees and gene structure are shown in Fig 1A and 1B . Family A comprises four affected individuals , three male and one female ( Fig 1A; A . II . 3 , 4 , 5 , 7 ) , and four healthy siblings born to consanguineous parents of Palestinian origin . Exome sequencing of the last affected fetus A . II . 7 revealed a homozygous missense mutation c . 1300G>T ( p . D434Y ) in the NUP88 gene [NM_002532 . 5] ( Fig 1A ) , absent in relevant databases ( dbSNP , Ensembl , UCSC , TGP , ExAC , HGMD , gnomAD ) . Sanger sequencing revealed identical homozygous missense mutation in the third affected fetus ( Fig 1A , A . II . 5; S1A Fig ) . Both parents and unaffected siblings A . II . 1 , A . II . 2 and A . II . 6 are heterozygous carriers of the mutation , unaffected sibling A . II . 8 carries two intact alleles of NUP88 after in vitro fertilization and preimplantation diagnostic ( Fig 1A ) . DNA was unavailable from the first and second miscarriage ( A . II . 3 and A . II . 4 ) , but clinical phenotypes resemble those of the two affected individuals A . II . 5 and A . II . 7 ( Table 1 ) . In Family B , one affected son was born to healthy unrelated parents of European descent . Exome sequencing in the affected individual , his parents and his two unaffected sibs ( S1B Fig ) revealed that the individual is compound heterozygous for two NUP88 mutations , i . e . a nonsense c . 1525C>T ( p . R509* ) and a single amino-acid deletion c . 1899_1901del ( p . E634del; Fig 1A; B . II . 2 ) , absent in relevant databases . Parents and healthy siblings were heterozygous carriers of the one or the other of the mutations , thus confirming correct segregation consistent with recessive inheritance ( Fig 1A ) . The missense substitution p . D434Y and deletion p . E634del affect evolutionary highly conserved NUP88 residues ( Fig 1B ) indicating functional relevance . Accordingly , SIFT/Provean , Polyphen-2 , and MutationTaster predicted both mutations to be disease causing or potentially pathogenic ( S1 Table ) . To gain further insights into the impact of the NUP88 mutations on NUP88 protein function , we performed structural modelling as the crystal structure of human NUP88 is not known . Models obtained ( see Methods ) predicted the N-terminal domain ( NTD ) to form a 7-bladed ß-propeller , set up in a ( 4 , 4 , 4 , 4 , 4 , 4 , 3 ) arrangement of ß-strands and no Velcro lock as typical for classical ß-propellers ( Fig 2A ) . Around 60 residues precede the ß-propeller and are located at the bottom or side of the propeller thereby shielding 2–4 blades in their vicinity ( Fig 2A ) . The model reveals high similarity to the PDB deposited structures of Nup82 from Baker’s yeast and Nup57 from Chaetomium thermophilum ( Fig 2B ) . The most prominent differences are a loop region and a helix-turn-helix ( HTH ) motif emanating from blades 4 and 5 , respectively ( Fig 2B ) . Models obtained for NUP88’s C-terminal domain ( CTD ) exhibited low reliability , but the CTD , in analogy to its yeast homolog , is likely composed of extended α-helices ( Fig 2C ) that form trimeric coiled-coils , either in cis or in trans . In this context , an arrangement with its complex partners NUP214 and NUP62 in trans is most likely , as described for the yeast counterpart of the complex [23 , 24] . According to the model structure , the p . D434Y mutation is located in the loop of a HTH motif between the two outermost ß-strands of blade 6 ( Fig 2B , overall view; Fig 2D , magnification ) . The mutation likely leads to a decrease in the interaction with one of the neighboring proteins , thereby leading to a destabilization of the complex . The nonsense mutation c . 1525C>T resulting in p . R509* is located just after the ß-propeller in the linker region to the CTD resulting in a complete loss of all α-helices . Thus , the interaction of NUP88 with its complex partners is likely reduced to only propeller interactions , if the protein is not completely lost due to nonsense mediated decay of the mRNA . The p . E634del mutation is located in the middle of the CTD sequence and predicted to lie in the last fifth of an extended helix . The deletion results in a frame-shift of the remainder of the α-helix , which shifts the following residues by about a third of a helical turn and thus disrupts the interaction pattern of all following residues , which , as a consequence , decreases the overall stability of the interactions within this helix bundle . FADS is a developmental disorder and to study the function of NUP88 in vertebrate development , we used a zebrafish ( D . rerio ) model . We first examined the spatial expression of nup88 during embryonic development . The single zebrafish nup88 orthologue ( ENSDARG00000003235 ) encodes a protein of 720 amino acid translated from a single 2410 bp transcript . The predicted translated gene product shares 63% identity and 75% similarity with human NUP88 . Whole-mount in situ hybridization ( WISH ) and RT-PCR analysis in wild-type AB zebrafish showed that nup88 transcripts are maternally deposited early in development ( S2A and S2B Fig , four-cell-stage embryos ) and then ubiquitously expressed at 5 hours post fertilization ( hpf ) . By 24 hpf , while expressed ubiquitously , particularly high levels of nup88 mRNA were detected in highly proliferative frontal regions of the embryo , i . e . the central nervous system , brain , eye and anterior trunk . At 72 hpf , nup88 transcript levels are decreasing in these frontal regions and only slightly higher than in other regions of the zebrafish larvae . Similar expression patterns in the developing zebrafish have been described for the two NUP88-binding partners , NUP98 and NUP62 [25 , 26] . To study the impact of nup88 deletion on zebrafish development , we used the nup88sa2206 allele generated by the Zebrafish Mutation Project [27 , 28] . Heterozygous nup88sa2206 carriers were outcrossed for four generations with wild-type AB zebrafish prior to phenotypic analysis . The nup88sa2206 allele is characterized by a nonsense mutation , c . 732T>A ( Fig 3A ) , resulting in a premature stop codon at amino acid 244 . nup88 mRNA levels are reduced by about 90% in 5 dpf nup88 mutants ( see below ) , suggesting that the mRNA is subjected to nonsense-mediated decay . For the purpose of this study , nup88sa2206/sa2206 is therefore referred to as nup88-/- . During early stages of development and up to 3 dpf , no marked differences in morphological features of nup88-/- compared to nup88+/+ and nup88+/- siblings were observed . Starting at 4 dpf , phenotypic alterations became visible: smaller head and eyes , lack of a protruding mouth , downwards curvature of the anterior-posterior axis , abnormal gut and aplastic swim-bladder ( Fig 3B ) . Further analyses of the cranial abnormalities revealed that nup88-/- larvae exhibit severe defects in the ventral viscerocranium formed by seven cartilaginous pharyngeal arches [29 , 30] . In nup88-/- larvae , the posterior pharyngeal arches 3–7 were dramatically reduced , distorted or even absent ( Fig 3C ) . The reduced size of head and eyes correlated with an increase in apoptosis in the head of nup88-/- embryos ( Fig 3D ) . Apoptotic cells , as assessed by acridine orange staining , were readily detected in the eyes , the brain and the anterior trunk of 35 hpf mutant embryos , but not in other parts of the body ( S2C Fig ) . Together , these data indicate that nup88 mutants are phenotypically similar to the large class of jaw and branchial arch zebrafish mutants , designated as the flathead group [31 , 32] . Disruption of nup88 furthermore led to impaired survival with lethality occurring at or after 9 dpf ( Fig 3E ) . To address the question whether NUP88 mutations identified in the familial cases of FADS affect NUP88 function , we performed phenotypic rescue experiments in zebrafish . Two of the three mutated residues in the uncovered FADS cases are conserved between human and zebrafish ( Fig 1B ) , hence we introduced the corresponding mutations on zebrafish expression constructs by site-directed mutagenesis . Human c . 1300G>T , p . D434Y corresponds to c . 1240G>T , p . D414Y in zebrafish and human c . 1899_1901del , p . E634del to zebrafish c . 1837_1839del , p . E613del . Human p . R509 is not conserved in zebrafish , therefore we inserted a stop codon at c . 1468-1470>TGA , p . H490* , a residue in a similar position as human R509 . Subsequently synthetic mRNA corresponding to each variant was microinjected into one-cell stage nup88-/- mutants and their rescue capacity was assessed by evaluating the eye size as well as the number and morphology of the pharyngeal arches . Injection of wild-type ( WT ) nup88 mRNA partially rescued the developmental defects of the 5 dpf mutant larvae as indicated by significant restoration of the eye size ( Fig 4A and 4B ) and a significant increase in the number of pharyngeal arches ( Fig 4A and 4C ) . In addition , the arches resembled the morphologically wild-type structures ( Fig 4A and 4D ) . In contrast to WT nup88 mRNA , injection of nup88 D414Y mRNA , nup88 H490* mRNA as well as nup88 E613del mRNA failed to suppress the nup88-/- phenotypes ( Fig 4A–4D ) , indicating that these nup88 mutant transcripts are functionally null . Although the results suggest that the encoded variant proteins are functionally inactive , the lack of rescue could also be the consequence of transcript or protein instability . To further investigate the implication of NUP88 in the etiology of FADS , we next determined whether locomotor function was impaired by loss of Nup88 using locomotion and touch-evoked escape assays in nup88-/- zebrafish . Zebrafish embryos develop spontaneous muscle contractions at 18 hpf [33] , therefore we first analyzed the coiling behavior of nup88-/- embryos as compared to nup88+/+ and nup88+/- embryos at 22–24 hpf . We did not detect problems in coiling behavior in nup88-/- embryos at this developmental stage ( S3A Fig and S1 Movie ) . Next , we analyzed spontaneous swimming activity at 4 dpf ( Fig 5A ) and found that only about 35% of the nup88-/- larvae showed spontaneous movement as compared to ~83% of nup88+/+ and about 73% of nup88+/- larvae ( Fig 5B ) . Moreover , those moving nup88-/- larvae displayed drastically reduced motor activity , traveled shorter distance ( Fig 5C ) and initiated swim bouts less often ( Fig 5D ) . In contrast , statistically significant differences in the mean velocity were not observed in nup88 mutant larvae ( Fig 5E ) . We next performed touch-evoked escape response assays at 3 dpf and 4 dpf . At 3 dpf , percentages of responsive animals and response duration were not significantly different between wild-type and nup88 mutant larvae ( S3B Fig ) . At 4 dpf , percentages of responsive animals were also not significantly different between wild-type and nup88 mutant larvae ( Fig 5F ) , however , the response duration among the larvae that moved was significantly reduced in homozygous nup88 mutants in comparison to wild-type zebrafish . Interestingly also heterozygous nup88 mutants showed a shortened response duration , although less significant ( Fig 5G and 5H and S2–S4 Movies ) . Next we wanted to assess whether the NUP88 mutations identified in the individuals with FADS interfere with the recruitment of NUP88 to NPCs . Due to a lack of relevant cell lines and the limited availability of tissue samples from affected fetuses , we performed immunofluorescence microscopy of GFP-tagged NUP88 proteins . Upon expression in HeLa cells , wild-type NUP88 and all mutants were co-localizing with the NPC marker mAb414 , although recruitment of the NUP88 p . R509* and p . E634del mutants to NPCs appeared reduced compared to wild-type NUP88 and the p . D434Y mutant ( S4A Fig ) . Moreover , all forms of NUP88 also localized partially to the cytoplasm , as previously seen for NUP88 overexpression [2 , 12] , and NUP88 p . R509* to the nucleoplasm ( S4A Fig ) . To define the effect of the mutations in NUP88 on the interface with its binding partners NUP214 , NUP98 and NUP62 , we employed GFP trap affinity purification in combination with Western blot analysis of lysates from HeLa cells expressing the GFP-NUP88 mutants . We found that NUP88 and the p . D434Y mutant co-purified NUP214 and NUP62 , while the p . R509* and the p . E634del mutant did not so ( Fig 6A ) . Binding of NUP214 to GFP alone was similar as compared to the p . R509* and the p . E634del mutants , indicating some non-specific binding of the NUP214 and/or the antibodies to GFP . The disrupted interaction between NUP88 and NUP214 , however , did not impair NUP214 localization at NPCs ( Fig 6B ) , whereas NUP62 association with NPCs was reduced in cells expressing NUP88 E634del ( Fig 6C ) . Our GFP trap assays further showed that NUP98 associated with NUP88 and all mutant forms ( Fig 6A ) . Consequently , NUP98 association with NPCs appeared unaffected in cells expressing NUP88 mutants ( Fig 6D ) . Furthermore , the disease-related mutations in NUP88 did not affect the organization of the nuclear envelope as revealed by immunofluorescence analysis of lamin A/C ( LA/C;S4B Fig ) , Western blots analysis of protein levels of lamin A/C and GST-pull-down down assays with NUP88 and the p . D434Y mutant ( S4C and S4D Fig and S1 Methods ) . Moreover and consistent with data from higher vertebrates , NUP88 appears to not be essential for NPC integrity: immunolabelling of NPCs using a monoclonal antibody ( mAb414 ) recognizing a subset of nucleoporins appears unaffected by the loss of nup88 in D . rerio brain sections ( S4E Fig and S1 Methods ) and in muscle histology sections of individual B . II . 2 ( S4F Fig ) . Additionally , depletion of NUP88 from HeLa cells using siRNAs had no visible effect on the distribution of lamin A/C ( LA/C ) and the NE marker proteins emerin , Nesprin-1 , Nesprin-2 , Sun1 , and Sun2 ( S5 Fig ) . As NUP88 is critically involved in CRM1-dependent nuclear export of proteins , we further asked whether the mutations in NUP88 affect nuclear import and/or export , but we observed no defects in general nuclear protein import or export ( S6A Fig ) or the three CRM1 targets mTOR , p62/SQSTM and TFEB ( S6B Fig ) . Impeded formation of AChR clusters at the neuromuscular junction ( NMJ ) is one cause of FADS . Given the central role of rapsyn in AChR clustering and in FADS [17 , 18 , 21] , we therefore asked whether a loss of NUP88 function would negatively affect rapsyn and depleted NUP88 by siRNAs from HeLa and C2C12 cells and monitored protein levels of rapsyn by Western blotting . As shown in Fig 7A , depletion of NUP88 from HeLa cells in fact coincided with a decrease in rapsyn levels . Similarly , reduced rapsyn levels were observed in C2C12 cells depleted for NUP88 using siRNAs ( Fig 7B and 7C ) and shRNAs ( Fig 7C ) . Quantification revealed that siRNA treatment led to reduction of NUP88 by 80–90% and at the same time a reduction of rapsyn by 40–60% ( Fig 7C ) . In contrast to that , NUP88 downregulation had no effect on MuSK levels , another key player in FADS [20] , and no effect on known NUP88 targets , such as CRM1 and NF-κB levels ( Fig 7A and 7B ) . Muscle biopsy of affected individual B . II . 2 and immunohistochemistry on paraffin sections furthermore showed weaker staining and irregular distribution of rapsyn in the cytoplasm in comparison to biopsy samples from a control fetus ( Fig 7D ) . Rapsyn is known to not only localize to the plasma membrane , but also to the cytoplasm [34–36] . Rapsyn protein levels could not be determined in zebrafish due to a lack of antibodies , but qRT-PCR analyses revealed a reduction of rapsn mRNA by about 20% , while nup88 mRNA levels were reduced by about 90% ( Fig 7E ) . Consistent with reduced rapsyn levels , we observed impaired AChR clustering in fast-twitch muscle fiber synapses , but not in myoseptal synapses of the 5 dpf zebrafish trunk ( Fig 7G ) . Quantification of the size of individual AChR cluster in WT and mutant zebrafish revealed that the diameter of the AChR clusters was significantly reduced in nup88-/- larvae as compared to nup88+/+ larvae ( Fig 7H ) . Interestingly , AChR cluster size was also reduced in nup88+/- larvae , both in comparison to WT and nup88-/- larvae . This reduced size of AChR clusters in the heterozygotes may account for the observed defects in touch-evoke response ( Fig 5H ) . In accordance with impaired neuromuscular junction formation as a consequence of loss of nup88 , muscle organization in zebrafish appeared indistinguishable in electron micrographs from nup88+/+ and nup88-/- larvae ( Fig 7F ) . Similarly , in affected fetus B . II . 2 skeletal muscle structure was , based on the autopsy report , intact . Thus , both in vitro and in vivo evidence support the notion that loss-of-function of NUP88 has a negative effect on rapsyn , which likely affects AChR clustering and proper formation of neuromuscular junctions .
Here , we have identified biallelic homozygous and compound heterozygous mutations in NUP88 as a cause of fetal akinesia . We demonstrate that the mutations in NUP88 lead to a loss-of-function phenotype , which coincides with reduced spontaneous motor activity and touch-evoked escape response in zebrafish . Consistent with the fact that the mutations in NUP88 affect different regions of the protein , we observed distinct effects of the mutants on binding to NUP214 and NUP62 in GFP-trap assays , whereas binding to NUP98 appears indistinguishable between wild-type and mutant forms of NUP88 ( Fig 6A ) . This suggests that impaired interaction with partner nucleoporins may contribute , but are unlikely to be causative for NUP88 malfunction in FADS . Our data further suggest that NUP88 malfunction in FADS is at least in part due to dysfunctional rapsyn , a known key player in FADS , and consequently impaired AChR clustering and neuromuscular junction formation . Muscle integrity , in contrast , appears grossly unaffected by a loss of NUP88 . Genetic disruption of nup88 in zebrafish led to pleiotropic morphological defects , including micrognathia , smaller head and eyes , distortion of the body axis , and aplastic swim bladder ( Figs 3 and 4 ) , and to impaired locomotor behavior ( Fig 5 ) . These phenotypes parallel defects observed in human fetuses affected by FADS , such as reduced fetal movement , micrognathia , joint contractures and lung hypoplasia ( for detailed comparison see Table 2 ) . The reduced head and eye size likely originates from increased apoptosis of neuronal cells ( Fig 3D ) . nup88 inactivation affected spontaneous movement of zebrafish from 4 dpf onwards , which resembles the onset of symptoms in the affected fetuses at about week 18 of gestation . The impaired touch-evoked response seen in the mutant zebrafish further matches the absence of reflex response observed in at least some fetuses ( Family A , I . 2 and B . F . , personal communication ) . Rescue experiments in zebrafish with wild-type or nup88 mutants revealed that none of the mutant mRNA could restore a wild-type-like phenotype ( Fig 4 ) , indicating that NUP88 mutations in individuals affected with FADS are loss-of-function variants . The zebrafish model developed here therefore provides a valuable in vivo system to further test nup88 deficiency . This in turn will be key in understanding the role of NUP88 in the etiology of FADS and its function in embryonic development . Reduced rapsyn levels might partly cause fetal akinesia upon loss of NUP88 . Rapsyn is one of the many contributing proteins required for the correct assembly of the AChR and is particularly involved in AChR assembly and localization to the cell membrane [37 , 38] . We observed reduced rapsyn protein levels in the absence of functional NUP88 in cellular assays using human and mouse cell lines in combination with siRNA and shRNA-mediated depletion of NUP88 ( Fig 7 ) . Due to a lack of cell lines derived from affected individuals , we could analyze rapsyn only in histological sections from muscle of one affected fetus , which revealed a weaker staining for and a perturbed intracellular localization of rapsyn ( Fig 7D ) . Consistent with aberrant rapsyn expression , AChR clustering in trunk regions of nup88+/- and nup88-/- zebrafish was impaired ( Fig 7G and 7H ) . Rapsyn protein levels could not be determined in zebrafish due to a lack of antibodies , but qRT-PCR analyses revealed a reduction of its mRNA by about 20% ( Fig 7E ) . Our data therefore suggest that absence of functional NUP88 causes fetal akinesia at least in part through misregulation of rapsyn expression . How loss of NUP88 results in reduced rapsyn levels on a mechanistic level remains to be seen . Moreover , this likely is not the only pathway by which NUP88 acts in NMJs , as effects on only one cellular pathway would be indeed ( i ) very surprising for a nucleoporin per se , ( ii ) irreconcilable with the cranial defects observed in human and zebrafish , and ( iii ) not in line with the broad central nervous system expression pattern of nup88 in zebrafish . Our data , however , demonstrate ( i ) that the nup88 spectrum of phenotypes indeed include locomotor defects and that therefore NUP88 deficiencies might result in FADS in humans , and ( ii ) that human alleles are dysfunctional . We further observed that ( iii ) NMJ defects correlate to this phenotype . Whether this is the primary cause of akinesia is impossible to determine given the pleiotropic effects of Nup88 deficiency , but at least they are sufficient to explain the phenomenon . Further mechanistic details will be subject for future studies .
Exome sequencing in one affected baby ( A . II . 7 ) was performed at the Clinical Exome Sequencing ( CES ) at University of California , Los Angeles and the sequencing report is on hand . In total 22 , 843 DNA variants were identified , including 21 , 625 single nucleotide substitutions and 1 , 218 small deletions/insertions ( 1–10 bp ) : the data were consistent with a high quality genomic sequence and fall within normal human genomic variation quality parameters . Estimated from these data , about 93% of the exome was reliably sequenced with at least 10x coverage . In total , 5 homozygous and 329 rare heterozygous protein-altering variants of uncertain clinical significance were identified across 313 genes . A rare autosomal-recessive model of inheritance with homozygous causative mutations due to consanguinity of the parents was assumed . After applying appropriate filters , a novel homozygous variant c . 1300G>T , p . D434Y in the NUP88 gene [NM_002532 . 5] was identified in the subject’s DNA . This variant had not been previously observed in the general population and was predicted to be deleterious/probably damaging by three in silico prediction algorithms ( S1 Table ) . Additionally , the homozygous mutation c . 1300G>T , p . D434Y in NUP88 was confirmed in a second affected fetus A . II . 5 by Sanger sequencing . Exome sequencing was performed on DNA from the proband B . II . 2 , both healthy sisters and both parents from the European family , as outlined previously [39] . Exome enrichment was performed on DNA using an Ampliseq Whole Exome kit , ( Thermo Fisher Scientific ) . Briefly , a total of 100 ng of DNA was amplified in 12 separate PCR pools , each containing ~25 , 000 primer pairs . After amplification , the individual reactions were pooled and digested to degrade the PCR primers . Next , barcoded sequencing adaptors were ligated and the library was purified using AMPure beads ( Beckman Coulter ) , amplified and purified again and analyzed on a 2100 Bioanalyzer ( Agilent Technologies ) . Libraries were diluted to 18–26 pM and attached to Ion Sphere Particles ( ISPs ) using an Ion Proton Template 200 V3 kit and sequenced on a P1 sequencing chip for 520 flows on an Ion Proton sequencer ( Ion Sequencing 200 kit V3 ) . Two samples were pooled and sequenced on a single chip . Following sequencing , reads were trimmed to remove low quality bases from the 3’ end and mapped to the human genome reference sequence ( HG19 ) using tmap ( Torrent Suite 4 . 2 ) . Variant calling was performed using the Torrent Variant Caller with custom settings optimized for whole exomes and the data was annotated using ion Reporter 4 . 0 . Variants were filtered with ANNOVAR [40] against ENCODE GENECODE v . 19 , 1000genomes ( threshold >0 . 5% ) , dbSNP138 common databases and against a list of in-house common variants . Genes with variants that fitted an autosomal recessive inheritance pattern and co-segregated with disease in the family were prioritized . NUP88 was the only candidate gene that harbored variants compatible with an autosomal recessive inheritance pattern in this family ( S1B Fig ) . Individual B . II . 2 was compound heterozygous for an in-frame 3 bp deletion in exon 14 ( c . 1899_1901del , p . E634del ) and a nonsense mutation c . 1525C>T , p . R509* in exon 11 . Mutation c . 1899_1901del mutation was inherited on the maternal allele and the c . 1525C>T paternally . Healthy sister B . II . 1 is heterozygous for the maternal mutation , healthy sister B . II . 3 is a heterozygous carrier of the paternal mutation ( Fig 1A , pedigree Family B ) . Human NUP88 protein sequence Q99567 ( UniProt database ) was used for structural investigation and modelling using an N-terminal and C-terminal part , according to predictions . Database search comparisons using HHPRed [41–45] revealed as best hits for the N-terminal domain ( NTD ) : Nuclear pore complex proteins Nup82 from S . cerevisiae and Chaetomium thermophilum ( PDBid: 3pbp A , 3tkn_A and 5cww_B , respectively ) and for the C-terminal domain ( CTD ) Nup57 from Chaetomium thermophilum ( PDBid: 5cws E ) and Nup54 from Homo sapiens ( PDBid: 5ijn F ) . The high similarity to yeast and Chaetomium Nup82 , the structures of which have been solved in part [46–48] strongly indicate that the NTD also forms a ß-propeller and predictions suggest that the CTD is in an all helical arrangement ( PsiPred and JPred4 ) [49 , 50] . In order to gain insight into the putative fold and the location of the disease-related residues identified in hsNUP88 , four servers were used for structure prediction . Phyre2 , Robetta , I-Tasser , RaptorX [51–57] were supplied with either full length sequence or only the NTD ( residues 1–495 ) or CTD ( residues 496–741 ) in case of residue limitations or implausible models resulting from full length submission . All modelling was performed using standard settings . All figures were generated using Pymol . Zebrafish ( Danio rerio ) were raised and bred at 28°C on a 14 h/10 h light/dark cycle . Embryos and larvae were raised in egg water ( 0 . 3 g/l Instant Ocean Salt , 75 mg/l CaSO4; 1 mg/l Methylene Blue ) . The line carrying nup88sa2206 allele was obtained from Zebrafish Mutation Project , Knockouts for Disease Models ( http://www . sanger . ac . uk/sanger/Zebrafish_Zmpgene/ENSDARG00000003235#sa2206 ) . Heterozygous fish for nup88 were out crossed for four generations with wild-type fish before analysis . All animal experiments were performed in accordance with the rules of the State of Belgium ( protocol approval number: CEBEA-IBMM-2017-22:65 ) . Capped messenger RNA was synthesized using the mMESSAGE mMACHINE kit ( Ambion ) . The following expression plasmids were generated and used in this study: the full-length zebrafish nup88 ORF was cloned from 48 hpf cDNA and recombined into BamHI-XhoI digested pCS2 using In-fusion cloning ( TaKaRa ) . nup88 mutants corresponding to the sequences identified in human fetal akinesia cases were generated by site-directed mutagenesis using QuikChange Lightning Site-Directed Mutagenesis Kit ( Agilent Technologies ) . All primer sequences are listed in S2 Table . mRNAs ( 300 pg ) were injected at the one-cell stage . nup88sa2206 genotyping was performed by RFLP assay using MseI restriction of a 150 bp PCR product . Real-time PCR was done at various stages of embryonic and larval development ( 4-cell to 5 dpf ) . Primer sequences are listed in S2 Table . Staging of embryos was performed according to [58] . Alcian Blue staining and histology were performed as described elsewhere [59] . All images were acquired using an Olympus SZX16 stereomicroscope and an Olympus XC50 camera using the imaging software Cell* after embryo anesthesia with a low dose of tricaine . Live embryos at 24 hpf , 36 hpf , 48 hpf were dechorionated and immersed in egg water containing 5 μg/ml acridine orange . They were incubated at 28 . 5°C for 15 min in the dark and then thoroughly washed with egg water . Embryos were mounted in low-melting agarose for positioning and immediately imaged using an Axio Observer Z1-1 microscope . Images were processed using Zeiss Zen software . Spontaneous tail coiling of 22–24 hpf embryos , placed in the grooves of an agarose chamber , was recorded for 2 min at 30 frames per second . Coiling events were scored manually . To analyze spontaneous locomotion , 4 dpf larvae were placed into 96-well plates ( one larva per well ) . Their behavior was recorded at 30 frames per second for 5 min and quantified using EthoVision XT 8 . 5 software ( Noldus ) . Touch-induced escape responses were analyzed in 4 dpf head-restrained larvae . Larvae were first embedded in 2% low melting point agarose , and then the agarose surrounding their tail was removed with a blade . Escape behavior was induced by touching the tail of larvae with a plastic pipette tip and recorded at 300 frames per second . Duration of behavioral responses was quantified with ImageJ . Total RNA of 5 dpf nup88 mutant and WT zebrafish larvae was extracted using TRIzol reagent ( Invitrogen , Carlsbad , CA , USA ) according to the manufacturer’s protocol . Genomic DNA was eliminated prior to reverse transcription using PrimeScript RT Reagent Kit with gDNA Eraser ( Perfect Real Time; RR047A , TaKaRa , Dalian , China ) . Total RNA ( 5 μg ) was next reverse-transcribed in 20 μl final volume according to manufacturer’s protocol . Primers for target genes ( S2 Table ) were designed with Realtime PCR Tool by Integrated DNA Technologies , Inc ( https://eu . idtdna . com/scitools/Applications/RealTimePCR ) , and sequences were submitted to BLAST ( http://blast . ncbi . nlm . nih . gov/Blast . cgi ) . Quantitative PCR was then performed using SYBR Premix Ex Taq TM II ( Tli RNaseH Plus ) , Bulk kit ( RR820L; Takara , Dalian , China ) . 1 μl of cDNA was used in each 20 μl-PCR well with 300 nM primers pair final concentration . Each sample was assayed in triplicate and samples not reverse-transcribed were used as negative controls . Eight housekeeping genes were tested according to [60] and the most stable combination was determined using qbase+ qPCR analysis software ( https://www . qbaseplus . com ) . actb2 , gapdh and ybx1 were the best combination of housekeeping genes at this stage of larval development and used for normalization . Differences in amplification curves between the target genes and housekeeping genes were identified by comparing standard curve slopes . Real-Time PCR was performed using Applied Biosystems StepOnePlus Real-Time PCR System and PCR analyses were performed with qbase+ system software . For whole-mount immunostaining , zebrafish embryos were fixed in 4% PFA for 3 h at room temperature ( RT ) , washed and permeabilized with proteinase K ( 40μg/ml ) at 37°C for 1 h . For blocking , larvae were gently shaken in PBS containing 10% of normal goat serum , 1% DMSO and 0 . 8% Triton X-100 for 1 h and then , incubated with the primary anti-acetylcholine receptor antibody ( mouse mAb35 ( DSHB , Hybridoma Bank; 1:100 ) overnight at 4°C . After washing , larvae were incubated with the secondary antibody ( anti-mouse IgG Alexa Fluor 594 , 1:1000 ) for 2 h at RT and washed six times for 15 min in PBST prior to mounting and confocal imaging ( Zeiss LSM710 ) . Size of AChR cluster were determined using ImageJ . Skeletal muscle of 5 dpf nup88+/+ and nup88-/- zebrafish larvae were analyzed by transmission electron microscopy on longitudinal and transversal ultrathin sections . Briefly , larvae were fixed in 2 . 5% glutaraldehyde and 0 . 1 M sodium cacodylate buffer , pH 7 . 4 overnight at 4°C and post-fixed in 1% osmium tetroxide , 1 . 5% ferrocyanide in 0 . 15 M cacodylate buffer for 1 h at RT . After serial dehydration in increasing ethanol concentrations , samples were embedded in agar 100 ( Agar Scientific Ltd . , UK ) and left to polymerize for 2 days at 60°C . Ultrathin sections ( 80 nm thick ) were collected using a Leica EM UC6 ultramicrotome and stained with uranyl acetate and lead citrate . Images were recorded on a Tecnai10 electron microscope ( FEI ) equipped with an Olympus VELETA camera and processed using the AnalySIS software . Brain and skeletal muscle paraffin-embedded blocks were obtained from autopsy of individual B . II . 2 . Control tissues of a fetus of the same age without neuromuscular disorder were used for IHC-P . Written consent form for use of paraffin samples for functional analysis was signed by the parents . This study was performed according to the guidelines of the local ethics committee ( Ghent University Hospital ) , which does not require formal review for the use of autopsy samples . Blocks were cut with a microtome and 5 μm sections were mounted . Sections were deparaffinized and rehydrated . Heat-induced epitope retrieval was performed using Tris/EDTA , pH 9 . 0 for all tissues when using antibodies against rapsyn and with sodium citrate , pH 6 . 0 for anti-nucleoporin antibody mAb414 . 0 . 3% H2O2 for 20 min was used to block samples endogenous peroxidase followed by washes with PBS . Tissues were permeabilized using PBS containing 0 . 5% Triton X-100 for 5 min at RT and subsequently blocked with PBS containing 1% BSA for 30 min . Primary antibodies were diluted in blocking buffer and incubated overnight at 4°C in a humidification chamber . Samples were washed in PBS for 10 min . Biotinylated goat anti-rabbit-Ig ( 1:400; DAKO-E0432 ) or biotinylated goat anti-mouse ( 1:400; DAKO-E0433 ) were added for 30 min . After washing in PBS , samples were incubated 30 min with streptavidin-HRP ( DAKO-P0397 ) diluted 1:300 in PBS and thoroughly washed with PBS . DAKO Liquid DAB+ Substrate Chromogen System K3468 ( 1ml substrate + 1 drop DAB ) was added on samples and left until a brown coloring appeared ( maximum 6 min ) . After several PBS washes , the samples were counterstained with hematoxylin for 15 seconds , washed thoroughly under running tap water to remove excess staining agent and mounted for subsequent observations with Aquatex . Images were acquired with an Olympus BX41 microscope and were processed using the Olympus cellSense software . For all constructs , human NUP88 was amplified by PCR . All constructs were verified by DNA sequencing . GFP-NUP88 was produced as described previously [2] . FLAG-NUP88 was cloned into KpnI/XbaI cut pFLAG-CMV2 ( Sigma-Aldrich ) . GFP-NUP88 and FLAG-NUP88 mutants were generated by site-directed mutagenesis using the QuikChange Lightning site-directed mutagenesis kit ( Agilent Technologies ) following the manufacturer’s instructions . Primers are listed in S2 Table . The following polyclonal antibodies were used in this study for Western blotting ( WB ) , immunofluorescence ( IF ) and immunohistochemistry ( IHC ) : rabbit anti-NUP214 ( Abcam , ab70497; WB 1:5000 , IF 1:500 ) , rabbit anti-lamin A ( Sigma-Aldrich , L1293; WB 1:500 ) , rabbit anti-rapsyn ( Novus Biologicals , NBP1-85537; WB 1:500; IHC 1:200 ) , rabbit anti-MuSK ( ThermoScientific , PA5-14705; WB 1:1000 ) , rabbit anti-GAPDH ( Cell Signaling , 2118; WB 1:10 . 000 ) , rabbit anti-actin ( Sigma-Aldrich , A2066; WB 1:1000 ) , rabbit anti-Nesprin1 ( Sigma-Aldrich HPA019113; IF 1:250 ) , rabbit anti-Nesprin2 ( a kind gift of Dr . Iakowos Karakesisoglou; IF 1:50 ) , rabbit anti-Sun1 ( kind gift of Dr . Ulrike Kutay; IF 1:1000 ) , rabbit anti-Sun2 ( Sigma-Aldrich , HPA001209; IF 1:200 ) , rabbit anti-emerin ( Abcam , ab153718; IF 1:500 ) , rabbit anti-NUP98 ( Abcam , ab45584; WB 1:2000 ) . The following monoclonal antibodies were used in this study: mouse anti-Nup88 ( BD Transduction Laboratories , 611896; IF 1:500 ) , mouse mAb414 ( Covance MMS-120R; IF 1:2000; IHC 1:100 ) , mouse anti-NUP62 ( Clone 53 , BD Transduction Laboratories , 610497; WB 1: 3000 , IF 1:500 ) , mouse anti-lamin A/C ( Abcam , ab8984; WB 1:300 , IF 1:30 ) , mouse anti-FLAG ( Sigma-Aldrich F-3165; IF: 1:200 ) , rat anti-NUP98 ( Sigma-Aldrich , N1038; WB IF 1:1000 ) , mouse anti-AChR ( mAb35 , DSHB , Hybridoma Bank; IF 1:100 ) , rat anti-GFP ( 3H9 , Chromotek; WB 1:1000 ) . Secondary antibodies for immunofluorescence were the corresponding goat anti-mouse-IgG Alexa 568 ( 1:1000; Invitrogen ) , goat anti-rabbit IgG Alexa 568 ( 1:1000; Invitrogen ) . Secondary antibodies were either alkaline phosphatase coupled antibodies from Sigma/Aldrich and used at 1:20 . 000 or HRP coupled antibodies from Cell Signaling Technology at 1:8000 . All experiments were conducted in HeLa cells ( provided by Robert D . Goldman , Feinberg School of Medicine , Northwestern University , Chicago , USA ) grown in Dulbecco's modified Eagle's medium ( DMEM ) , supplemented with 10% fetal bovine serum ( FBS ) plus penicillin and streptomycin . Cells were transfected with plasmids using Turbofect transfection reagent ( Thermo Scientific Fermentas , St . Leon-Rot , Germany ) , Lipofectamine 2000 ( Life Technologies Invitrogen , Gent , Belgium ) or jetPRIME ( Polyplus , Illkirch , France ) and with siRNAs using Lipofectamine RNAiMAX ( Life Technologies Invitrogen ) following the instructions of the manufacturer . Smart-pool small interfering RNAs were obtained from Dharmacon ( GE Healthcare Europe , Diegem , Belgium ) : human NUP88 ( L-017547-01-0005 ) , mouse NUP88 ( L-054949-01-0005 ) , and non-targeting siRNAs ( D-001810-10 ) . HeLa cells were grown on glass coverslips , transfected , fixed in 4% PFA in PBS for 5 min , permeabilized with 0 . 5% Triton-X-100 in PBS for 5 min and then fixed again . Blocking was performed with 2% BSA/0 . 1% Triton-X-100 in PBS for 30 min at RT . Primary antibodies were incubated at 4°C over-night in a humidified chamber . Secondary antibodies were incubated 1 h at RT in the dark . Excess antibodies after primary and secondary antibody staining were removed by three washing steps using 0 . 1% Triton-X-100 in PBS for 5 min . Cells were imaged using a Zeiss LSM 710 ( Zeiss , Oberkochen , Germany ) confocal laser scanning microscope with Zeiss Plan-Apochromat 63x/1 . 4 oil objective . Images were acquired using the microscope system software and processed using Image J and Adobe Photoshop ( Adobe Systems , Mountain View , CA ) . HeLa cells , grown in a 10 cm dish , were transfected with GFP and GFP-NUP88 wild-type and mutant variants , respectively . Cells were grown for 48 h at 37°C in a humidified atmosphere with 5% CO2 . To harvest cells , growth medium was aspirated off , 1 ml of ice-cold PBS was added to cells and cells were scraped from dish . The cells were transferred to a pre-cooled tube , centrifuged at 500 xg for 3 min at 4°C and the supernatant was discarded . The cell pellet was washed twice with ice-cold PBS . Pellets were subsequently lysed with 200 μl of ice-cold lysis buffer ( 10 mM Tris/HCl , pH 7 . 5 , 150 mM NaCl , 0 . 5 mM EDTA , 0 . 5% NP-40 , protease-phosphatase inhibitor ) using a syringe before incubating on ice for 30 min . The tubes were centrifuged for 15 min at 16 . 000 xg at 4°C and the supernatant was transferred into a new reaction tube . Bradford assay was used to determine the protein concentration of the lysates and 300 μg of protein lysate adjusted to 500 μl in dilution buffer ( 10 mM Tris/HCl , pH 7 . 5 , 150 mM NaCl , 0 . 5 mM EDTA , protease-phosphatase inhibitor ) was added to 25 μl of GFP-Trap_MA beads ( ChromoTek , Planegg-Martinsried , Germany ) . Beads were prewashed twice with dilution buffer . The beads and the lysates were incubated 1h at 4°C on an end-to-end rotor . The magnetic beads were then washed three times in dilution buffer containing 150 mM , 250 mM and 500 mM NaCl following the addition of 20 μl of 2x SDS-sample buffer ( 120 mM Tris/HCl , pH 6 . 8 , 20% glycerol , 4% SDS , 0 . 04% bromophenol blue , 10% β-mercaptoethanol ) and boiling at 95°C for 10 min . The eluates were subsequently loaded on to 7% polyacrylamide gels and Western blot was carried out ( see below ) . HeLa and C2C12 ( provided by Vincent Mouly , The Pitié-Salpêtrière Hospital , Institute of Myology , Paris , France ) cells were lysed in lysis buffer ( 50 mM Tris-HCl , pH 7 . 8 , 150 mM NaCl , 1% Nonidet-P40 and protease inhibitor cocktail tablets ( Roche , Basel Switzerland ) ) . 20 μg of protein were loaded and separated by sodium dodecyl sulfate-polyacrylamide ( 5% or 7% ) gel electrophoresis ( SDS-PAGE ) . The proteins were transferred onto a PVDF membrane ( Immobilon-P , Millipore ) and the membranes were blocked with TBS containing 0 . 1% Tween 20 and 5% non-fat dry milk for 1 h . The membranes were then incubated for 1 h in blocking solution containing a primary antibody followed by washing 3x in TBS containing 0 . 1% Tween 20 and 5% non-fat dry . The membranes were next incubated with secondary antibodies for 1 h , washed 3x in TBS and developed . X-ray films were scanned and processed using ImageJ .
|
Fetal movement is a prerequisite for normal fetal development and growth . Fetal akinesia deformation sequence ( FADS ) is the result of decreased fetal movement coinciding with congenital malformations related to impaired fetal movement . FADS may be caused by heterogenous defects at any point along the motor system pathway and genes encoding components critical to the neuromuscular junction and acetylcholine receptor clustering represent a major class of FADS disease genes . We report here biallelic , loss-of-function mutations in the nucleoporin NUP88 that result in lethal FADS and with this the first lethal human developmental disorder due to mutations in a nucleoporin gene . We show that loss of Nup88 in zebrafish results in defects reminiscent of those seen in affected human fetuses and loss of NUP88 affects distinct developmental stages , both during human and zebrafish development . Consistent with the notion that a primary cause for FADS is impaired formation of the neuromuscular junction , loss of Nup88 in zebrafish coincides with abnormalities in acetylcholine receptor clustering , suggesting that defective NUP88 function in FADS impairs neuromuscular junction formation .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"fish",
"neuromuscular",
"junctions",
"medicine",
"and",
"health",
"sciences",
"hela",
"cells",
"nervous",
"system",
"biological",
"cultures",
"vertebrates",
"electrophysiology",
"neuroscience",
"animals",
"biological",
"locomotion",
"animal",
"models",
"osteichthyes",
"developmental",
"biology",
"mutation",
"model",
"organisms",
"experimental",
"organism",
"systems",
"nonsense",
"mutation",
"cell",
"cultures",
"embryos",
"eyes",
"research",
"and",
"analysis",
"methods",
"embryology",
"animal",
"studies",
"life",
"cycles",
"head",
"cell",
"lines",
"zebrafish",
"eukaryota",
"anatomy",
"synapses",
"neurophysiology",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"ocular",
"system",
"cultured",
"tumor",
"cells",
"larvae",
"organisms"
] |
2018
|
Biallelic mutations in nucleoporin NUP88 cause lethal fetal akinesia deformation sequence
|
River blindness ( onchocerciasis ) causes severe itching , skin lesions , and vision impairment including blindness . More than 99% of all current cases are found in sub-Saharan Africa . Fortunately , vector control and community-directed treatment with ivermectin have significantly reduced morbidity . Studies in Mali and Senegal proved the feasibility of elimination with ivermectin administration . The treatment goal is shifting from control to elimination in endemic African regions . Given limited resources , national and global policymakers need a rigorous analysis comparing investment options . For this , we developed scenarios for alternative treatment goals and compared treatment timelines and drug needs between the scenarios . Control , elimination , and eradication scenarios were developed with reference to current standard practices , large-scale studies , and historical data . For each scenario , the timeline when treatment is expected to stop at country level was predicted using a dynamical transmission model , and ivermectin treatment needs were predicted based on population in endemic areas , treatment coverage data , and the frequency of community-directed treatment . The control scenario requires community-directed treatment with ivermectin beyond 2045 with around 2 . 63 billion treatments over 2013–2045; the elimination scenario , until 2028 in areas where feasible , but beyond 2045 in countries with operational challenges , around 1 . 15 billion treatments; and the eradication scenario , lasting until 2040 , around 1 . 30 billion treatments . The eradication scenario is the most favorable in terms of the timeline of the intervention phase and treatment needs . For its realization , strong health systems and political will are required to overcome epidemiological and political challenges .
Elimination of neglected tropical diseases ( NTDs ) has recently emerged on the global health agenda and gained prominence with the release of the global plan to combat NTDs by the World Health Organization ( WHO ) [1] . In 2012 , WHO issued a roadmap towards the elimination of 17 NTDs [2] , and stakeholders from the public and private sectors pledged to contribute to the control , elimination , and eradication of ten NTDs through the London Declaration on NTDs [3] . The second WHO report on NTDs further elaborated the roadmap [4] , and the London Declaration follow-up report showed the substantial progress that had already been achieved through the stakeholder partnership approach [5] . One of the NTDs targeted for elimination is onchocerciasis ( river blindness ) . This is a parasitic disease caused by filariae that are transmitted by blackflies . Severe itching , skin lesions , and vision impairment including blindness are its most notable symptoms . Onchocerciasis is endemic in parts of Africa , Latin America , and Yemen , but over 99% of all current cases are found in sub-Saharan Africa [6] where onchocerciasis has historically been a serious public health problem and hindered socioeconomic development in endemic areas [7] . However , many infections are asymptomatic , and vector control and community-directed treatment with ivermectin have significantly reduced morbidity . Specifically , the Onchocerciasis Control Program ( OCP ) , which was implemented in West Africa from1975 to 2002 , and the African Programme for Onchocerciasis Control ( APOC ) , which has supported onchocerciasis control activities in sub-Saharan countries since 1995 and continued the OCP’s activities where needed , have decreased the burden of disease to such an extent that it is no longer a public health problem in most endemic areas [8] . In Latin America , the Onchocerciasis Elimination Program for the Americas ( OEPA ) implemented since 1993 has brought the disease close to elimination . Colombia and Ecuador announced the elimination of onchocerciasis after WHO verification in 2013 and 2014 , respectively [9 , 10] . Treatment has also been stopped in seven foci in Guatemala and Mexico where it has been replaced by surveillance to detect possible recrudescence [11] . Regional elimination in Latin America is expected to be achievable by 2020 if the regular treatment of a sufficient proportion of the nomadic Yanomami in the border area between Brazil and Venezuela can be achieved [12] . In Yemen , onchocerciasis is endemic in a limited number of communities . Elimination in the near future is considered technically feasible , and a national action plan aiming at elimination by 2015 was developed in 2010 [13] . Currently , political instability and security concerns that limit access to endemic areas hamper its implementation [4] . Studies in Mali and Senegal have proved the feasibility of onchocerciasis elimination through ivermectin treatment in some hyper-endemic foci in West Africa [14 , 15] . This has provided additional momentum and arguments for a shift in the strategic goal from control to elimination also in Africa . The decision to invest in elimination and eradication efforts should be informed by broad assessments considering biological and technical feasibility , financial and economic costs , health and economic gains , capacity of and impacts on health systems , and societal and political willingness to cooperate [16] . An approach to such an assessment has been proposed in the form of eradication investment cases in 2010 [17] . Tediosi and colleagues have examined the approach with focus on three NTDs including onchocerciasis [18] . With reference to this approach , we have developed and compared alternative scenarios , namely , staying in a control mode versus moving toward elimination and subsequent eradication . In the present paper , we describe the scenarios to achieve control , elimination , and eradication of onchocerciasis , predict the timeline of stopping treatment at country level , and estimate the number of required ivermectin treatments over the next 30 years with focus on Africa .
We developed scenarios , describing all required activities and resources that are expected to lead to the goals of control , elimination , and eradication , if effectively implemented and sustained as long as required , based on current standard practice , the results of large-scale studies , and available historical data . To clearly distinguish these alternative scenarios , we referred to the definitions of control , elimination , and eradication endorsed and recommended by the WHO Strategic and Technical Advisory Group for NTDs [19] . The ultimate goals of the scenarios were defined as follows: 1 ) control scenario: continuing community-directed treatment with ivermectin ( CDTi ) to keep the prevalence under a locally acceptable level; 2 ) elimination scenario: scaling up CDTi to all endemic areas where feasible aiming at the reduction of disease incidence to zero; and 3 ) eradication scenario: including strategies and tailored interventions to overcome operational challenges in endemic areas with feasibility concerns in addition to CDTi with the aim of reducing the global disease incidence to zero ( Table 1 ) . From an operational perspective , the control and elimination scenarios are designed to target endemic areas where interventions appear feasible without major challenges , whereas the eradication scenario is an optimal situation . To make the eradication scenario feasible , intensive efforts to improve operational capacity and to increase political willingness would be required to overcome epidemiological and political challenges . We assume effective treatment would be implemented through tailored approaches in those areas , and regular surveillance would be maintained during and after the intervention phase until eradication has been verified . Referring to the general principles for developing scenarios outlined by Tediosi and colleagues [18] , the key components of scenarios were identified at project level . Scenarios were further revised by verifying the realism of assumptions in consultation with a technical advisory group consisting of policymakers , onchocerciasis epidemiologists , public health experts , health economists , and donors . Key components for developing the scenarios are defined as follows and the developed scenarios are described in Table 1 . The number of required ivermectin treatments to achieve the goals of the control , elimination and eradication scenarios in endemic African regions was predicted by multiplying the estimated population living in endemic areas with the treatment coverage rate and the CDTi frequency per year for the required duration of treatment at project level . The capacity of drug manufacturers to supply the required number of ivermectin was assumed to be sufficient considering Merck’s commitment to donate ivermectin until elimination is achieved globally [41] . The time horizon for predicting the number of treatments was 2013 to 2045 . The start year was set considering the most recent version of the APOC databases available for analysis was for 2012 . The end year was chosen based on the prediction that the last project in the eradication scenario would stop CDTi in 2040 , and that after stopping CDTi , at least three years would be required to confirm local elimination . In the control and elimination scenarios , the last projects were expected to continue CDTi beyond 2045 . In the S1 Table , the relevant data regarding the key components of the scenarios , which were used for estimating the timelines and the number of required ivermectin treatments , are presented at project level . Parameters used for the scenario analysis were subject to considerable uncertainty and the impact of the uncertainty was examined for the target population , the timeline when CDTi is expected to be stopped , and the number of required ivermectin treatments . The impact of a single parameter’s uncertainty was assessed with one-way deterministic sensitivity analysis ( DSA ) . Considering the final estimates are driven by the joint effects of multiple parameters , multivariate probabilistic sensitivity analysis ( PSA ) was conducted with all the variables examined in the one-way DSA . The included parameters were population growth rate , treatment coverage , treatment duration , CDTi start and end years , and the assumptions for selecting target projects . For DSA , the parameter uncertainty ranges were determined based on available data , expert opinion or both . For PSA , statistical distributions were chosen considering the characteristics of parameters , and fitted to available data . Simulations were run 1 , 000 times for each scenario .
The control scenario targeted hyper-and meso-endemic areas in all endemic African countries . Under the elimination scenario , CDTi was extended to hypo-endemic areas where CDTi is feasible in addition to hyper-and meso-endemic areas . Countries that include projects with feasibility concerns have been identified to be the Central African Republic , the Democratic Republic of the Congo , and South Sudan due to political instability , and Gabon due to the high prevalence of L . loa in areas with a low prevalence of onchocerciasis . In these four countries , hypo-endemicity areas were therefore excluded from the elimination scenario . The eradication scenario targeted all hyper- , meso- , and hypo-endemic areas . The endemic countries in Africa were categorized into two control programs in which they participate or participated , APOC and OCP , respectively ( Table 2 ) . The control scenario included 27 countries , and potential new projects were predicted to cover around 3% of the total population in the entire target area , or 4 . 7 million of 144 million ( Fig 1 ) . The elimination scenario included the same 27 countries , and new projects were predicted to cover at most 17% of the population in the entire target area ( 167 million ) . Depending on the number of new projects in potential hypo-endemic areas , the population in new project areas ranged from 12 . 1 million to 27 . 8 million ( 7% and 17% ) . The eradication scenario included one more country , Gabon , and the total population in the entire target area was estimated at around 176 million of which 21% at maximum live in new project areas with a range of 12 . 1 million to 36 . 5 million people ( 7% to 21% ) depending on the number of new projects in potential hypo-endemic areas . In the control scenario , most endemic countries outside West Africa were predicted to continue CDTi beyond 2045 ( Fig 2 ) . The most influential parameter determining the expected year of ending CDTi was the extension of treatment duration due to insufficient treatment coverage ( Fig 3 ) . For the elimination and eradication scenarios , the final year of CDTi represents the year of ending the intervention phase at country level assuming no recrudescence would occur . In the elimination scenario , all endemic countries except the four countries with feasibility concerns were expected to finish the intervention phase by 2028 at the latest and those four countries were expected to continue CDTi beyond 2045 ( Fig 2 ) . In the eradication scenario , all endemic countries were expected to reach the end of the intervention phase by 2040 assuming sufficient treatment would be delivered sustainably in the four countries with epidemiological and political concerns . For the elimination and eradication scenarios , one-way DSA ( Fig 3 ) showed that any delay in starting and ending CDTi and low treatment coverage would result in the intervention phase to end later than expected; on the contrary , high treatment coverage would expedite the progress of the intervention phase and lead to an earlier end of the intervention phase . The need for ivermectin treatments was concentrated in the first half of the time horizon for the elimination and eradication scenarios , as 80% of all potential projects were stopped safely by 2031 and 2025 , respectively . In the control scenario , it took until 2038 for the same proportion of the total projects to stop CDTi ( Fig 4 ) . The cumulative number of required ivermectin treatments over 2013–2045 was estimated at 2 . 63 billion ( 95% central range: 2 . 41 billion-2 . 99 billion ) for the control scenario . Specifically , 1 . 48 billion ( 1 . 51bn-1 . 57bn ) treatments were predicted to be required until 2025 and 1 . 15 billion ( 0 . 90bn-1 . 41bn ) treatments over 2026–2045 ( Table 3 ) . According to the simulation of the elimination scenario , the required number of ivermectin treatments over the whole period was around 1 . 48 billion ( 1 . 42bn-1 . 79bn ) . Compared to the control scenario , the total number of required treatments in the elimination scenario was lower by 1 . 15 billion ( 44% ) : 0 . 45 billion ( 0 . 36bn-0 . 55bn ) until 2025 and 0 . 69 billion ( 0 . 38bn-0 . 92bn ) from 2026 to 2045 ( Table 3 , Fig 5 ) . The eradication scenario required an even smaller number of ivermectin treatments for the whole period , 1 . 30 billion ( 1 . 18bn-1 . 51bn ) , which was 0 . 18 billion ( 0 . 03bn-0 . 49bn ) , or 12% , lower than that under the elimination scenario and 1 . 32 billion ( 0 . 97bn-1 . 75bn ) , or 50% , lower than that under the control scenario ( Fig 5 ) . In one-way DSA ( Fig 6 ) , the most influential parameter on the cumulative number of required ivermectin treatments was the delay in ending CDTi in all scenarios . For the control scenario , the second most influential parameter was the number of projects with extended CDTi duration due to insufficient treatment coverage . For the elimination and eradication scenarios , it was the number of potential new projects in hypo-endemic areas .
The key changes for shifting from the control mode to elimination and subsequent eradication are the scale-up of CDTi to hypo-endemic areas and the implementation of regular epidemiological and entomological surveys along with ongoing surveillance . For successful implementation of these , overcoming the existing feasibility issues related to the co-endemicity with L . loa , the insecure political situation , and weak health systems will be critical . We found that , if this could be accomplished , regional elimination in Africa could be achieved as early as 2040 , and consequently all endemic countries including Latin Americas and Yemen would be in the post-elimination phase until eradication has been verified . We found that achieving elimination would reduce treatment needs by 43% compared to the control mode for the period 2013–2045 . The driver of this remarkable difference is that CDTi could be stopped for the majority of projects based on regular surveillance , while it would have to continue for at least 25 years under the control scenario . The eradication scenario is predicted to require an even smaller number of ivermectin treatments than the elimination scenario , as hypo-endemic areas with feasibility concerns were assumed to have a shorter treatment period through effective treatment via tailored approaches as well as CDTi , whereas those areas would be under the control mode in the elimination scenario . This finding implies that saved ivermectin drugs could be used for other disease programs , for instance , mass drug administration ( MDA ) for lymphatic filariasis ( LF ) . The uncertainty about the target population in the elimination and eradication scenarios was mainly driven by uncertainty in the number of potential new projects in hypo-endemic areas , as some of those areas might not be actually endemic . Parasitological surveys are therefore needed to determine the current infection status of those areas . Setting up a new project requires operational planning , human resource mobilization , and startup costs . To move towards elimination without delay and to save human and financial resources , the rapid mapping of potential hypo-endemic areas should be a priority to confirm areas to set up new projects and to develop elimination strategies for those areas . The main driver of the number of required ivermectin treatments was the delay in stopping CDTi . This finding implies that maintaining high treatment coverage to avoid the extension of treatment duration and continuous monitoring and evaluation to decide a proper time to stop CDTi would lead to faster elimination and prevent unnecessary efforts to deliver drugs . We assumed no recrudescence in our analysis . However , if recrudescence occurs , the duration of CDTi would need to be extended , local elimination would be delayed , and the number of required treatments would increase . Recrudescence might occur because of human or vector migration , interrupted drug distribution due to political instability , and residual transmission from not-treated endemic areas due to incomplete or inconsistent geographic coverage . We did not adjust for alternative treatment approaches for areas where L . loa is highly endemic but onchocerciasis is hypo-endemic . Suggested treatment approaches for these areas include anti-Wolbachia therapy with macrofilaricidal drugs , high doses of albendazole , and the test-and-treat strategy [43 , 44] . These approaches would expedite elimination and increase the demand for other drugs while reducing the need for ivermectin . Our modeling did not incorporate the impact of changing the CDTi frequency on the treatment duration . It has been suggested to increase the frequency of CDTi to reduce the prevalence and transmission of onchocerciasis faster compared to the annual CDTi [35] . A recent study by Coffeng and colleagues shows that six-monthly ivermectin treatment could reduce the required treatment duration by 40% based on a dynamical transmission model [38] . In practice , increasing the CDTi frequency would require collaboration between policymakers , health workers , and community volunteers and new strategies on how to mobilize human and financial resources , given limited resources and competing health programs . Under the control scenario , annual CDTi could mean overtreatment for projects that had more than 15–20 years of treatment , for example , some areas in West Africa where ivermectin administration has been implemented since the 1990s . For these areas , less frequent CDTi could be an alternative for morbidity control , which would require a smaller number of ivermectin and less human and financial resources . However , less frequent CDTi might lead to a loss of local expertise , human resources , and community compliance over the time interval without CDTi and , consequently , to the decrease of treatment coverage below the required level , which could expose the areas to the risk of recrudescence . We did not incorporate possible delays in ending CDTi due to co-endemicity with LF . In areas where LF is co-endemic with onchocerciasis , an assessment whether both diseases have reached the thresholds to stop treatment will be needed in order to stop CDTi . In practice , no delay is expected in most cases as MDA for LF , which relies on albendazole and ivermectin , usually requires fewer cycles to reach the point of transition to the post-treatment phase . However , LF mapping or anti-LF MDA have not started in about a third of the 35 endemic countries in Africa [45] . We did not take into account the possibility of drug resistance , as no confirmed cases of ivermectin resistance have been reported from endemic countries so far . However , if ivermectin resistance were to happen as suggested by Bourguinat and colleagues through studies on the effects of ivermectin on the genetics of Onchocerca volvulus [46] , the entire efforts for onchocerciasis treatment could be endangered , as current strategies heavily rely on ivermectin . The long time horizon of 2013–2045 poses challenges in predicting technological , political , and economic changes . New treatment and diagnostic tools could be game changers in achieving elimination . Ivermectin is a microfilaricidal drug which requires many years of treatment and has a risk of eliciting severe adverse reactions in L . loa patients . Macrofilaricidal drugs that are safe and effective for general population use , are easy to administer in communities , and have a shorter treatment period than ivermectin could substantially change treatment strategies and expedite elimination . Several macrofilaricidal drugs for human use have been or currently are under development , e . g . , doxycycline [47] , emodepside [48] , moxidectin [49] , and flubendazole [50] . The need for diagnostic techniques that are capable of detecting infections early , are easy to use in the field , and are affordable would greatly facilitate surveillance when early detection of new infections is paramount . The skin snip method , currently the most common diagnostic method , has low sensitivity for detecting very light infections , and can result in a delay in detecting recrudescence . Several diagnostic techniques , e . g . , OV-16 ( ELISA and Rapid Test ) and the DEC patch test [51 , 52] , that may prove more sensitive and practical , have been developed . Unexpected political unrest might hamper the elimination programs , as it interrupts interventions and weakens political support . Industrialization along with economic growth may have a significant impact . For instance , the construction of dams can flood existing breeding sites of blackflies or create new ones , and deforestation can greatly alter the composition or density of blackfly populations . Political will across the whole spectrum of stakeholders from global and national policymakers to community members will be particularly critical during the “last mile” towards elimination and subsequent eradication [53] . Countries sharing borders spanning endemic areas would need to effectively collaborate to enable prompt responses to or prevent possible recrudescence . Regular meetings have been held between Guinea/Sierra Leone/Liberia , Togo/Benin , and Benin/Nigeria [54] , and this proves such mechanism can work . Similar collaborative relationships would need to be fostered for other endemic countries . APOC has announced that it would transform to a new regional entity by 2016 that would support integrated country-driven programs to eliminate onchocerciasis , LF , and other preventive chemotherapy NTDs ( soil-transmitted helminthiasis , schistosomiasis , trachoma ) in Africa [55 , 56] . Successful launching of this new regional entity might provide a more collaborative environment for sustainable interventions and post-treatment surveillance for NTDs in the region . Continuous support from community members is essential for onchocerciasis elimination in Africa . National policymakers would need to keep empowering community drug distributors , as their role is critical for successful CDTi and will continue to be so until eradication has been achieved .
|
River blindness ( onchocerciasis ) is transmitted by blackflies and causes severe itching , skin lesions , and vision impairment including blindness . More than 99% of all current cases are found in sub-Saharan Africa where the disease has historically hindered socioeconomic development in endemic areas . The treatment goal is shifting from control to elimination in Africa as morbidity has significantly decreased through vector control and community-directed treatment with ivermectin . Studies in Mali and Senegal proved that elimination is feasible with ivermectin administration . Given limited resources , national and global policymakers need a rigorous analysis of investment options from epidemiological , economic , and societal aspects . For this , we developed control , elimination , and eradication scenarios and compared treatment timelines and drug needs over the next 30 years . We found that the elimination and eradication scenarios would require a shorter treatment phase and a smaller amount of ivermectin than the control scenario , mainly because community-directed treatment with ivermectin could be ended earlier thanks to regular active surveillance .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Control, Elimination, and Eradication of River Blindness: Scenarios, Timelines, and Ivermectin Treatment Needs in Africa
|
We introduce a computational model to simulate chromatin structure and dynamics . Starting from one-dimensional genomics and epigenomics data that are available for hundreds of cell types , this model enables de novo prediction of chromatin structures at five-kilo-base resolution . Simulated chromatin structures recapitulate known features of genome organization , including the formation of chromatin loops , topologically associating domains ( TADs ) and compartments , and are in quantitative agreement with chromosome conformation capture experiments and super-resolution microscopy measurements . Detailed characterization of the predicted structural ensemble reveals the dynamical flexibility of chromatin loops and the presence of cross-talk among neighboring TADs . Analysis of the model’s energy function uncovers distinct mechanisms for chromatin folding at various length scales and suggests a need to go beyond simple A/B compartment types to predict specific contacts between regulatory elements using polymer simulations .
The human genome contains about 2 meters of DNA that is packaged as chromatin inside a nucleus of only 10 micrometers in diameter [1] . The way in which chromatin is organized in the three-dimensional space , i . e . , the chromatin structure , has been shown to play important roles for all DNA-templated processes , including gene transcription , gene regulation , DNA replication , etc [2–4] . A detailed characterization of chromatin structure and the physical principles that lead to its establishment will thus greatly improve our understanding of these molecular processes . The importance of chromatin organization has inspired the development of a variety of experimental techniques for its characterization . For example , using a combination of nuclear proximity ligation and high-throughput sequencing , chromosome conformation capture and related methods quantify the interaction frequency in three-dimensional space between pairs of genomic loci [5 , 6] , and have revealed many conserved features of chromatin organization . A consistent picture that is emerging from these experiments is the formation of chromatin loops and topologically associating domains ( TADs ) at the intermediate scale of kilobases to megabases , and the compartmentalization of chromatin domains that are millions of base pairs apart in sequence [7–11] . Many of the findings from these cross-linking experiments are now being validated and confirmed with microscopy imaging studies that directly probe spatial contacts [12–20] . Polymer modeling has played a critical role in our understanding of the genome organization and in interpreting features of Hi-C contact maps [21] . In particular , due to its deviation from the value of an equilibrium globule [6] , the power-law exponent of the contact probability between pairs of genomic segments as a function of the genomic separation has attracted the attention of numerous research groups [22–28] . Of the many mechanisms that have been proposed , the non-equilibrium extrusion model [29–31] , which assumes that cohesin molecules function as active enzymes to inch along the DNA and fold the chromatin until encountering bound CTCF molecules , has gained wide popularity [32] . Notably , this model succeeds in explaining the flanking of CCCTC-binding factor ( CTCF ) and cohesin binding sites at the boundaries of chromatin loops and TADs [7 , 9–11 , 33] . On the other hand , phase separation , which is emerging as the key mechanism for organizing numerous membraneless organelles [34–36] , has been suggested as the driving force for chromosome compartmentalization [37–39] . Since polymer molecules that differ in chemical compositions are known not to intermix [40] , micro-phase separation can contribute to the formation and compartmentalization of chromatin domains with distinct histone modification profiles . Finally , besides these mechanism-based modeling strategies , data-driven approaches have also been quite successful in reconstructing chromosome structures directly from Hi-C data and revealing structural features of both interphase and metaphase chromosomes [41–45] . In parallel , bioinformatics studies have provided powerful tools in addressing potential biases in Hi-C data [46–48] , and offered numerous insights in our understanding of genome organization . In particular , correlating one-dimensional genomics and epigenomics data with 3D contacts has been rather informative and has led to many proposals on the molecular mechanism of chromatin folding [4 , 49–54] . Furthermore , using advanced machine learning techniques , numerous groups have developed predictive models to identify specific contacts between regulatory elements [55–58] . Though not able to construct the whole contact map and 3D chromosome structures , these machine learning approaches have achieved the level of resolution and specificity needed to study functionally important contacts within a TAD . On the other hand , it remains challenging to quantitatively study such functionally important contacts using polymer modeling approaches , though significant progress towards that direction is being made [39 , 59–63] The difficulty in predicting contacts between specific regulatory elements using polymer models is at least twofold . First , existing phase separation models based on A/B compartments or six subcompartments are inadequate for such purposes , despite their success in recapitulating the long-range block-wise patterns observed in Hi-C . As chromosome compartments are defined based on contact patterns revealed by Hi-C at a coarse resolution from 50kb to 1 Mb , they tend to group many regulatory elements together as one “active” type and fail to capture the distinction among them [6 , 7 , 47] . The ambiguity of these compartments significantly limits the accuracy of polymer models built upon them . To study enhancer-promoter interactions , one must introduce new chromatin types at a higher resolution to achieve the required specificity . How to define these types and how many types are needed remain unclear . Secondly , even with our current understanding of chromatin folding mechanisms , developing a quantitative polymer model to predict contact probability between pairs of genomic loci is still a non-trivial task . In particular , robust and efficient schemes are needed to derive parameters of polymer models to ensure their accuracy . In this paper , we report the development of a predictive and transferable polymer model to simulate the structure and dynamics of chromosomes at five kilo base resolution . This model takes combinatorial patterns of epigenetic marks and genomic location and orientation of CTCF binding sites as input , and can be parameterized from Hi-C data with a robust and efficient maximum entropy approach [64 , 65] . A key innovation of this model is its use of chromatin states to capture the wide variety of regulatory elements and to probe their interactions . Computer simulations of this model provide a high-resolution structural characterization of chromatin loops , TADs , and compartments , and succeed in quantitatively reproducing contact probabilities and power-law scaling of 3D contacts as measured in Hi-C and super-resolution imaging experiments . Many significant enhancer-promoter contacts can be captured in simulated contact maps as well . As the model incorporates ingredients from both the extrusion and the phase separation mechanism , its success in quantitative predictions of genome organization provides strong support for such mechanisms . In the meantime , detailed analysis of the model parameters further reveals a significant difference between the interactions that stabilize TAD and those that drive compartmentalization , providing additional insight into chromatin folding not appreciated in existing modeling efforts . Finally , we demonstrate that the model is transferable across chromosomes and cell types , setting the stage for de novo prediction of the structural ensemble for any given chromatin segment using only one-dimensional sequencing data that is available for hundreds of cell types .
We introduce a predictive model to study cell-type specific 3D chromatin folding . This model takes a sequence of chromatin states derived from genome-wide histone modification profiles and a list of CTCF binding sites as input . We selected these genomic features due to their known roles in organizing the chromatin at various length scales ( Fig 1A ) . At the core of this model is an energy function—a force field—that is sequence specific and ranks the stability of different chromatin conformations . Starting from the input for a given chromatin segment , we use molecular dynamics simulations to explore chromatin conformations dictated by the energy function and to predict an ensemble of high-resolution structures . These structures can be compared directly with super-resolution imaging experiments or converted into contact probability maps for validation against genome-wide chromosome conformation capture ( Hi-C ) experiments . As shown in Fig 1B , a continuous genomic segment is represented as beads on a string in this model . Each bead accounts for five-kilo bases in sequence length and is assigned with a chromatin state derived from the underlying combinatorial patterns of 12 key histone marks . Chromatin states are known to be highly correlated with Hi-C compartment types [39 , 54 , 66] and , therefore , will help model large-scale chromosome compartmentalization . In the meantime , chromatin states can go beyond traditional A/B compartments or subcompartments to provide polymer models with the specificity needed for studying interactions between regulatory elements . We define a total of 15 chromatin states , identified using a hidden Markov model [67] , to distinguish promoters , enhancers , heterochromatin , quiescent chromatin , etc ( see Methods ) . Detailed histone modification patterns for these chromatin states are shown in Fig 1C . We note that 15 is large enough to capture the diversity of epigenetic modifications while still being small enough to ensure a sufficient population of each state for a robust inference of interaction parameters between them ( Figure A1 in S1 Supporting Information ) . We further studied a hidden Markov model with 20 states , and found that further increasing the number of states does not lead to a discovery of additional epigenetic classes with significant populations ( Figure A2 in S1 Supporting Information ) . A polymer bead is further labeled as a CTCF site to mark chromatin loop boundaries if both CTCF and cohesin molecules are found to be present in the corresponding genomic region . We define the orientation of these CTCF sites by analyzing the underlying CTCF motif and the relative position of CTCF molecules with respect to cohesin . Details for the definition of CTCF binding sites are provided in Methods . The potential energy for a given chromatin configuration r is a sum of three components , and UChrom ( r ) = U ( r ) + UCS ( r ) + UCTCF ( r ) . U ( r ) is a generic polymer potential that is included to ensure the continuity of the chromatin , and to enforce excluded volume effect among genomic loci . UCS ( r ) is a key innovation of the chromatin model , and is crucial to capture the formation of TADs and compartments . It quantifies the chromatin state specific interaction energies between pairs of loci . As detailed in Section: Physical principles of chromatin organization and Methods , we used a general form for UCS ( r ) to capture its dependence on genomic separation . UCTCF ( r ) is inspired by the loop extrusion model [29–31] , and facilitates the formation of loop domains enclosed by pairs of CTCF binding sites in convergent orientation ( Fig 1A ) . Both UCS ( r ) and UCTCF ( r ) contain adjustable parameters that can be derived from Hi-C data following the optimization procedure developed by one of the authors [64 , 65] . Segments of chromosomes 1 , 10 , 19 and 21 from GM12878 cells were used for parameterization to ensure a sufficient coverage of all chromatin states ( see Figure A1 in S1 Supporting Information ) . Detailed expressions for the potential energy , and the parameterization procedure are provided in Methods and in the S1 Supporting Information . Using the parameterized energy function , we simulated the ensemble of chromatin structures and determined the corresponding contact probability map for a 20 Mb region of chromosome 1 from GM12878 cells . As shown in Fig 2A , the simulated contact map is in good agreement with the one measured by Hi-C experiments from Ref . [7] and reproduces the overall block-wise checkerboard pattern that corresponds to the compartmentalization of chromatin domains . A zoomed-in view along the diagonal of the contact map provided in Fig 2B and 2C further suggests that chromatin TADs and loops are also well reproduced . Similar comparisons for other chromosomes used in parameterizing the model are provided in Figure B in S1 Supporting Information . We note that the length 20 Mb was chosen for computational efficiency , but the model can be easily generalized to longer chromatin segments ( see Figure C in S1 Supporting Information ) . To go beyond the visual inspection and quantify the correlation between simulated ( GM-Sim ) and experimental ( GM-Exp ) contact maps , we calculated the Pearson correlation coefficient ( PCC ) between the two for chromosome 1 and found that it exceeds 0 . 96 . Importantly , this number is higher than the PCC ( 0 . 94 ) between GM-Sim and Hi-C data from IMR90 cells ( IMR-Exp ) . Breaking down the PCC at different genomic separations also supports that GM-Sim is more correlated with GM-Exp at all ranges than with IMR-Exp ( Figure D in S1 Supporting Information ) . In addition , we also determined the stratum-adjusted correlation coefficient ( SCC ) that takes into account the distance-dependence effect of contact maps by stratifying them according to the genomic distance [68] , and obtained 0 . 7 for GM-Sim/GM-Exp , and 0 . 66 for GM-Sim/IMR-Exp . Therefore , SCC analysis also validates our model’s ability in reproducing Hi-C contact maps and in capturing the distinction between cell types . We note that the magnitude of SCC can be sensitive to the smoothing parameter used in its calculation and should be interpreted with caution ( Figure E in S1 Supporting Information ) . We further examined the agreement between simulated and experimental contact maps using multiple feature-specific metrics . First , we define the contact enhancement for a pair of genomic loci as the ratio of their contact probabilities over the mean contacts averaged over a locally selected background region ( see Figure F1 in S1 Supporting Information ) . The contact enhancement for chromatin loops from chromosome 1 is always larger than one , indicating a strong enhancement of spatial colocalization between loop anchors . Furthermore , over 74% of the loop pairs exhibit a contact enhancement that is larger than the 90th percentile of the distribution for random genomic pairs . These random pairs are selected regardless of CTCF occupancy but with comparable sequence separations as those found in chromatin loops . Therefore , if we use the 90th percentile of the random distribution as a threshold ( 1 . 16 ) and predict every convergent CTCF pairs as loops , the prediction will have a false negative rate of 26% , and a false positive rate less than 10% . The false positive value is an upper bound since most of the random pairs are not flanked with convergent CTCF . The sensitivity of chromatin loop predictions on the threshold is shown in Figure F2 in S1 Supporting Information . It is worth pointing out that the contact enhancement for chromatin loops calculated using Hi-C data is in general larger than simulated values and separates better from that for random pairs ( Figure F3 in S1 Supporting Information ) . The overlap between the two distributions in our simulation is due to that random pairs include a significant fraction of convergent CTCF pairs whose contacts are enhanced as a result of the potential UCTCF ( r ) . Many of these pairs , however , are not recognized as loops in Hi-C , and more advanced algorithms than simple binding site orientations are probably needed to identify loop forming CTCF pairs [69] . To go beyond CTCF mediated contacts and evaluate our model’s ability in reproducing strong interactions between genomic loci , we selected statistically significant contact pairs from simulated and experimental contact maps for chromosome 1 using the software Fit-Hi-C [48] ( Figure G in S1 Supporting Information ) . As a quantitative metric , we define the matching score as the percent of experimental pairs that can be found in the list extracted from simulation . The reverse matching score can be similarly defined as the percent of simulated pairs found in the experimental list . The matching score for the top 1000 chromatin contacts is determined to be 46% and 52% for the reverse matching . To examine specific interactions between regulatory elements , we performed a similar analysis by selecting the top 100 enhancer ( state: EnhW1 ) -promoter ( state: PromD1 ) pairs with highest contact probabilities based on simulated and experimental contact maps . We find that over 70% of experimental pairs are captured in our simulation for chromosome 1 . These results suggest that our model based on chromatin states and CTCF mediate interactions is able to reproduce a large fraction of significant contacts detected in Hi-C experiments . Further improving the model’s ability in predicting functionally important pairs would potentially require considering the effect of other proteins , such as YY1 that are known to mediate chromatin interactions [70] , and will be an interesting future direction . We next determined the correlation coefficients between the top five eigenvectors for simulated and experimental contact matrices . As shown in Figure H in S1 Supporting Information , the contact maps reconstructed using only these eigenvectors recapitulate the formation of TADs and compartments observed in the original maps . The high correlation between simulated and experimental eigenvectors ( with PCC at approximately 0 . 8 ) supports that the corresponding features are well captured by the computational model , and confirms the qualitative observations from Fig 2 and Figure B in S1 Supporting Information . To more closely examine the quality of simulated TADs , we calculated the insulation profile by sliding a uniform 500kb × 500kb square along the diagonal of the contact matrix and averaging over all contacts within the square . The minima of this profile can be used to identify TAD boundaries as inter-TAD contacts are sparser compared to intra-TAD contacts , resulting in a drop in the insulation score profile as the sliding window crosses TAD boundaries [71] . The PCC between experimental and simulated insulation profiles for chromosome 1 is 0 . 7 . We find that the matching score for TAD boundaries is 80% and 100% for the reverse matching . As another independent validation , we determined TAD boundaries using the software TADbit [43] , and found that the simulated results again match well with experimental ones ( see Figure I in S1 Supporting Information ) . To demonstrate the transferability of the computational model across chromosomes and cell types , we performed additional simulations for chromosomes from GM12878 , K562 , and Hela cells , whose Hi-C data were not included during the parameterization procedure . As shown in Fig 3 and Figure J in S1 Supporting Information , these de novo predictions are in good agreement with experimental results as measured by PCC ( Fig 3B ) and SCC ( Fig 3C ) between experimental and simulated contact maps , matching score between TAD boundaries detected from the insulation profile ( Fig 3D ) and from TADbit ( Figure K1A in S1 Supporting Information ) , PCC between experimental and simulated insulation profiles ( Figure K1D in S1 Supporting Information ) , matching score between significant contacts detected using Fit-Hi-C ( Fig 3E ) , matching score between interacting enhancer-promoter pairs ( Figure K2C in S1 Supporting Information ) , correlation coefficients of the top five eigenvectors ( Fig 3F and Figure H in S1 Supporting Information ) , and false negative rate of loop predictions ( Fig 3F ) . Furthermore , the model succeeds in revealing the cell-type specificity of Hi-C contact maps , and the simulated contact maps are always more correlated with the corresponding experimental data from the same cell type than with those from IMR90 cells ( light colors in Fig 3B and 3C ) . The matching scores between experimental and simulation results are also significantly higher than those calculated between experimental and control data ( light colors in Fig 3D and 3E ) , which were obtained by randomly shuffling the size of loops/enhancer-promoter pairs/TADs along the chromosome while keeping their total number unchanged . The success of these de novo predictions supports that the chromatin-state-based model introduced here provides a consistent description of the 3D genome organization across cell types . We next analyze the simulated 3D structural ensembles to gain additional insights on chromatin organization . Consistent with previous experimental and theoretical studies [37 , 72 , 73] , our model reproduces the clustering of active chromatin state and their preferred location at the exterior of chromosomes ( Figure L in S1 Supporting Information ) . Super-resolution imaging experiments probe chromatin organization in 3D space to quantify spatial distances between genomic segments . These 3D measurements can be compared directly with simulated chromatin structures , and thus provide a crucial validation of the computational model parameterized from Hi-C experiments with independent datasets . To understand the overall compactness of various chromatin types , we selected a set of active , repressive and inactive chromatins and determined their radiuses of gyration from the ensemble of simulated structures . These different chromatin types are identified using two key histone marks H3K4me2 and H3K27me3 ( Fig 4A ) . The complete list of chromatin domains with their genomic locations is provided in the Extended Data Sheet . As shown in Fig 4B , the radius of gyration increases at larger genomic separation following a power law behavior in all cases with exponents of 0 . 34 , 0 . 31 and 0 . 23 for the three chromatin types respectively . These scaling exponents are in quantitative agreement with imaging measurements performed for Drosophila chromosomes [12] and support the notion that active chromatins adopt less condensed conformations to promote gene activity . Consistent with the imaging study performed on chromosome 21 from IMR90 cells [13 , 20] , we also observe a strong correlation between Hi-C contact probabilities and spatial distances for pairs of genomic loci ( Fig 4C ) . One of the most striking features revealed by high-resolution Hi-C experiments is the formation of chromatin loops anchored at pairs of convergent CTCF sites [7 , 10 , 74 , 75] . Microscopy studies that directly visualizes 3D distances using fluorescence in situ hybridization ( FISH ) methods further find that these loops are dynamic , and despite their high contact frequencies , loop anchors are not in close contact in every cell [16 , 41 , 76] . Consistent with their dynamic nature , chromatin loops in our simulation adopt flexible conformations as well . As shown in Fig 5A , for the loop formed between chr1:39 . 56–39 . 73 Mb , we observe a large variance in the probability distribution of its end-to-end distances . Additional results for other loop pairs are provided in Figure M in S1 Supporting Information . Two example configurations of the loop domain with distance at 0 . 08 and 0 . 24 μm are shown in the inset . A systematic characterization of all the loops identified in Ref . [7] for the simulated chromatin segment shows that the conformational flexibility is indeed general , though there is a trend in decreasing variance for loops with larger contact probabilities ( Fig 5B ) . We also emphasize that though higher contact probabilities , in general , corresponds to smaller end-to-end distances , their relationship is not strictly monotonic . The opposite correlation can be seen in numerous cases in Fig 5B . Such seemingly paradoxical observations have indeed been found in previous experimental studies that compare 3C with FISH experiment [16 , 77] , and can naturally arise as a result of dynamical looping or loop extrusion [78] . Compared to chromatin loops , TADs are longer and are stabilized by a complex set of interactions [79] . The analysis of their structural ensemble is less straightforward , and the end-to-end distance may not be sufficient for a faithful description of their conformational fluctuation [80] . It is desirable to analyze TAD structures using reaction coordinates that not only help to distinguish different clusters of chromatin conformations , but can also provide insight into the mechanism of TAD folding and formation . Borrowing ideas from protein folding studies , we approximate these reaction coordinates using collective variables with slowest relaxation timescales as determined following the diffusion map analysis [81 , 82] . Progression along these variables approximates well the most likely transition between two sets of structures and can , therefore , shed light on the pathway for conformational rearrangements . Diffusion map analysis has been successfully applied to a variety of systems to provide mechanistic insights on the conformational dynamics involved in protein folding , ligand diffusion , etc . [83 , 84] . We applied the diffusion map technique to the predicted structural ensemble of the genomic region chr1:34–38 Mb from GM12878 cells that consists of three visible TADs . As shown in Fig 6 , several basins are observed in the probability distribution of chromatin conformations projected onto the first two reaction coordinates , suggesting the presence of multiple stable TAD structures , rather than a unique one . Conformational heterogeneity in TADs has indeed been observed in a recent super-resolution imaging study that characterizes single cell chromatin structures [20] . To gain physical intuition on the reaction coordinates and insight on the transition between TAD structures , we calculated the corresponding contact maps at various values of these coordinates . As shown in the top panel , reaction coordinate one captures the formation of contacts between TAD1 and TAD3 while the structures for all three TADs remain relatively intact . On the other hand , progression along reaction coordinate two ( left panel ) leads to significant overlaps between TAD1 and TAD2 . Interaction between TAD2 and TAD3 can also be observed along a third coordinate as shown in Figure N in S1 Supporting Information . Example structures for the three TADs in various regions are also provided on the right panel . These results are consistent with the notion that TADs are stable structural units for genome organization [79] , but also suggest the presence of significant cross-talk among neighboring TADs [85] . Though the exact molecular mechanism and driving force for chromatin folding remain elusive , it is becoming increasingly clear that different molecular players are involved in organizing the chromatin at various length scales [49 , 60 , 86 , 87] . For example , transcription factors and architectural proteins are critical in stabilizing the formation of chromatin loops and TADs [4 , 33 , 79] . On the other hand , nuclear compartments , such as the nucleolus and the nuclear envelope , contribute to chromatin compartmentalization and mediate contacts among chromatin domains separated by tens of Mb in sequence [50 , 88] . We expect that these different molecular mechanisms will give rise to distinct interaction energies at various genomic length scales . For example , for the same pair of chromatin states , as the genomic separation between them is varied , the interaction energy that stabilizes their contact should vary . In the following , we examine the dependence of inferred contact energies on genomic separation to reveal the principles of genome organization . Fig 7A presents the derived contact energies among chromatin states UCS ( r ) at various genomic separations ( 500kb , 1 . 5 Mb , 4 Mb and 10 Mb from left to right ) , with blue and red for attractive and repulsive interactions respectively . A notable feature for all four length scales is the clear partition of chromatin states into at least two groups that correspond to well-known active and repressive chromatins respectively . For example , attractive interactions are observed among the top half chromatin states that include promoters ( PromD1 , PromU ) , enhancers ( TxEnh5 , Enhw1 ) and gene body ( Tx ) , and for the bottom half that includes inactive chromatin ( Quies ) , polycomb repressed domain ( ReprPC ) and heterochromatin ( Het ) . The unfavorable interactions among active and repressive chromatins will drive their phase separation shown in Fig 2D and Figure L in S1 Supporting Information . Partitioning of chromatin states into active and inactive groups is also evident from the dendrogram shown in Fig 7B , and the eigenvectors for the largest in magnitude eigenvalue of the interaction matrices shown in Fig 7C . Despite their overall similarities , the interaction energies at various genomic separations differ from each other . To quantify their differences , we determined the pairwise Pearson correlation coefficients between the interaction matrices . As shown in Fig 7C , the interactions that are responsible for TAD formation ( ~ 1 Mb ) indeed differ significantly from those that lead to chromatin compartmentalization ( ~ 10 Mb ) , as evidenced by the low correlation among them . Strikingly , the correlation coefficient between interaction matrices at 4 Mb and 10 Mb exceeds 0 . 9 , indicating the convergence of chromatin interactions at large genomic separation . We further compared the complexity of the interaction matrices by calculating the ratio of the first n eigenvalues over the sum of all eigenvalues . Fig 7D plots this complexity measure as a function of n , and absolute values of the eigenvalues were used to calculate the measure . For all three matrices with genomic separation larger than 1 Mb , we find the top first six eigenvectors can explain a large fraction of their complexity ( over 80% ) . This observation is consistent with the success of our previous effort in modeling chromatin organization with six compartment types [37] . However , more eigenvectors are needed , especially for short range in sequence interactions , to capture the full matrix complexity . These results together highlight the presence of distinct mechanisms that fold the chromatin at various genomic separations , and argues the importance of using sequence length dependent contact energies .
We introduced a novel computational model for studying 3D genome organization by integrating bioinformatics analysis with polymer modeling . This integration brings together the best of both worlds and results in a powerful predictive tool . Similar to bioinformatics approaches , our model succeeds in identifying cell-type specific interactions between regulatory elements . As in polymer modeling , the availability of 3D chromosome conformations makes it possible to characterize contacts between any genomic segments and construct the whole contact map , to study global properties of the genome organization that involve many-body interactions , and to explore the physical mechanism and driving force of genome folding . This predictive model presents a significant improvement from our previous effort in simulating chromatin structures [37] by switching the input from compartment types to chromatin states . In particular , unlike compartment types that are results from clustering Hi-C contact matrices [7] , chromatin states are defined as combinational patterns of histone modification profiles . Uncoupling the input from Hi-C data is critical to ensure that the model is genuinely predictive . Furthermore , chromatin states allow us to model chromatin structures at a much higher resolution ( 5kb ) to provide a detailed structural characterization of chromatin loops and TADs , and to resolve long-range specific contacts between promoters and enhancers . On the other hand , chromatin models based on compartment types are inherently limited to 50kb [37 , 39] , a resolution at which compartment types can be robustly derived from Hi-C data [7] . Finally , as shown in Fig 7 , the novel sequence-separation dependent contact potential developed here enables a rigorous assessment of the number of “types” needed for modeling chromatin structures , and suggests that the six compartment types are insufficient for an accurate description of TAD formation . Since the data required to define chromatin states are available for hundreds of cell types via the epigenome roadmap project [89] , we anticipate a straightforward application of the model developed here to characterize the differences of chromatin structures across cell types and to understand the role of 3D genome organization in cell differentiation and cell fate establishment . Histone modifications have long been recognized as crucial for the genome’s function [90] . The “histone code” hypothesis was proposed to rationalize the presence of numerous types of histone marks and the importance of their combinatorial roles [91] . However , a mechanistic understanding of the relationship between these chemical modifications and the functional outcome remains lacking [92] . The success of the computational model introduced here in predicting chromatin structures argues for the importance of histone modifications in organizing the genome . It is tantalizing to hypothesize that the histone code can be understood from a structural perspective . Epigenome engineering experiments that perturb histone modifications at specific genomic locations will be helpful to elucidate further whether the relationship between 1D histone modifications and 3D genome organization is causal .
The energy function of the chromosome model , which can be rigorously derived following the maximum entropy principle [64 , 65] , adopts the following form UChrom ( r ) =U ( r ) +∑I , J∑i∈I∑j∈JαIJ ( |j-i| ) f ( rij ) +∑K , L∑K≤k<l≤L[αCh , Ch+αC , Ch+αC , C ) f ( rkl] . U ( r ) defines the generic topology of the chromosome as a confined polymer with excluded volume effect . The second term incorporates the sequence length dependent contact energies αIJ ( |j − i| ) between pairs of loci i , j characterized with chromatin states I , J respectively . As discussed in the main text , the dependence of contact energies on sequence length separation is crucial to reproduce the hierarchical genome organization , and to detect independent mechanisms of chromatin folding at different length scales . f ( rij ) measures the contact probability between a pair of loci i and j separated by a distance rij , and is defined as follows f ( r ) ={12[1+tanh ( σ ( rc-r ) ) ] , ifr≤rc12 ( rcr ) 4 , forr>rc# Where rc = 1 . 76 and σ = 3 . 72 . As shown in Figure O in S1 Supporting Information , compared to a simple hyperbolic tangent function used in previous studies [64 , 65] , the new expression decays to zero for large distances r at a slower rate . This new form is motivated by the power law relationship between spatial distances and Hi-C contact probabilities observed in Ref . [13] . Finally , the last term , inspired by the recently proposed extrusion model [29–31] , is included to model the formation of chromatin loops . In particular , the genomic segment enclosed by a pair of convergent CTCF binding sites experiences a condensing potential due to the binding of cohesin molecules . We limit this potential to convergent CTCF pairs that are separated by no more than 4 CTCF binding sites with 5’– 3’ orientation or 4 CTCF binding sites with 3’– 5’ orientation to mimic the finite processivity of cohesin molecules [30] . For generality , three different potentials are used for CTCF-CTCF interaction ( αC , C ) , CTCF-chromatin interaction ( αC , Ch ) and chromatin-chromatin interaction ( αCh , Ch ) . The explicit mathematical expression for UChrom ( r ) is provided in the SI . It contains a total of 1883 parameters . This seemingly large number is a result of our use of chromatin states and the dependence of their interaction energies , αIJ ( |j − i| ) , on genomic separation . Both of these two features are innovations of our model to predict specific contacts between enhancers and promoters , and to capture the different biological mechanisms for TAD formation and chromosome compartmentalization . We emphasize that since a specific experimental constraint can be defined for each one of these parameters , their values can be derived robustly and efficiently using the iterative maximum entropy algorithm introduced by Zhang and Wolynes [64] . As proven before , the value of these parameters in principle is unique [76] . Numerical values of the parameters are provided in the Extended Data Sheet . After a careful analysis of the interaction energies shown in Fig 7 , however , we believe that the number of parameters could potentially be significantly reduced without sacrificing the model accuracy . In particular , the number of chromatin states used here is probably “too many” since the complexity of the interaction energy matrices can be well explained with the top 10 eigenvectors . Furthermore , the interaction energies also converge at larger genomic separation , making its dependence on |j − i| unnecessary . These insights will prove useful for future chromatin modeling efforts . We carried out constant temperature simulations to predict chromatin structures consistent with the energy function UChrom ( r ) using the molecular dynamics software package LAMMPS [93] . For each contact map presented in the manuscript , a total of eight independent 40-million-timestep long simulations were performed to ensure sufficient statistics . On an Intel Xeon E5-2690 v4 2 . 6GHz node with 14 cores , each one of such simulations takes approximately 30 hours to finish . More details on the simulation are provided in the supporting information . To enable a quantitative comparison between simulated chromatin structures with microscopy imaging data , we estimate a 5kb long genomic segment with a width of 30 nm and a length of 60 nm based on a high-resolution chromatin structure characterized by cryogenic electron microscopy ( Cryo-EM ) technique [94] . Experimental contact maps at 5kb resolution from Ref . were downloaded using the Gene Expression Omnibus ( GEO ) accession number GSE63525 ( see Extended Data Sheet ) . We used the combined contact matrices constructed from all read pairs that map to the genome with a MAPQ> = 30 . The raw matrices were then normalized with the KR method using the normalization vector provided in the same dataset . To convert the contact matrices into probabilities , we further divided each matrix element with the diagonal value Cii = 1035 obtained from averaging over all chromosomes . With this probability conversion , all the genomic segments that are within in 5kb along the sequence will on average have a contact probability of 1 . Since in the computational model , a 5kb segment has a diameter of σ = 30 nm , this probability conversion is equivalent of specifying the contact probability as 1 for genomic loci that are within a spatial distance of 30 nm . Such a probability definition is indeed consistent with the contact function f ( r ) defined in Eq . [3] and plotted in Figure O in S1 Supporting Information . A key input of the computational model is the sequence of chromatin states that captures the variation of epigenetic modifications along the genome sequence . Following Ref . [67] , we defined chromatin states as the set of unique combinatorial patterns of histone marks . Using a multivariate hidden Markov model that maximizes the posterior probability of assigning a hidden state to each genomic segment given the sequence of observed histone modifications [95] , we derived 15 chromatin states from genome-wide profiles of 12 key histone marks collected from six cell types that include GM12878 , K562 , HeLa , H1hesc , Huvec and Hepg2 . A single set of chromatin states is crucial to ensure the transferability of the parameterized force field across cell types . The dataset used for chromatin state inference is listed in the Extended Data Sheet . Detailed histone modification patterns for these chromatin states are shown in Fig 1C . With the set of chromatin states specified , every five-kilo-base long segment can then be assigned to a chromatin state based on its histone modification profiles , and a sequence of chromatin states for the entire chromatin segment can be defined as the simulation input . To capture the formation of chromatin loops , we compiled a list of CTCF-binding sites along the chromatin of interest using cell-type specific ChIP-Seq data . Starting from the peak profile downloaded from ENCODE ( see Extended data sheet ) , we identified the center of binding for each peak of both CTCF and cohesin subunit Rad21 . As both CTCF and cohesin molecules are found at the boundaries of most chromatin loops , we selected loop forming CTCF binding sites as those that have at least one Rad21 molecule located within 50bp of their genomic locations . We then determined the orientation of each CTCF-binding site as follows . We first attempted to align the binding sites to the set of CTCF motifs compiled in Refs . [7] and [96] ( see Extended data sheet ) . If the alignment succeeds and a motif is found within 100bp of the binding site , the orientation of the binding site was then assigned based on the DNA sequence of that motif . If no motif can be aligned , the orientation of the CTCF-binding site is determined using the genomic location of its binding center relative to that of the nearest binding center of Rad21 . For example , we assign the orientation as 5’– 3’ if the nearest Rad21 binding center is in the downstream of the CTCF binding site; otherwise , the orientation is assigned as 3’– 5’ . The above procedure will result in a list of oriented CTCF sites at single base resolution . From this list , we defined a 5kb-long bead in the computational model as a CTCF site if there is at least one CTCF binding site falls into the genomic region enclosed by that bead . If all the CTCF sites within the 5kb region have the 5’– 3’ orientation , then the bead is assigned with the 5’– 3’ orientation; similarly , if all the CTCF sites within the 5kb region have the 3’– 5’ orientation , then the bead is assigned with the 3’– 5’ orientation . If CTCF sites with both orientations are present , then the bead is assigned with dual orientation as well . For molecular systems that exhibit a separation of timescales , it is often desirable to approximate their dynamics at long time limit with a handful of slow variables . The time evolution of these slow variables should be Markovian and independent of the fine details of the high dimensional system to capture the dynamical behavior of the system on a coarsened timescale . Mathematically it has been proven that an optimal choice of these slow variables is the first few eigenfunctions of the backward Fokker–Planck diffusion operator [81] . Diffusion map is a data-driven approach that approximates these eigenfunctions and therefore the slow variables by defining a random walk process on the simulation data [97] . In particular , for N chromatin configurations selected from the simulated structural ensemble , we first constructed a transition probability matrix K for the random walk by defining its elements as Kij=exp ( -dij2ϵiϵj ) . The eigenfunctions of the above transition matrix can be shown to converge to that of the Fokker–Planck operator in large N limit . The distance between two configurations dij was calculated as the mean difference of their corresponding contact probability maps . We followed the algorithm proposed in Ref . [82] to normalize the matrix and to estimate ϵi . From the normalized transition matrix , we then determined its eigenfunctions and used the top two with the smallest non-zero eigenvalues as the reaction coordinates shown in Fig 6 ( see Figure N in S1 Supporting Information for eigenvalues ) .
|
Three-dimensional genome organization is expected to play crucial roles in regulating gene expression and establishing cell fate , and has inspired the development of numerous innovative experimental techniques for its characterization . Though significant progress has been made , it remains challenging to construct chromosome structures at high resolution . Following the maximum entropy approach pioneered by Zhang and Wolynes , we developed a predictive model and parameterized a force field to study chromatin structure and dynamics using genome-wide chromosome conformation capture data ( Hi-C ) . Starting from one-dimensional sequence information that includes histone modification profiles and CTCF binding sites , this model predicts chromosome structure at a 5kb resolution , thus establishing a sequence-structure relationship for the genome . A significant advantage of this model over comparable approaches is its ability to study long-range specific contacts between promoters and enhancers , in addition to building high-resolution structures for loops , TADs and compartments . Furthermore , the model is shown to be transferable across chromosomes and cell types , thus opens up the opportunity to carry out de novo prediction of genome organization for hundreds of cell types with available epigenomics but not Hi-C data .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"chromosome",
"structure",
"and",
"function",
"histone",
"modification",
"chromosome",
"mapping",
"mathematics",
"materials",
"science",
"algebra",
"epigenetics",
"molecular",
"biology",
"techniques",
"macromolecules",
"structural",
"genomics",
"chromatin",
"research",
"and",
"analysis",
"methods",
"polymers",
"polymer",
"chemistry",
"gene",
"mapping",
"chromosome",
"biology",
"gene",
"expression",
"chemistry",
"chromatin",
"modification",
"molecular",
"biology",
"eigenvectors",
"cell",
"biology",
"linear",
"algebra",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"materials",
"chromosomes"
] |
2019
|
Predicting three-dimensional genome organization with chromatin states
|
Gaining insights into genetic predisposition to age-related diseases and lifespan is a challenging task complicated by the elusive role of evolution in these phenotypes . To gain more insights , we combined methods of genome-wide and candidate-gene studies . Genome-wide scan in the Atherosclerosis Risk in Communities ( ARIC ) Study ( N = 9 , 573 ) was used to pre-select promising loci . Candidate-gene methods were used to comprehensively analyze associations of novel uncommon variants in Caucasians ( minor allele frequency~2 . 5% ) located in band 2q22 . 3 with risks of coronary heart disease ( CHD ) , heart failure ( HF ) , stroke , diabetes , cancer , neurodegenerative diseases ( ND ) , and mortality in the ARIC study , the Framingham Heart Study ( N = 4 , 434 ) , and the Health and Retirement Study ( N = 9 , 676 ) . We leveraged the analyses of pleiotropy , age-related heterogeneity , and causal inferences . Meta-analysis of the results from these comprehensive analyses shows that the minor allele increases risks of death by about 50% ( p = 4 . 6×10−9 ) , CHD by 35% ( p = 8 . 9×10−6 ) , HF by 55% ( p = 9 . 7×10−5 ) , stroke by 25% ( p = 4 . 0×10−2 ) , and ND by 100% ( p = 1 . 3×10−3 ) . This allele also significantly influences each of two diseases , diabetes and cancer , in antagonistic fashion in different populations . Combined significance of the pleiotropic effects was p = 6 . 6×10−21 . Causal mediation analyses show that endophenotypes explained only small fractions of these effects . This locus harbors an evolutionary conserved gene-desert region with non-coding intergenic sequences likely involved in regulation of protein-coding flanking genes ZEB2 and ACVR2A . This region is intensively studied for mutations causing severe developmental/genetic disorders . Our analyses indicate a promising target region for interventions aimed to reduce risks of many major human diseases and mortality .
The demographic transition on aging of populations in developed countries requires strategies , which could extend healthspan and lifespan , and compress morbidity [1–3] . Breakthroughs in genome-wide sequencing and high-throughput genotyping raised enthusiasm for advancing the progress in the field by discovering genes influencing various health-related traits . To accelerate the progress , one necessarily faces the need to deal with genetic predisposition to complex , inherently heterogeneous , age-related traits , i . e . , traits that are characteristic of the elderly people in modern societies . Heterogeneity is the result of various processes . Of the most familiar sources of heterogeneity , genome-wide association studies ( GWAS ) commonly handle those associated with evolutionarily selected genetic patterns in populations [4] and complex etiologies of human health-related phenotypes [5] . Age-related traits are a special case of heterogeneous phenotypes , because they emerge as a new widespread phenomenon in humans [6] , especially given substantially shorter lifespan of our even recent predecessors [7] , and because they are characteristic of the post-reproductive period , when the role of evolutionary selection in these traits is elusive [6 , 8–12] . These factors imply that “diseases are not shaped by selection , ” [6] , i . e . , evolution did not fix the molecular basis of age-related traits . The latter makes the analyses of genetic influence on such traits be a more challenging task than that of fitness-related traits ( e . g . height [13] ) . An important challenge is a special type of heterogeneity attributed to the elusive role of evolution in shaping the genetic basis of age-related traits . This heterogeneity is the result of age-related processes in an organism and compositional changes in a population in changing environment [9 , 14 , 15] . The age-related heterogeneity implies the potential existence of gene-endophenotype-phenotype pathways ( i . e . , mechanisms mediating the effects between genes and age-related traits ) and that these mechanisms can change with age , time , and population composition even if the same genetic variant and trait are considered . This heterogeneity can naturally contribute to non-replication of genetic effects in different populations even in case of populations of the same ancestry and phenotypes [16–18] . The elusive role of evolution in fixing molecular basis of age-related traits can also benefit genetic analyses because it can enhance the basis of pleiotropic influences on different traits , including “apparently distinct” ones [19] . Statistical benefit is that pleiotropic analysis may improve power [20] . Substantive benefit is that pleiotropic influences on apparently distinct traits are a part of an attractive gerontological idea which has been conceptualized as geroscience [21] . This concept assumes that age [22] and aging [23] can be major risk factors of geriatric diseases of distinct etiologies [24] . Detecting such pleiotropy and developing gene-based interventions may strengthen strategies for reducing burden of not just one disease but a major subset of them to efficiently extend healthspan and lifespan [23 , 25 , 26] . Given specific properties of age-related traits , common methods of GWAS of these traits may be insufficient and more comprehensive methods typical for candidate gene studies may be needed . One strategy to improve genetic analyses of age-related traits is to use genome-wide scan for pre-selection of promising SNPs and more comprehensive methods for detail analyses of these SNPs in large samples . Following this approach , we selected promising SNPs from GWAS of the Atherosclerosis Risk in Communities ( ARIC ) Study ( see Methods ) . Then we conducted detail analysis of associations of SNPs at a promising locus in band 2q22 . 3 with risks of major diseases including coronary heart disease ( CHD ) , heart failure ( HF ) , stroke , diabetes , cancer , and neurodegenerative diseases ( ND , dementias including Alzheimer’s type ) and risks of death using the ARIC , the Framingham Heart Study ( FHS ) , and the Health and Retirement Study ( HRS ) . We leveraged the analyses of pleiotropy , age-related heterogeneity , and causal inferences . In causal mediation analyses , we used biomarkers ( body mass index [BMI] , total [TC] and high-density lipoprotein [HDL-C] cholesterols; triglycerides [TG] , systolic [SBP] and diastolic [DBP] blood pressures ) as endophenotypes for diseases or death and diseases as endophenotypes for death . We show that the minor allele increases risks of death by about 50% ( p = 4 . 6×10−9 ) , CHD by 35% ( p = 8 . 9×10−6 ) , HF by 55% ( p = 9 . 7×10−5 ) , stroke by 25% ( p = 4 . 0×10−2 ) , and ND by 100% ( p = 1 . 3×10−3 ) . This allele also significantly influences each of two diseases , diabetes and cancer , in antagonistic fashion in different populations . Combined significance of the pleiotropic effects was p = 6 . 6×10−21 .
Table 2 shows the estimates of risks of major human diseases or death for carriers and non-carriers of the minor allele . This allele is highly significantly associated with risks of CHD in ARIC ( HR = 1 . 74 , p = 1 . 1×10−7 ) and death in FHS ( HR = 1 . 64 , p = 1 . 3×10−6 ) . Leveraging these pleiotropic associations , the global null hypothesis that neither of these associations is true evaluated using the Fisher’s combined probability test [27] is p = 4 . 7×10−12 . This result implies that the probability of being a false finding in this case is much smaller than that defined by the genome wide significance ( pGW = 5×10−8 ) . Multiple testing correction for 15 tests with other phenotypes , which are not included in the Fisher’s test ( see Table 2 ) , does not alter this result , p = 4 . 7×10−12 × 15 = 7 . 1×10−11 << pGW . The minor allele is also nominally significantly ( p<5 . 0×10−2 ) associated with risks of death in two other studies ( ARIC and HRS ) and with risks of HF ( ARIC , FHS ) , stroke ( FHS ) , diabetes ( ARIC , HRS ) , and cancer ( FHS , HRS ) . It increases risks of all diseases and death ( Table 2 ) , except cancer in HRS . All non-significant effects were also detrimental . FHS sample includes two cohorts of participants from different generations ( the FHS original cohort and the FHS offspring [FHSO] cohorts ) which may be a natural source of age-related genetic heterogeneity ( see the Introduction ) . Evaluation of risks of the selected diseases and death in each cohort separately ( Table 3 ) shows that weak and highly non-significant association of the rs222826_T allele with risk of diabetes in the entire FHS sample ( HR = 1 . 05 , p = 7 . 6×10−1; Table 2 ) is due to antagonistic effects of this allele on risks of diabetes in the FHS original ( HR = 0 . 71 , p = 2 . 3×10−1 ) and FHSO ( HR = 1 . 36 , p = 1 . 5×10−1 ) cohorts . Formal test shows that multiplicative interaction of the minor allele with these FHS cohorts is significant ( HR = 2 . 5 , p = 1 . 1×10−2 ) . Explicating this heterogeneity , the effect size in the FHSO ( HR = 1 . 36 , p = 1 . 5×10−1 ) became the same as in ARIC ( HR = 1 . 35 , p = 4 . 2×10−2 ) and HRS ( OR = 1 . 35 , p = 9 . 9×10−3 ) . Important result is that we observe the same effect sizes in three cohorts of younger individuals who were born in the same time period around 1930s-1940s ( Table 1 ) . The opposite effect is observed in the FHS original cohort for individuals from substantially older generation born around 1910s ( Table 1 ) . Kaplan-Meier survival curves ( Fig 1A–1C ) suggest that in ARIC and each FHS cohort the rs222826_T allele carriers can be at antagonistic risks of CHD at different ages . They can be protected against early onset CHD and be at risk of later onset CHD . These antagonistic effects are most pronounced in the FHS original cohort . This heterogeneity implies that the estimates of the risks in this cohort are biased because of disproportional hazards . Correcting for this heterogeneity by focusing on individuals with onset of CHD at 65 years and older ( Fig 1D ) , the effect size in the FHS original cohort becomes nearly the same as in ARIC and attains suggestive significance , HR = 1 . 50 , p = 7 . 9×10−2 . Given old ages of the genotyped participants of the FHS original cohort , we tested the association of the rs222826_T allele with ND . Table 3 shows that this allele is significantly associated with ND as well . To address this question , we conducted causal mediation analysis ( see Methods ) . First , we evaluated the associations of the minor allele with the selected biomarkers in each study . The analyses show significant and marginally significant associations of this allele with HDL-C ( β = -3 . 8 , p = 4 . 7×10−3 ) and TG ( β = 6 . 6 , p = 3 . 3×10−3 ) in ARIC and BMI ( β = -3 . 4 , p = 5 . 7×10−2 ) in the FHS original cohort ( S1 Table ) . Then , we evaluated mediating roles of these biomarkers in effects between rs222826 and risks of the selected diseases and death . The analysis in ARIC ( Table 4 ) showed significant indirect effects of HDL-C and TG in associations of rs222826 with the risks of majority of diseases and death , except cancer ( HDL-C and TG ) and diabetes ( TG ) . The sizes of these indirect effects were , however , substantially smaller than those of “direct” effects ( direct effect means the effect not through the selected mediator ) . Accordingly , significant mediating effects of lipids explained only a small fraction of the total genetic effects ( they are given in Tables 2 and 3 ) . The analyses in the FHS original cohort show significant indirect effects of the rs222826_T allele on risks of HF , diabetes , and ND through BMI ( Table 4 ) . As in ARIC , these effects also represented a small fraction of the total effects . However , unlike ARIC , the HRs for the indirect effects were less than one for all diseases . Because the minor allele showed protective effect against diabetes in this cohort ( HR = 0 . 71 , p = 2 . 3×10−1 , Table 3 ) , this indirect effect implied that the association of the minor allele with BMI partly mediated ( explained ) the association of this allele with diabetes . For HF and ND , conditioning on BMI amplified detrimental effects between the rs222826_T allele and these diseases ( compare Tables 3 and 5 ) because of explicating a fraction of BMI-related genetic heterogeneity . Table 6 shows that CHD , HF , and diabetes significantly mediate the risk of death for the rs222826_T allele carriers in ARIC . Diabetes explained 12 . 5% , —i . e . , 4% ( HR = 1 . 04 , see Table 6 ) of 32% ( HR = 1 . 32 , see Table 2 ) , —of the death risk . CHD or HF explained 28 . 1% of the death risk . Mediating effect of either of these diseases ( CHD , HF , or diabetes ) was also highly significant ( HR = 1 . 11 , p<8 . 0×10−3 ) explaining about 34 . 4% of the death risk . In FHS ( Table 6 ) , only stroke ( FHS_C1 ) and diabetes ( FHSO ) showed small , marginally significant mediating effects between the rs222826_T allele and risk of death . In HRS , the minor allele showed small but significant indirect effects on risk of death through diabetes and cancer ( Table 6 ) . Indirect effect through diabetes was of mediating nature explaining a small fraction ( 6 . 7% ) of the total risk of death . Cancer showed moderating effect amplifying the total risk of death for the minor allele carriers in additive and multiplicative approximations ( Table 5 ) . Explicating age-related heterogeneity in the above sections helped in gaining further insights into genetic predisposition to diseases and death and , as a result , in improving estimates of the effects of the minor allele on these outcomes . Accordingly , pooling genetic effects in different populations should leverage these insights . Table 7 shows the results of meta-analyses leveraging information from the analyses of age-related heterogeneity in FHS ( Table 3 ) and causal inferences ( Table 5 ) . These analyses leveraged also potential substantial basis of antagonistic effects for diabetes ( see sections “Explicating age-related genetic heterogeneity in the FHS” above and “The role of endophenotypes” in the Discussion ) by pooling evidences for the effects in different FHS cohorts disregarding the effect directions . Meta-analysis of detrimental effects for diabetes , which are seen in younger individuals only , gives HR = 1 . 35 , p = 3 . 3×10−4 . Further , significant antagonistic effects for cancer in different studies imply significant heterogeneity ( e . g . , evidenced by non-overlapping 95% CI in Table 2 for cancer estimates in FHS and HRS ) . Accordingly , the result of meta-analysis in this case was presented in Table 7 disregarding the effect directions . An estimate for cancer in more homogeneous samples of ARIC and FHS is HR = 1 . 23 , p = 1 . 4×10−2 . For comparison , Table 7 also provides estimates without leveraging all this information . Table 7 shows pleiotropic associations of the minor allele with risks of all major diseases and death . Combining p-values for these pleiotropic associations into a single p-value using the Fisher’s test [27] , the global null hypothesis that neither of these associations is true is p = 6 . 6×10−21 . Following meta-analysis without leveraging additional information the estimate is still highly convincing p = 8 . 3×10−14 . This test provides inflated estimates because it disregards potential correlation of the association signals . However , this is a reasonable approximation for the combined significance of the pleiotropic effects of the minor allele because this p-value aggregates the estimates from independent studies .
Our initial findings ( Table 2 ) show consistent and significant associations of the minor allele ( rs222826_T in ARIC and FHS and rs222827_A in HRS ) with risks of death in each study but dissimilar associations of this allele with diseases . Dissimilar associations in different studies may reflect differences in biodemographic structures in these studies [28] . The results of the analyses of age-related heterogeneity in the FHS ( Table 3 ) support this mechanism of non-replication of the genetic associations with risks of diabetes and CHD . Indeed , explicating antagonistic effects between the rs222826_T allele and risk of diabetes in two FHS cohorts shows a striking result that the detrimental effect of this allele is actually the same in three younger cohorts ( ARIC , FHSO , and HRS ) of individuals born in about the same time period around the 1930s-1940s . The effect is opposite ( protective ) in substantially older population with mean birth year around the 1910s . For CHD , we observe antagonistic risks at different ages , which cause biased estimates in the models based on the assumption of proportional hazards ( Fig 1 ) . Explicating this age-related heterogeneity by focusing on onset of CHD at later ages in the FHS original cohort detected replicating signal . Importantly , the effect size in this case becomes comparable with the effect size in the association of the rs222826_T allele with CHD in ARIC . Accumulating evidence suggests the importance of age-related heterogeneity in genetic effects . The analyses highlighted the role of age in genetic regulation of BMI [29] , sensitivity of the effects of longevity alleles to birth cohorts [30 , 31] , sensitivity of genetic associations with lipids to chronological age [32 , 33] , changes in the allele frequencies with age [34 , 35] , antagonistic risks of diseases and death [16 , 32] . An important result of these analyses is that they highlight potential non-stochastic mechanisms , which can contribute to non-replication of genetic associations . Clearly , biodemographic factors are not the only ones that can cause non-replication; other factors ( e . g . , GxG interactions [14] ) may play a role . Causal mediation analyses showed that three biomarkers-endophenotypes ( HDL-C and TG in ARIC and BMI in the FHS original cohort ) significantly moderated the effects between the rs222826_T allele and risks of several diseases and death . They , however , accounted for a small fraction of the genetic effects implying that the major fraction of these effects is not through the selected biomarkers . As expected , lipids showed significant mediating effects explaining a fraction of the total detrimental effects on cardiovascular diseases in ARIC . Favorable association of the rs222826_T allele with BMI in the FHS original cohort partly explained its favorable association with risk of diabetes in the same cohort ( Table 4 ) . This result emphasizes real nature of the protective ( though insignificant ) effect of this allele on diabetes . It also shows that favorable associations with diabetes and BMI are characteristic for older people from early birth cohorts ( represented by the FHS original cohort ) . The lack of favorable associations with BMI and the presence of detrimental associations with diabetes in younger cohorts ( see the above section ) may indicate change in the mechanisms connecting this allele with diabetes in older and younger cohorts [10] . This change is consistent with recent trend on increase of incidence of diabetes [36] . We also found that BMI was a significant moderator in the FHS original cohort amplifying effects between the rs222826_T allele and HF or ND . This moderation effect implies that favorable association of this allele with BMI ( S1 Table ) can partly mitigate detrimental effects of this allele on HF and ND ( compare Tables 3 and 5 ) . Interestingly , this analysis suggests that BMI can be involved in a pathway linking the rs222826_T allele with ND . This is in line with findings in large epidemiological studies reported on association of BMI with ND [37] , that is also seen in the FHS original cohort ( Table 5 ) . The mediation analysis of indirect effects of the rs222826_T allele on risks of death through diseases-endophenotypes ( Table 6 ) provided strongest evidence for significant mediating effects of CHD , HF , and diabetes in ARIC . Combined mediating effect of these diseases explained about 34 . 4% of the death risk . Other diseases in ARIC , FHS , and HRS either mediated substantially smaller fractions of the total effect or did not mediate it at all . We found that the rs222827_A allele and cancer increase the risk of death in HRS additively ( Table 5 ) . However , cancer patients who carry and do not carry this allele show the same survival . This result indicates that detrimental effect of this allele on risk of death can be partly mitigated by ( unknown ) genetic and/or environmental factors . Thus , the results from the causal mediation analyses indicate that most of the effects between the minor allele and risks of diseases and death are only partly explained by the selected endophenotypes . These results suggest that such a wide impact of this allele on phenotypes with major contribution to healthspan and lifespan may indicate connections of this variant with some fundamental biological mechanisms ( see below ) that is in line with the concept of geroscience [21] . Modulation of the effects by age-related heterogeneity and endophenotypes suggests a role of other factors ( other genes and/or environment ) in the effects of this allele . A “side effect” of gaining insights into intermediate mechanisms connecting genes with major phenotypes contributing to healthspan and lifespan discussed above is improving statistical estimates ( Table 5 ) . Indeed , Table 7 shows improvement in overall significance of the combined pleiotropic effect of the minor allele by seven orders of magnitude from p = 8 . 3×10−14 to p = 6 . 6×10−21 . The minor allele increases risks of death by about 50%; this estimate is genome-wide significant ( p = 4 . 6×10−9 ) . In addition , the same allele increases risks of CHD by 35% ( p = 8 . 9×10−6 ) , HF by 55% ( p = 9 . 7×10−5 ) , stroke by 25% ( p = 4 . 0×10−2 ) , and ND by 100% ( p = 1 . 3×10−3 ) . This allele is also associated with risks of diabetes ( p = 1 . 6×10−4 ) and cancer ( p = 1 . 8×10−3 ) . Most of its effects are detrimental as it increases risks of diabetes in younger generations from ARIC , FHSO , and HRS by 35% ( HR = 1 . 35 , p = 3 . 3×10−4 ) and risk of cancer in ARIC and FHS by 23% ( HR = 1 . 23 , p = 1 . 4×10−2 ) . The rs222826 ( and its proxy in HRS , rs222827 , which are 90 bp apart ) SNP is an intergenic variant with MAF of about 2 . 5% in each of three Caucasian populations in ARIC , FHS , and HRS . This SNP is located on chromosome 2 in band q22 . 3 , which harbors gene desert region ( S1 Fig ) . Studies show that gene deserts ( which make up ~25% of the genome [38] ) exhibit characteristics suggestive of functional importance [39] . Functional role of gene deserts is supported by the fact that some of them are evolutionary conserved suggesting their essential role in regulation of core vertebrate genes [40 , 41] . The rs222826/rs222827 SNPs are within an evolutionary conserved gene desert region [40–42] , which contains intergenic regulatory sequences likely involved in regulation of the expression of protein-coding flanking genes ZEB2 ( zinc-finger , E-box-binding homeobox-2 ) ( -1 . 6 Mb ) and ACVR2A ( activin receptor type-2A ) ( +1 . 7 Mb ) . Function of ZEB2 can be directed in a tissue- and age-dependent manner by long- ( 1 . 2 Mb ) and short- ( 62 Kb ) distance enhancers suggesting a conserved regulatory string of enhancers for ZEB2 and possibly ACVR2A [43] . Other long-range enhancers for ZEB2 were also observed [44 , 45] . These SNPs are also in LD with SNPs from nearby regulatory regions ( e . g . , r2 = 1 with rs222809; S2 Fig ) . In addition , gene expression may be also regulated through non-coding RNAs [46–48] . The ZEB2 gene functions as a regulator of transcription interacting with activated SMADs in the TGF-β signaling pathway and ACVR2A is part of a receptor complex that binds and activates SMAD transcriptional regulators . Accordingly , these genes are linked through SMAD proteins and TGF-β signaling . The ZEB2 gene is one of key regulators of epithelial-to-mesenchymal transition playing a critical role in the development of neural crest and is involved in the development of other organs that are not derived from the neural crest . This gene mediates multiple pathways related to inflammation , aging and carcinogenesis [49] . The ACVR2A gene takes part in many distinct pathways by mediating the functions of members of TGF-β superfamily which are involved in a variety of biological functions including development and tissue homeostasis and associated with a wide range of human diseases [50 , 51] . Various mutations in ZEB2 ( e . g . , haplo-insufficiency , gene inactivation and deletions ) and deletions at 2q22-24 are associated with a Mowat–Wilson syndrome , a complex developmental disorder involving a range of physical symptoms as well as severe intellectual disorders [45 , 52 , 53] . Detrimental effects caused by the deletions in chromosomal region harboring rs222826/rs222827 and by mutations in flanking genes strengthen functional role of this evolutionary conserved region [44 , 54] . Potential functional importance of this gene desert is supported by the results of our analyses showing extensive pleiotropic effects on major human diseases and strong effect on human survival .
The ARIC Study participants [55] ( aged 45–64 at baseline in 1987 ) were randomly selected and recruited at four field centers across the U . S . We used data from four available examinations . Measurements of biomarkers were available in all examinations . Data on onsets of diseases and survival were available through 2004 . Genotyping for 12 , 771 ARIC participants ( N = 9 , 633 whites ) was conducted using Affymetrix 6 . 0 arrays ( 1 , 000K SNPs ) . The FHS design has been previously described [56–58] . We used data from 28 examinations of the FHS original cohort ( aged 28–62 years at baseline in 1948 ) , 8 examinations of the FHS Offspring ( FHSO ) cohort ( aged 5–65 years at baseline in 1970 ) , and one examination of the 3rd Generation ( 3rd Gen ) cohort ( aged 21–71 at baseline in 2001 ) . Measurements of biomarkers were available at multiple examinations in the FHS/FHSO and the baseline in the 3rd Gen cohort . Data on onsets of diseases and survival were available through 2011 . Biospecimens were mostly collected in the late 1980s and through the 1990s from surviving participants [59 , 60] . Genotyping of 9 , 167 white FHS participants was conducted using Affymetrix 500K arrays [58] . The HRS design has been previously described [61] . We used available information on biomarkers measured in 2006–2008 and on survival during follow up from 2006 ( time of biospecimen collection ) through 2013 . The data on onsets of diseases was not available . The HRS genotyped about 2 . 5M SNPs for 12 , 507 subjects ( N = 9 , 736 whites ) using the Illumina HumanOmni 2 . 5 Quad chip . The focus of the analyses was on risks of major human diseases available in the data including coronary heart disease ( CHD ) , heart failure ( HF ) , stroke , diabetes , cancer , and neurodegenerative diseases ( ND , dementias including Alzheimer’s type ) , and risk of death . Biomarkers represented by the traditional risk factors for cardiovascular diseases were used for causal mediation analyses ( see below ) . They included body mass index ( BMI ) , total cholesterol ( TC ) , high density lipoprotein cholesterol ( HDL-C ) , triglycerides ( TG ) , systolic blood pressure ( SBP ) , and diastolic blood pressure ( DBP ) . Because genetic variants may play a complex role in age-related traits ( see the Introduction ) , traditional GWAS techniques , including those designed to evaluate pleiotropic associations [20] , may not necessarily address complexity of genetic influence on such traits [28] . Accordingly , the focus of this paper was on comprehensive analyses using more detailed candidate-gene-like techniques . GWAS was used as a tool to preselect variants , which showed promising pleiotropic properties . Below we detail the analyses sketched in the flowchart in Fig 2 . This manuscript was prepared using controlled-access data obtained though dbGaP ( accession numbers phs000007 . v22 . p8 , phs000280 . v2 . p1 , phs000428 . v1 . p1 ) . Phenotypic HRS data are available publicly and through restricted access from http://hrsonline . isr . umich . edu/index . php ? p=data .
|
Biomedical research and medical care are traditionally focused on individual health conditions in order to postpone , ameliorate , or prevent the accumulation of morbidities in late life . An attractive idea is to find factors , which could reduce burden of not just one disease but a major subset of them to efficiently extend healthy lifespan . Here we focus on the analyses of genetic predisposition to risks of major human age-related diseases and mortality . The analyses highlight a locus in band 2q22 . 3 associated with risks of coronary heart disease , heart failure , stroke , diabetes , cancer , neurodegenerative diseases , and death . Our analyses indicate a promising target region for interventions aimed to reduce risks of many major human diseases and mortality .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"neurodegenerative",
"diseases",
"biomarkers",
"diabetes",
"mellitus",
"endocrine",
"disorders",
"coronary",
"heart",
"disease",
"mathematics",
"statistics",
"(mathematics)",
"alzheimer",
"disease",
"cardiology",
"research",
"and",
"analysis",
"methods",
"stroke",
"endocrinology",
"mathematical",
"and",
"statistical",
"techniques",
"dementia",
"heart",
"failure",
"mental",
"health",
"and",
"psychiatry",
"metabolic",
"disorders",
"biochemistry",
"cerebrovascular",
"diseases",
"neurology",
"meta-analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"vascular",
"medicine",
"genetics",
"of",
"disease",
"statistical",
"methods"
] |
2016
|
Pleiotropic Associations of Allelic Variants in a 2q22 Region with Risks of Major Human Diseases and Mortality
|
Transcription factor ( TF ) proteins recognize a small number of DNA sequences with high specificity and control the expression of neighbouring genes . The evolution of TF binding preference has been the subject of a number of recent studies , in which generalized binding profiles have been introduced and used to improve the prediction of new target sites . Generalized profiles are generated by aligning and merging the individual profiles of related TFs . However , the distance metrics and alignment algorithms used to compare the binding profiles have not yet been fully explored or optimized . As a result , binding profiles depend on TF structural information and sometimes may ignore important distinctions between subfamilies . Prediction of the identity or the structural class of a protein that binds to a given DNA pattern will enhance the analysis of microarray and ChIP–chip data where frequently multiple putative targets of usually unknown TFs are predicted . Various comparison metrics and alignment algorithms are evaluated ( a total of 105 combinations ) . We find that local alignments are generally better than global alignments at detecting eukaryotic DNA motif similarities , especially when combined with the sum of squared distances or Pearson's correlation coefficient comparison metrics . In addition , multiple-alignment strategies for binding profiles and tree-building methods are tested for their efficiency in constructing generalized binding models . A new method for automatic determination of the optimal number of clusters is developed and applied in the construction of a new set of familial binding profiles which improves upon TF classification accuracy . A software tool , STAMP , is developed to host all tested methods and make them publicly available . This work provides a high quality reference set of familial binding profiles and the first comprehensive platform for analysis of DNA profiles . Detecting similarities between DNA motifs is a key step in the comparative study of transcriptional regulation , and the work presented here will form the basis for tool and method development for future transcriptional modeling studies .
Transcription factor ( TF ) proteins usually recognize a small number of DNA targets via the formation of sequence-specific and nonspecific molecular interactions . Understanding the evolution of TF DNA-binding preferences will not only provide useful insights on the mechanism of DNA recognition , it will also allow more accurate prediction of genomic regulatory elements , which still constitutes a major hurdle in understanding cellular gene regulatory networks . Furthermore , high-throughput studies , such as microarray and ChIP–chip , generate a number of DNA motifs that are putative targets of usually unknown TFs . In this study , we present an alignment and comparison platform that is optimized for DNA motifs , thereby allowing for their efficient analysis and enabling and formalizing their evolutionary study . This platform is called STAMP , for similarity , tree-building , and alignment of DNA motifs and profiles . TF DNA-binding preferences are usually modeled via frequency matrices , derived from alignments of known sites ( see Methods ) . Typically , these position-specific scoring matrices ( PSSMs ) assume independency between the base positions [1] . It has been recognized that structurally related TFs often share similarities in their DNA-binding motifs , although the extent to which this happens depends on the TF family [2–5] . Generalized binding models or familial binding profiles ( FBPs ) , a term coined by Sandelin and Wasserman [6] , constitute an “average” binding specificity of a family of TFs ( see Figure 1 for an illustration of the FBP concept ) . FBPs can be incorporated in pattern-finding algorithms as prior knowledge in order to bias them towards motifs from a particular TF family [6–8] . This is useful if the investigator expects to find motifs from a particular class of TFs . The use of FBPs as prior information focuses the motif search on biologically relevant patterns , offering a way to improve upon the currently limited performance of DNA motif-finders [9] . FBPs have been used to infer the identity of the TF family bound to predicted novel motifs [6 , 8 , 10] , and to remove degeneracy between related motifs in the motif repositories [11–13] . More recently , FBPs have been used to help estimate the binding specificity of regulatory proteins from ChIP–chip data [14] . The early studies introducing FBPs demonstrated their potential in regulatory DNA analysis . However , the methods employed to compare and align DNA-binding motifs , a key aspect in constructing FBPs , have not been thoroughly studied . Currently , the construction of FBPs is based on ( semi ) -empirical clustering methods and ad hoc distance metrics . Ungapped local motif alignments [6 , 7] or enumeration of subsequence frequencies across related motifs' members [10 , 11] are typically used to compare PSSMs , although it is not yet known if these strategies are optimal . Even the definition of binding motif families and subgroups is currently problematic . Structural information and protein sequence comparisons have been previously used to guide manual clustering of TF binding profiles [6 , 8] , although automatic methods have been recently introduced [7] . For more reliable FBP construction methods , and in order to expand the area of applications for generalized binding models , a detailed evaluation of a variety of motif alignment strategies is required . Motif evolution and the motif dependence on the proteins' structural properties need to be investigated . Sandelin and Wasserman [6] did an important first step when they created a set of 11 FBPs for the nonzinc finger families . During the FBP construction , they noticed that the bZIP family exhibited two different DNA-binding patterns , so they partitioned it into CCAAT/enhancer binding protein ( C/EBP ) - and cAMP response element-binding ( CREB ) -related proteins . However , their FBP clustering was done manually and other ( sub ) family characteristics may have been missed . On the other side of the spectrum , two families may have similar binding preferences , and although they may not belong to the same structural group , it may be reasonable to cluster their binding profiles together , so that the overall number of false positive predictions is reduced . Finally , a detailed analysis of the structural properties of the protein–DNA complexes , together with automatic clustering and classification results , is hoped to shed more light on the evolution of the DNA preferences and their utility in prediction and classification studies . Schones et al . have compared the effectiveness of three profile distance metrics [13] . We expand their study by evaluating combinations of six distance metrics , three pairwise alignment methods , two multiple-alignment strategies , and two tree-building algorithms . In addition , we develop a new statistic for automatically deciding the optimal number of clusters in a given motif tree . We use this statistic on the tree obtained from the optimal distance metrics and alignment strategies combination to generate a new set of FBPs without prior knowledge of TF structural classification . The new collection of FBPs exhibits better TF classification accuracy than previous manually derived clusters [6] and identifies similarities and differences in the binding preferences of TF ( sub ) families . Although species-specific binding preference may emerge for some TFs [15] , in general structurally related TFs often share similarities in their DNA-binding preferences . Exploring this trend , Narlikar and Hartemink built a Bayesian TF structural family classifier based on the DNA motifs [10] . We found that the same accuracy of this sophisticated method can be achieved with simple motif similarity searches when the appropriate alignment algorithms are used . Correctly predicting the TF structural class for novel motifs will be a crucial step in the interpretation of experiments that aim to systematically estimate regulatory motifs in entire mammalian genomes ( e . g . , [16] ) .
All columns from the known PSSM models in the TRANSFAC database [17] were compared with each other using the six metrics presented in Table 1 ( see Methods ) . Figure 2 shows the great variability in the range and distribution of the scores . Both Pearson's correlation coefficient ( PCC ) and average log likelihood ratio ( ALLR ) have negative expected values , which makes them especially suitable for use in alignment algorithms ( although , negative mean values can be obtained from any metric by subtracting an appropriate number ) . Three of the methods , namely PCC , sum of squared distances ( SSD ) , and p-value of chi-square ( pCS ) , have peaks of very small variance , with PCC having two distinct peaks . Comparison of JASPAR [18] columns gave similar results ( unpublished data ) . Each of the six metrics was tested for its ability to discriminate between columns randomly sampled from two distinct distributions: an information content–specific distribution and a background ( reference ) distribution . For information content , I , Figure 3 plots the percent of the -sampled columns that were included in the area around when a false discovery rate ( FDR ) of 1% was reached . For lower information content , ALLR and SSD perform best at discriminating columns sampled around and . The finding that ALLR is a better discriminator than pCS metric may seem to contradict the findings of Schones et al . [13] . However , apart from our sampling size being larger , their evaluation focused on the whole motif level . As we will see later , the advantageous performance of ALLR in column-to-column comparisons does not seem to extend to motif-to-motif comparisons . It is difficult to construct an unbiased artificial dataset for the evaluation of motif alignment strategies . However , an indication of performance may be gained from similarity searches of a motif against all motifs in a database . Generally , in databases with good representation , the best match to a given motif is expected to be a motif associated with a member of the same structural class [6] . The “best-hit” approach can be used to assess the relative effectiveness of column-scoring metrics and alignment-method combinations by finding the proportion of motifs that match another member of the same structural class using each strategy . One hundred and five combinations of similarity metrics , alignment methods , and gap penalty values were tested over two datasets: the JASPAR- and TRANSFAC-derived models ( see Methods ) . The top 15 and bottom 15 performing strategies/combinations are presented in Table 2 , and a full table of results is available in Table S1 . Smith–Waterman local alignments populate the list of best performing results , indicating that they are generally better than Needleman–Wunsch global alignments in motif alignment applications . The results also suggest that the PCC and SSD metrics are on average more effective than the AKL , pCS , and ALLR metrics ( including ALLR_LL ) for whole-motif comparison . The best-performing combination is Smith–Waterman local alignment using the SSD metric and gap open = 1 ( average accuracy 0 . 811 ) , whereas Smith–Waterman local alignment using the PCC metric also scores highly ( seventh-best score; average accuracy 0 . 805 ) . The first strategy that uses the AKL metric appears at position 28 of the list , and the first strategy using ALLR_LL appears at position 37 . Strategies using the standard ALLR metric first appear at position 48 , and strategies using the pCS metric first appear at position 60 . In the study of Narlikar and Hartemink [10] , a sparse Bayesian learning algorithm was used to predict the structural class of the TF that binds to a given DNA motif . The test dataset in that study consisted of the six largest motif families in TRANSFAC , and Narlikar and Hartemink's algorithm was able to correctly predict the TF structural family for 86% of these DNA-binding motifs . By analyzing the same dataset , we found that the best-hit approach ( with ungapped Smith–Waterman alignment and the PCC metric ) yields practically the same performance ( 87% ) . The best-hit searches perform better in predicting the bZIP , C4 zinc finger , and Forkhead families , whereas the Bayesian learning is better in the more diverse ( in terms of DNA binding motifs ) C2H2 zinc finger , homeodomain , and bHLH families ( Table 3 ) . We will later show that the appropriate clustering of DNA motifs can help improve the classification accuracy . Sandelin and Wasserman manually constructed FBPs for those ten nonzinc-finger structural families for which four or more motifs exist in the JASPAR database ( 71 motifs ) [6] . One of the ten families , bZIP , produced two distinct FBPs: one related to C/EBP and one related to the CREB proteins . The set of 71 profiles provides an appropriate dataset for testing tree-building strategies and automatic clustering methods . In this study , an agglomerative ( UPGMA ) and a divisive ( SOTA ) strategy were compared ( see Methods ) . The most effective alignment strategy for the nonzinc-finger JASPAR dataset , the ungapped Smith–Waterman alignment , was used with each of the six similarity metrics in conjunction with UPGMA or SOTA . Performance was measured as the average homogeneity of the families at each leaf node on the tree and with respect to the tree growth . For a given node , a performance average homogeneity score of 1 denotes that only one family is represented in the motifs clustered at that node ( perfect homogeneity ) ; a score of 0 . 5 denotes that two equally represented families are clustered in that node; etc . The point at which the average homogeneity of the leaf nodes reaches 1 is the point where motifs have been successfully separated on the basis of structural class ( although a class might be split into multiple nodes ) . As can be seen in Figure 4 , UPGMA generally performs better than the neural tree method ( SOTA ) , regardless of the similarity metric . Ungapped Smith–Waterman alignment using the PCC metric is by far the most successful metric on this dataset , managing to separate all ten structural families ( 100% average homogeneity ) with only 25 leaf nodes . In addition , ungapped Smith–Waterman alignment using the SSD metric achieves 95% average homogeneity with 26 leaf nodes . The actual tree resulting from the combination of ungapped Smith–Waterman alignment , the PCC metric , and the UPGMA tree construction method displays a high degree of separation of the TF structural classes ( Figure 5 ) . Estimating the optimal number of data clusters on a tree of binding motifs is of significant interest in classification and familial binding property analysis applications . It is well-known , however , that this is an inherently arbitrary procedure; different criteria on where to “split” the tree will give different estimates of cluster number . A number of statistics have been described that aim to estimate the optimal number of clusters ( e . g . , [19–21] ) , usually by seeking an optimal balance between intercluster and intracluster variability . We used a subset of the JASPAR motifs to understand how different metrics perform in determining the optimal number of clusters in a tree . For this purpose , only closely related members of the ETS , REL , Forkhead , high mobility group ( HMG ) , and MADS families were used . The statistics we compared were: the Gap statistic of Hastie , Tibshirani , and Walter ( HTW ) [22] , the standard Calinski and Harabasz ( CH ) [19] , and a derivative ( CHlog ) we developed ( Table 4 ) . When tested on the tree of the five well-defined families , the standard CH statistic didn't yield any local maximum number of clusters . We believe this is because this statistic performs well when the number of clusters is small compared with the number of points . Otherwise , the within-cluster difference goes quickly to zero , driving the CH to infinity . The CHlog , by design , avoids this problem . HTW and CHlog both gave the correct optimal number of clusters in the five-family dataset . When tested on the full nonzinc-finger JASPAR dataset ( 71 motifs ) ( Figure 5 ) , however , HTW and CH yielded no optimum number of clusters , whereas CHlog gave 17 ( Figure 6 ) , which is a reasonable number ( compared with 11 of Sandelin and Wasserman ) . For the most part , the identified clusters of motifs seemed reasonable both in terms of binding motif conservation ( manual examination ) and in terms of TF subfamily classification . More information is provided in the Discussion section . The suitability of CHlog was also tested in trees with two , three , four , and six well-defined clusters ( from the above set of JASPAR PSSMs ) and always yielded the correct answer . We test the hypothesis that the 17 automatically generated FBPs are a more accurate representation of motif diversity in the JASPAR set than the 11 manually constructed motifs using leave-one-out cross-validation ( LOOCV ) . In this test , we treat the set of FBP clusters as static multiple alignments and remove the contribution of each motif from its appropriate FBP in turn . The removed motif is then compared against all regenerated FBPs , and treated as correctly classified if it most closely matches the FBP from which it was withdrawn . LOOCV using Sandelin and Wasserman's 11 manually defined FBPs results in nine misclassifications , or a classification performance of 62/71 = 87% ( this performance rate was also reported in [6] ) . By comparison , in our dataset of 17 automatically defined FBPs , LOOCV resulted in two misclassifications , which , combined with the two unclassifiable singleton clusters , suggests a classification performance of 94% ( 67/71 ) . JASPAR also contains three zinc-finger motifs that were not used in the Sandelin and Wasserman FBP construction . One of them , C2H2 , includes TF proteins with highly divergent patterns of contacts ( see Discussion ) . The other two , DOF ( a C4 zinc-finger family ) and GATA , have quite conserved DNA-binding patterns . We repeated the above analysis by including the DOF and GATA motifs in JASPAR ( four motifs in each family ) . Our method determined 18 clusters ( including two singletons ) , and a LOOCV test resulted in 72/79 correct classifications ( 91% ) . Sandelin's and Wasserman's method on the 13 clusters ( the previous 11 and the two zinc fingers ) resulted in 60/79 correct classifications in the LOOCV test ( 76% ) . Compared with the 71-motif tree ( Figure 5 ) , the new tree is identical in 15 of the clusters ( Figure 7 ) . The main difference is that the two-member cluster of the heterodimeric bHLH proteins , TAL1-TCF3 and HAND1-TCF3 , is now split . The TAL1-TCF3 motif is part of a new cluster with the ( previously singleton ) FOXL1 and the GATA-1 zinc-finger protein motif . The remaining three GATA proteins formed a new cluster . All four DOF proteins are co-clustered with the IRF proteins . All described methods have been compiled in a software platform ( STAMP ) ( Mahony S , Benos PV , STAMP: A web tool for exploring DNA-binding motif similarities , unpublished ) . STAMP is modularly designed to allow any combination of column–column scoring metric , alignment method , tree-building algorithm , and multiple-alignment strategy to be used . Its potential uses range from simple motif database searches to identify the TF that may bind to a particular motif to a full-scale analysis of multiple-aligned genomic regions . In the section below , examples of both these uses are provided . STAMP is publicly accessible from http://www . benoslab . pitt . edu/stamp/ .
Six PSSM column similarity metrics were evaluated together with three pairwise alignment methods ( two gapped and one ungapped ) , two multiple-alignment strategies , and two tree-building strategies on motif datasets . The results showed that the Smith–Waterman local alignment algorithm used with the PCC or SSD metrics generally performs better in aligning the currently available PSSM models with models for which the associated TF belongs to the same structural family . We also discovered that the high efficiency of some metrics in column-to-column comparison does not extend to the alignment of whole motifs , which is a surprising and previously overlooked outcome . In the case of the ALLR metric , we believe that this inconsistency is due to the metric's very negatively skewed scoring distribution ( Figure 2 ) . Although such a distribution might be advantageous in distinguishing between PSSM columns with low information content , it also makes motif alignment more difficult , especially when the motifs contain low-scoring regions that can rigorously negate the overall score . The difficulties in using ALLR became more apparent with the Needleman–Wunsch global alignment , as low scores that are frequently observed in the beginning and/or at the end of the alignment could not be adequately subsidized by the positive scores in the alignable areas . The use of the alternative ALLR_LL metric ( i . e . , ALLR with a lower limit of −2 ) improved the results slightly . However , ALLR is the only metric that takes into consideration the background distribution when it compares two columns , which , in combination with the fact that it distinguishes better between columns of low information content ( Figure 3 ) , can be advantageous in identifying the correct PSSM model among closely related models ( e . g . , those belonging to the same family ) . Smith–Waterman local alignments were found to be more effective than Needleman–Wunsch global alignments for DNA motifs . This is expected given the current status of the motif databases . Motifs in existing databases usually result from some automated method that runs on a set of unaligned sequences recorded in these databases . These motifs frequently consist of a “core” area of columns with high information content , surrounded by columns of low ( er ) information content . On such motifs , local motif alignment methods will tend to perform better than global alignment methods . Structurally , this also makes sense , since binding sites are often recognized by a single structural sequence recognition element ( e . g . , α-helix ) either surrounded by or adjacent to a less-specific element that provides additional binding energy . Interestingly , ungapped algorithms or gapped algorithms with high gap-opening penalties generally performed better with the same metric ( Table 2 ) . This is probably due to the fact that the motifs for TF families in JASPAR and TRANSFAC share ungapped regions of similarity . However , the difference between ungapped/high penalty and gapped with lower-gap penalty algorithms is marginal in the JASPAR and TRANSFAC databases . Gapped alignment methods are expected to be more effective when aligning families of motifs that share common half-sites with variable length spacer regions , like many prokaryotic sites . An unexpected finding was that the current multiclass motif classifiers perform no better than simple best-hit similarity queries against a motif database when appropriate motif alignment methods are used . Interestingly , the two zinc-finger families ( C2H2 and C4 ) are predicted with the worst and the best efficiency , respectively , which reflects their binding geometries . The C4 factors form extensive networks of contacts along the length of an α-helix embedded in a B-DNA major groove . As a result , their target sequences are very conserved and thus their predictions easier . The C2H2 zinc fingers on the other hand contact the DNA helix at an angle using only few amino acid side-chains extending from the end of a less intimately associated helix . This results in less binding dependence upon individual amino acid sequences . Changes in certain “key” amino acid positions can drastically alter the DNA-binding specificity , thus yielding highly variable targets . bZIP factors bind to DNA as dimers in palindromic targets and they select individual half-sites in the process . The monomers readily dissociate from the dimeric form , binding DNA initially as half-site monomers [23] . This kinetic selection process , along with the intimate association of the recognition helix with the major groove ( similar to C4 Zn fingers ) likely provides exquisite selectivity . The same selectivity may not be realized for the bHLH proteins . Although these factors also select out half-sites , their monomers , in contrast to bZIPs , have very low dissociation rates and act more like covalently linked DNA-binding domains ( similar to C2H2 factors ) . Also , the angle of interaction between the recognition helix and the major groove is more obtuse for bHLH than for bZIP ( 23 versus 20 degrees ) , resulting in less interaction with half-sites ( 3 bp instead of 4 bp ) . These differences may explain why a bHLH behaves more like a C2H2 in target heterogeneity . The relatively poor prediction specificity exhibited for the Homeo HTH domain proteins stems from the binding of these factors to highly divergent targets ( usually recognizing mostly an “AT” motif ) , probably due to their dependence upon partner domains that are either dissociable ( like Hox-Pbx , Ubx-Exd , and Mat a1-alpha2 ) or covalently linked ( like Oct , Pax , and Pit ) . Member families of the monomeric wHTH subtype have strong consensus correlations ( like the ETS and Forkhead families with consensus GGAA and TAAACA , respectively ) . This results in higher prediction efficiency for Forkhead than for the more variable Homeo class . It is this idea of variability , perhaps dependent upon multimerization-related constraints , that may be a useful basis for the distinctions we observe . Another problem that may impose a limit in the classification efficiency of any method ( regardless of the TF family ) is the quality of the TFBS alignment and the resulting PSSM models . It is hoped that with the accumulation of new data , this will become less of a problem in the future . Sandelin and Wasserman [6] had previously built 11 FBPs from 71 nonzinc-finger PSSM models ( ten TF families ) available in the JASPAR database . Their manual clustering of the PSSM models was based on prior knowledge of the structural class of the corresponding TF . An exception to this general rule was the bZIP family , for which they constructed two FBPs ( CREB- and C/EBP-related ) after observing very different DNA patterns . The FBP corresponding to each structural class was calculated from a multiple-motif alignment where the contributions from outlying motifs were negatively weighted . The 11 familial binding profiles and the 71 motifs in the training set are available to view from the JASPAR database website ( http://mordor . cgb . ki . se/cgi-bin/jaspar2005/jaspar_db . pl ) . Sandelin and Wasserman's manual approach is suitable for relatively small sets of binding motifs where the structural class corresponding to each motif is known , and where the representatives from each component structural class bind a set of closely related target motifs . However , in the more general case , where families that bind diverse target motifs are included or where the structural class of certain motifs is unknown , it may be useful to attempt automatic generation of the appropriate familial binding motifs . We developed a fully automated method for PSSM clustering , based on the combinations of metric , alignment strategies , and tree building examined in this study . The advantages of the automatic clustering are obvious . By remaining ignorant to the prior knowledge of the structural class of each motif , we can find cases where motifs from diverse structural classes are more suitably grouped together , if they have similar binding preferences . Similarly , the automatic approach avoids the temptation of forcing together subfamilies of the same structural class with different binding preferences . Also , “outlier” PSSMs can be easily detected through an automatic clustering and subsequently be excluded from the FBP . The method we used to determine the optimal number of clusters is similar to the Calinski and Harabasz statistic [19] , but the intercluster and intracluster variability is calculated on the log-scale . We found this method to compare favourably with other methods on datasets with a known , well-defined small number of clusters and in the whole dataset . When applied on the 71 JASPAR PSSM models of our dataset , this method yielded 17 clusters , two of which are singletons and another two of which contain a pair of heterologous TFs each ( Figure 8 ) . Overall , the automatic clustering method divides the dataset into homogeneous clusters with respect to the structural group of the corresponding TF ( note that the clusters are based solely on the binding preferences of the TFs ) . This agrees with the general notion that structurally similar TFs tend to have similar binding specificities . The MADS domain proteins and the wHTH ETS proteins are two examples of TF families with very conserved DNA-binding preferences . Also , homeobox and nuclear receptor family clusters are homogeneous , although some members of these families can be found in other clusters . Interestingly , our algorithm split the patterns of the members of the ( so-called ) TRP family into Myb-related proteins and IRF proteins . This may not be surprising , since these two HTH-like proteins exhibit distinct DNA-binding geometries . Myb proteins contain three HTH domains , only one of which is involved in the target recognition ( with reported consensus YAAC[G/T]G ) . The IRF family consists of wHTH proteins that bind as homodimers via nonpalindromic direct repeats [24] or as monomers cooperatively with other proteins , like ETS [25] . The IRF motif ( Figure 8 ) contains a repeat of the commonly reported [A/G]NGAAA consensus , which we attribute to the homodimerization binding of these proteins . Our algorithm also correctly recognized three subfamilies in the major bHLH family . The six members of the bHLH-zip subclass ( e . g . , USF1 , MAX , etc . ) are clustered together , whereas the remaining four “standard” bHLH proteins ( Myf , NHLH1 ) and bHLH complexes ( HAND1-TCF3 , TAL1-TCF3 ) form two separate clusters . Examination of the FBPs of these clusters ( Figure 8 ) shows clearly that binding preferences are substantially different , reflecting their corresponding mode of DNA recognition . The bZIP binding motifs were also automatically split into two clusters , identical to the ( manually ) classified JASPAR FBPs: one with the CREB-like and one with the C/EBP-like proteins . We note the striking similarity between the bZIP/CREB and the nuclear receptor binding patterns . Still , since the only base position they differ in is one of high information content , our clustering method was able to distinguish between the two patterns . The HMG proteins are represented by three protein families that bind chromosomal DNA . The two families represented in the JASPAR database are HMGA/HMGI/Y and HMGB/SOX/SRY , whereas the HMGN family is not represented . The HMGA proteins are members of the AT-hook family of TFs [26] . The HMGB proteins are structurally distinct HMG-box proteins [27] . Both families prefer to bind to AT-rich sequences in the minor groove with low selectivity . This is probably the reason that our algorithm clustered them together with the Forkhead family , which also binds to AT-rich sequences , but in the major groove . There is no structural similarity among these classes of proteins , and the mode of their interaction with the DNA suggests that the target similarity is coincidental . In fact , the two motifs are not identical , but they show significant overlap in four highly informative nucleotide positions ( consensus: AACA ) ( Figure 9 ) . Nevertheless , for those that use the FBPs for predicting the TF that binds to a given DNA motif , this provides an example where generating individual FBPs might lead to misclassification due to coincidental target similarity . Thus , for prediction purposes , we propose to keep these two families in the same cluster ( FBP ) , since distinguishing between the two may be difficult ( Figure 9 ) . Notably , our algorithm identified another cluster composed of members of both families , suggesting there is a relationship between their motifs . Both these clusters contain members of the HMGB subgroup . The only member of the distinct HMGA subgroup , HMG-I/Y , clusters together with a homeodomain protein , Pax4 . Pax proteins have two covalently linked HTH domains separated by a long linker and they also bind AT-rich sequences . The HTHs bind to 5-bp and 6-bp recognition sequences with an interpositioned 6-bp spacer that interacts with the linker [28] . This explains the long recognition sequence revealed in the FBP in this cluster . When the two zinc-finger families were included in the analysis , the overall structure of the tree and the clusters remained the same , pointing to the stability of our multiple alignment and clustering algorithm . Most of the GATA proteins formed a new cluster , whereas all DOF proteins joined the cluster of the two IRF proteins . This is because the DOF consensus target sequence ( AAAG ) is part of the IRF motif ( Figure 8 ) . The cross-validation results in the extended family tree/clustering remained very high ( 91% compared with 94% in the smaller tree ) . The STAMP platform , introduced in this study , contains all tested algorithms and can be efficiently used in BLAST-like searches against a database of PSSM models . Various datasets ( including TRANSFAC and JASPAR motifs ) are currently supported . In the future , it would be useful to incorporate other similarity metrics , alignment methods , and tree-building algorithms into the platform in order to allow for further exploration of optimal methods . Note , however , that it may not be possible to implement all of the known tree-building algorithms for motif alignment . Other distance-based methods ( such as neighbour-joining [29] ) rely heavily on additivity of the distance metric , which was not possible to define using our comparison metrics . Parsimony-based methods [30] rely on the estimation of substitution rates between sites , which is also not easily definable for frequency matrices . However , a substitution matrix has recently been defined for DNA-binding consensus sequences [31] , so application of alternative tree-building methods may yet be possible in the DNA-binding motif domain ( albeit not for PSSMs per se ) . In summary , we expect that the methods and the results described in this study will facilitate the exploration of DNA-binding preference evolution amongst related transcription factors and will have a significant impact in many areas of gene research .
The structures we used in Figure 8 have the following Protein Data Bank ( http://www . pdb . org ) accession numbers: 9ANT , 2EZD , 1HRY , 1IF1 , 1DH3 , 1BC8 , 2NLL , 1H88 , 1PDN , 1SVC , 1SRS , 1MDY , 1AN2 . The HNF3_Mod structure was provided by Kirk L . Clark and Stephen K . Burley ( personal communication ) .
|
Transcription factors are primary regulators of gene expression . They usually recognize short DNA sequences in gene promoters and subsequently alter their transcription rate . It is known that structurally related transcription factors often recognize similar DNA-binding patterns ( or motifs ) . Comparison of these motifs not only provides insights into the evolutionary process they undergo , but it also has many important practical applications . For example , motifs that are found to be “similar” can be combined to form generalized profiles , which can be used to improve our ability to predict novel DNA signals in the promoters of co-expressed genes , and thus facilitate a more accurate mapping of gene-regulatory networks . However , to date there is no comprehensive platform that will allow for an efficient analysis of DNA motifs . Furthermore , the efficiency of the methods used to assign similarity between DNA motifs has not been thoroughly tested . This paper takes an important first step towards this goal by evaluating available comparison strategies as applied to DNA motifs and by generating an improved familial profile dataset .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Supporting",
"Information"
] |
[
"eukaryotes",
"computational",
"biology"
] |
2007
|
DNA Familial Binding Profiles Made Easy: Comparison of Various Motif Alignment and Clustering Strategies
|
Invertebrate stages of Leishmania are capable of genetic exchange during their extracellular growth and development in the sand fly vector . Here we explore two variables: the ability of diverse L . major strains from across its natural range to undergo mating in pairwise tests; and the timing of the appearance of hybrids and their developmental stage associations within both natural ( Phlebotomus duboscqi ) and unnatural ( Lutzomyia longipalpis ) sand fly vectors . Following co-infection of flies with parental lines bearing independent drug markers , doubly-drug resistant hybrid progeny were selected , from which 96 clonal lines were analyzed for DNA content and genotyped for parent alleles at 4–6 unlinked nuclear loci as well as the maxicircle DNA . As seen previously , the majority of hybrids showed ‘2n’ DNA contents , but with a significant number of ‘3n’ and one ‘4n’ offspring . In the natural vector , 97% of the nuclear loci showed both parental alleles; however , 3% ( 4/150 ) showed only one parental allele . In the unnatural vector , the frequency of uniparental inheritance rose to 10% ( 27/275 ) . We attribute this to loss of heterozygosity after mating , most likely arising from aneuploidy which is both common and temporally variable in Leishmania . As seen previously , only uniparental inheritance of maxicircle kDNA was observed . Hybrids were recovered at similar efficiencies in all pairwise crosses tested , suggesting that L . major lacks detectable ‘mating types’ that limit free genetic exchange . In the natural vector , comparisons of the timing of hybrid formation with the presence of developmental stages suggest nectomonads as the most likely sexually competent stage , with hybrids emerging well before the first appearance of metacyclic promastigotes . These studies provide an important perspective on the prevalence of genetic exchange in natural populations of L . major and a guide for experimental studies to understand the biology of mating .
Protozoan parasites of the genus Leishmania present a remarkable epidemiologic and clinical diversity , producing a spectrum of human diseases that pose serious health concerns throughout tropical and sub-tropical regions . Leishmania have a dimorphic asexual life cycle consisting of extracellular promastigotes that multiply and develop within the alimentary tract of the sand fly vector , and intracellular amastigotes that multiply within the phagolysosomes of their host macrophages . No sexual dimorphism has been described , and based on the strong linkage disequilibrium revealed in population genetic studies applied to several Leishmania species , these parasites have been argued to be essentially clonal [1] , [2] . This conclusion , however , must contend with a series of observations regarding the analysis of multilocus microsatellite , isoenzyme , or karyotype markers consistent with occasional sexual recombination and possible endogamy [3]–[8] . Many of these studies incorporated models that assumed Leishmania were ‘normal diploids’ , which is now known to be untrue . Most and probably all Leishmania strains show varying degrees of aneuploidy ( reviewed in [9] , [10] ) . Moreover , chromosome number variation appears to be highly variable in culture , with individual cells/chromosomes showing monosomic , disomic and trisomic chromosome numbers [11] . The rapidity and transient nature of variable segregation , loss and expansion of chromosomes numbers are expected to have a profound effect on levels of heterozygosity , and to date the impact of rapidity and frequency of these factors on Leishmania has not been adequately accounted for by the population genetic models in the studies cited earlier . Nonetheless , it is clear from the literature that in natural populations hybrid parasites have been observed , and in some cases are widespread [12] . Two recent reports involving experimental crosses in sand flies have provided the first direct demonstration that genetic exchange in Leishmania can occur . The first series of successful crosses involved co-infection of Phlebotomus duboscqi sand flies using two strains of L . major engineered to express heterologous drug selectable markers [13] . The doubly drug-resistant clones obtained were clear genomic hybrids , based on the presence of both parental alleles at seven unlinked loci . Importantly , the same parental lines that successfully mated in the sand fly midgut failed to generate hybrid genotypes following their co-inoculation and high density growth in axenic culture or in the mouse ear dermis . From these data it was concluded that L . major , like Trypanosoma brucei [14] , is capable of a sexual cycle consistent with a meiotic process , and that sex is confined to the extracellular stages of the parasite developing within the insect vector . A subsequent study by Sadlova et al . [15] used a fluorescent protein detection system to observe yellow promastigotes in P . perniciosus and Lutzomyia longipalpis midguts co-infected with RFP and GFP transgenic lines of L . donovani . The putative hybrids were observed in only two of the hundreds of co-infected flies examined , and the parasites could not be recovered and propagated to confirm their hybrid genotypes . A number of essential features of genetic exchange in Leishmania remain to be further substantiated or explored , including the sexual competency of diverse Leishmania species and strains , the timing and frequency of hybrid formation in the midgut , their developmental stage associations , and the ability of other phlebotomine vectors to support genetic exchange . In the current studies , we have expanded the analysis of the mating competency of L . major strains to include pairwise matings of multiple isolates distributed over the full geographic range of this species . Our genotype analysis of a large number of progeny clones substantiates their chromosomal inheritance of both parental alleles at multiple unlinked loci , consistent with a meiotic process , and their uniparental inheritance of kinetoplast DNA . A low but unexpected frequency of nuclear loci were also inherited from only one parent , suggesting loss of heterozygosity . The timing of hybrid recovery in a natural vector suggests that nectomonad promastigotes are the most likely mating competent forms . Finally , we have demonstrated the capacity of a new world vector , L . longipalpis , to efficiently support the Leishmania sexual cycle .
Four different L . major strains derived from primary isolates originating in Senegal , Israel , Iraq , and southern Russia , and stably transfected with various antibiotic resistance markers , were tested for their ability to generate hybrid genotypes following co-infection of P . duboscqi sand flies , a natural vector of L . major transmission in Africa . In these experiments a HYG marker was integrated into the LPG5A locus , a SAT marker was integrated into one cistron within the rRNA locus , and the BSD marker was integrated into the LPG5B locus , as described previously [13] , [16] . These loci were chosen as prior studies revealed no effects on parasite growth in vitro or in sand flies ( [13] and data not shown ) . For all parental lines , a number of clonal lines were obtained and screened for their ability to show survival and complete development in P . duboscqi . Mating tests were performed by mixing equal numbers of parental lines , and feeding in vivo as described previously [13] . At varying times following the infective feed , midguts were dissected , parasites recovered , and cultured in media selective for the doubly-drug resistant hybrids . Flies cannot be maintained aseptically , and a variable proportion of the cultures established from the dissected midgut homogenates was lost to either fungal or bacterial contamination . The frequency of hybrid recovery was calculated as the percentage of the number of ‘clean’ midguts yielding doubly-drug resistant promastigotes . For each infected midgut yielding a doubly-drug resistant population , only a single clone was selected for further analysis . All lines studied are therefore a product of an independent mating event . In all selected clones , the presence of both parental drug markers was confirmed by PCR using primers specific for the respective antibiotic resistance markers ( Table S1 ) ( Fig S1 ) . SNP typing of loci present on chromosomes unlinked to the integrated markers on the parental lines was performed subsequently . Our initial crossing experiments employed the same L . major parental lines , Fn/Sat and Lv39c5/Hyg , used to generate the set of hybrids originally reported [13] , in order to explore the timing of hybrid formation during L . major growth and development in P . duboscqi . Eight hybrids were recovered from the total of 112 clean midguts dissected on days 5 , 9 and 13 post-infection ( p . i . ) ( Table 1 ) . Importantly , no hybrids were recovered from day 5 midguts , while 3 hybrids were recovered from day 9 ( 8 . 6% ) , and 5 from day 13 midguts ( 13 . 9% ) . Two additional timing studies were carried out using Fn/Sat paired with a new parental line , Sd/Hyg . In total , 15 hybrids were recovered amongst the 215 clean midguts dissected on days 4–15 . Again , hybrids were recovered only at later time points , with no hybrids recovered on day 4 and only a single hybrid recovered on day 5 ( 2 . 4% ) . Flies dissected on days 6–8 yielded 7 hybrids ( 7 . 0% ) , while flies dissected on days 9–15 yielded 11 hybrids ( 10 . 3% ) . Thus , at least for mating pairs involving the Fn/Sat parental line for which the kinetics of hybrid formation in P . duboscqi was investigated , there was a significant correlation between the probability of hybrid recovery and the number of days of co-infection ( Fn/Sat×Lv39/Hyg cross , p = 0 . 024; Fn/Sat×Sd/Hyg cross , p = 0 . 034; Fig S2 ) . As all of the successful mating attempts up to this point employed the Fn/Sat line as one of the parents , additional crosses were undertaken in P . duboscqi that did not involve the Fn strain . From two independent crosses involving LV39c5/Hyg and Sd/Bsd , we were able to recover 4 hybrids ( 6 . 7% ) in flies dissected on days 8–10 p . i . A single experiment in flies co-infected with Lv39/Hyg and Ry/Sat yielded 5 hybrids ( 14 . 3% ) from flies dissected day 10 p . i . A summary of all of the new crosses that we have undertaken in P . duboscqi involving the various pairwise combinations of the 4 L . major strains is presented in Table 2 . For the sake of comparison , the summary table includes only flies that were co-infected for at least 8 days prior to dissection . Notably , the fraction of flies from which hybrids were recovered , ranging from 6 . 7–14 . 3% , did not vary significantly between crosses involving the different strains , including Fn vs both LV39c5 and Sd , or LV39c5 vs Sd , Ry and Fn . The findings suggest that L . major lacks detectable ‘mating types’ that might limit free genetic exchange . We addressed the question of whether there is a correlation between the intensity of infection in the midgut and the probability of hybrid recovery , by comparing the total number of promastigotes per gut at the time of dissection in flies yielding hybrids or not ( Fig . 1 ) . There was no significant difference between these groups . We were struck by the low infection levels even at late time points in many of the flies yielding hybrids ( <10 , 000 promastigotes/gut ) . In addition to infection levels , the association between hybrid formation and promastigote developmental stages was explored . During cyclical transmission in a competent vector , Leishmania promastigotes undergo a sequence of morphologic changes ( reviewed in [17] ) . For L . major , these distinct forms typically include procyclics , which appear as short , ovoid , slightly motile and rapidly dividing cells that develop in the abdominal midgut during the first 72 hr . prior to their transformation to long , slender promastigotes , termed nectomonads , 3–6 days post-feeding . These forms fill the abdominal midgut , with many becoming attached by their flagellum to the microvillar lining . By 4–5 days , most of the digested bloodmeal is excreted and nectomonads can begin to be found in the thoracic midgut . This forward migration is accompanied by their gradual transformation over the following week to shorter , broader haptomonads attached to the intima of the stomodeal valve , as well as short , slender , highly active , free swimming and non-dividing metacyclic promastigotes that are the infective stage egested by the fly . The proportion of these developmental forms at the time of dissection in each of the midguts from which hybrids were recovered from crosses involving Fn/Sat×Sd/Hyg or Fn/Sat×Lv39/Hyg is shown in Fig . 2 . Even at the earliest time of hybrid recovery in a single fly on day 5 , procyclics were no longer present and thus are not shown in the figure . This midgut , as well as 4 of the 5 midguts from which hybrids were recovered on days 6–8 , was dominated by nectomonads , and 3 had exclusively these forms . The transition to infections dominated by haptomonad and metacylic promastigotes was observed in one midgut by day 8 , but was otherwise delayed until day 9 . By day 13 , metacyclics were the predominant form in the midgut . Importantly , in 6 of the hybrid positive midguts dissected on days 5–9 , no metacyclics were observed , indicating that metacyclogenesis is not a prerequisite for mating to occur . The increased frequency of hybrid recovery in flies with mature infections might reflect the greater opportunity that nectomonad-related forms have to mate , particularly as they continue to replicate and become more densely packed in the microenvironment of the anterior midgut . However , the contribution of late developmental stages , including metacyclics , to hybrid formation , cannot be discounted . As referred to above , PCR tests confirmed that in each case the progeny clones were true genetic hybrids bearing the parental selectable markers ( Fig . S1 ) . We also examined the segregation of loci not linked to the chromosomes bearing the drug resistance markers using SNPs identified previously [13] from comparisons of the SCG genes located on chromosomes 2 , 7 , 21 , 25 , 31 , 35 , and 36 . Other nuclear markers were chosen based on SNPs previously identified by comparing different L . donovani isolates [18] and that we found to also identify allelic differences between some of the L . major sub-strains on chromosomes 14 , 34 . The remaining SNPs on chromosomes 4 , 9 , and 31 were identified by comparing gene alignments determined from the genome data bases of other Leishmania species , including L . major , L . infantum , L . braziliensis , and L . mexicana . The position and description of each SNP , and the primers and conditions used for their amplification , are summarized in Tables S1 and S2 . The SNPs present in the different parents and their hybrid progeny were analyzed by direct sequencing ( Fig S3 ) , and are summarized for each of the crosses in Tables 3–6 . Both parents were homozygous for every marker applied to the analysis of the hybrids generated by those particular parents . Of the 18 Fn/Sat×Sd/Hyg hybrids recovered , 15 were heterozygous at all 4 loci analyzed , while 2 were heterozygous at 3 loci , with the 4th marker sequence not determined ( Table 3 ) . Interestingly , homozygosity was found for the FnSd1 hybrid for the marker on chromosome 35 , and for the FnSd7d hybrid for the marker on chromosome 14 . For the 8 hybrids generated from the Fn/Sat×Lv39c5/Hyg cross , heterozygosity was observed at all loci for which sequences were available , with at least 3 and in some cases all 5 of the loci showing both parental alleles ( Table 4 ) . For the 4 LV39c5/Hyg×Sd/Bsd hybrids , 3 were heterozygous at 5 of the loci analyzed , while the LSD2 hybrid was heterozygous at 4 loci , but was homozygous at the markers on chromosomes 9 and 31 ( Table 5 ) . Interestingly , the inheritance patterns at these two loci were from different parents . Finally , in the cross of Ry/Sat×LV39c5/Hyg , all 5 hybrids were heterozygous at all 5 loci analyzed ( Table 6 ) . In summary , taking into account both the drug-selectable and additional unlinked markers , all 35 progeny clones inherited both parental alleles present in at least 5 loci located on 5 different chromosomes , and 97% of the nuclear markers analyzed were heterozygous for these alleles . The progeny thus appear to be full genomic hybrids , with the caveat that the plasticity in chromosome number ( or other process ) might in some instances have resulted in loss of heterozygosity ( LOH ) , reflected in the apparent uniparental inheritance at a few ( <3% ) of the marker loci analyzed . In our previous study LOH was not observed , most likely arising from the fact that fewer hybrids were studied . These conclusions are supported in preliminary genome-wide analysis of SNP inheritance , as deduced by whole genome deep sequencing ( J . Shaik , N . Akopyants , D . Dobson , P . Lawyer , D . Elnaiem , D . Sacks , and S . Beverley , in preparation ) . Figure 3 and Figure S4 show an analysis of the inheritance of homogozygous parental SNP for L . major chromosome 17 , showing biparental inheritance of 1314 SNPs for 5 representative hybrids . The inheritance of the kinetoplast organelle in the hybrids was examined by analyzing the segregation of polymorphic regions within the maxicircle kDNA [9] , [10] ( Fig S5 ) . In contrast to the inheritance of nuclear DNA , and consistent with our previous findings , maxicircle kDNA was inherited from a single parent ( Tables 3–6 ) . For the Fn/Sat×Sd/Hyg hybrids , 13 inherited both of the maxicircle kDNA markers analyzed from Fn , while 5 inherited the kDNA markers from Sd . The departure of this ratio from a random 50∶50 segregation is of borderline statistical significance ( p = 0 . 096; exact two-sided binomial test ) . For the 8 hybrids generated in the Fn/Sat×Lv39c5/Hyg cross , all inherited both Fn kDNA markers . For the 5 Ry/Sat×Lv39c5/Hyg hybrids , all inherited both kDNA allelic markers from the Ry parent . Finally , only one kDNA marker was analyzed for the Lv39c5/Hyg×Sd/Bsd hybrids , with 3 inheriting the Sd marker and one the Lv39 marker . DNA content analysis of the progeny clones revealed that while the majority resembled the ‘2n’ parents ( recalling that while Leishmania chromosomes are predominantly diploid , aneuploidy is quite common ) , a number of polyploid progeny were observed , including a total of 6 ‘triploid’ , and one ‘tetraploid’ hybrid ( Fig . S6 ) . DNA contents of intermediate amounts were not observed , although small differences cannot be ruled out . The ‘triploid’ clones remained ‘triploid’ following extensive serial in vitro passage and following recovery from mouse dermal lesions , whereas the single ‘tetraploid’ progeny clone remained ‘tetraploid’ in vitro , but only diploid cells that had lost their double drug resistance were recovered from infected mouse tissue . We used SNP genotyping combined with cleaved amplified polymorphic site ( SNP-CAPS ) analysis performed as previously described [13] , to determine the relative contribution of each parental genome in the triploid progeny . The analysis is based on a SNP in Glucose 6 phosphate dehydrogenase ( G6PD-LmjF34 . 0080 ) in position 975 ( Table S2 ) that results in a SACII restriction site in both Lv39/Hyg and Sd/Hyg and not in Fn/Sat . The bar graph in Fig S7 shows the ratio between the intensity of the uncut upper band from Fn/Sat and the middle band from either Lv39/Hyg or Sd/Hyg , and compares the various hybrids with controls generated by mixing the respective parental DNA at 1∶1 , 2∶1 or 1∶2 ratios . The SNP-CAPS analysis confirms the direct sequencing results indicating that all of the progeny clones examined are allelic hybrids for this marker , and suggests that for the triploid progeny , the extra genome complement is inherited from Lv39/Hyg ( Fig S7A ) or from Sd/Hyg ( Fig S7B ) , and in no instance from Fn/Sat , similar to the triploid hybrids previously described [13] . To confirm that mating in Leishmania can occur in another vector species , we pursued co-infections experiments in Lu . longipalpis , belonging to the genus Lutozomyia , and the chief vector of L . infantum transmission in the new world . The experiments were done using the two L . major strains Fn/Sat and Lv39c5/Hyg , which we know to be mating competent in P . duboscqi ( [13]and this work ) . Although Lu . longipalpis is not a natural vector of L . major , it is permissive for the complete development of L . major [19] , including the Fn/Sat and Lv39/Hyg lines ( data not shown ) . An initial series of co-infection experiments were analyzed only at late time points ( days 11–18 ) to optimize the chances of hybrid recovery ( Table 7 ) . In three experiments , a total of 10 doubly drug-resistant hybrids were recovered from 8–15% of the clean Lu . longipalpis midguts , similar to the percentage of hybrid recovery in P . duboscqi ( Table 2 ) . A kinetic experiment was undertaken to determine if the timing of hybrid formation might also be similar . Surprisingly , flies yielding hybrids were recovered in this experiment at a remarkably high frequency at all time points examined ( 40–65% ) , including already at day 3 . A total of 51 doubly drug-resistant lines were established in this single kinetic study . Another striking feature of this experiment was the virtual absence of fungal or bacterial contaminants in the cultures established from the dissected midguts . The 61 hybrids generated by the Fn/Sat×Lv39c5/Hyg crosses in Lu . longipalpis were cloned and genotyped using the markers previously described [13] . In addition to the drug resistance markers that were amplified by PCR , five different SNP markers on chromosomes 2 , 7 , 25 , 35 and 36 and two SNP markers on maxicircle kDNA were analyzed , in this instance by SNP-CAPS analysis ( Table 8 ) . All of the hybrids carried both drug resistance markers , and the majority inherited both parental alleles at all 5 chromosomal markers , with heterozygous alleles present at 90% of the nuclear markers analyzed . Ten of the hybrids , however , were homozygous at 2 , 3 or 4 of the 5 chromosomal marker loci analyzed . While for the kDNA markers uniparental inheritance was expected , it was surprising to find that all 61 of the hybrids had inherited their maxicircle kDNA from only the Fn parent . Finally , they all appeared to be ‘2n’ , which again distinguishes the hybrids generated in these crosses from those arising in the natural vector P . duboscqi in this and especially the previous study [13] . The midguts analyzed in the kinetic study of hybrid formation in Lu . longipalpis were also used to determine the infection intensity and developmental stage associations in flies yielding hybrids or not . No significant difference in the number of promastigotes per midgut was observed between these groups , and hybrids were again recovered from many flies that had very low numbers of parasites , even at late time points ( Fig . 4 ) . There was also no difference at any time point between the hybrid positive and negative flies in the average number of each developmental stage ( data not shown ) . Fig . 5 shows the relative proportion of each developmental stage at the time of dissection in the individual midguts yielding a hybrid . Because hybrids were recovered as early as day 3 , procyclic forms were still present and may have been involved in hybrid formation . Multiple attempts in subsequent experiments to investigate hybrid recovery at days 1–2 when procyclics would have been the predominant if not exclusive stage in the gut , were negative . These experiments remain inconclusive , however , as the frequency of hybrid recovery even at late time points was very low ( <5% ) , while the frequency of bacterial overgrowth in the midgut homogenate cultures was especially high ( >80% ) ( data not shown ) . In the experiment shown in fig . 4 , nectomonads were already the predominant stage in the majority of flies by day 3 , consistent with prior observations regarding the accelerated progression of promastigote stage differentiation in Lu . longipalpis as compared to a natural vector [19] , [20] . At day 7 , nectomonads were still the prevailing form , and were found in all hybrid positive midguts at day 10 , though in proportionately less numbers than haptomonads and metacyclics in most flies . Since the majority of the flies yielding hybrids at days 3 and 7 contained no metacyclics , the data again argue that metacyclogenesis is not required for mating competent forms to arise .
In this study we have extended and confirmed our original report regarding the capacity of Leishmania to undergo genetic exchange in the sand fly vector , analyzing in detail 96 independent hybrids . Using a system in which P . duboscqi sand flies , a natural vector of L . major transmission in West Africa , were co-infected with L . major strains expressing distinct drug selectable markers , we could confirm the mating capacity of strains originating throughout the geographic range of this species , from Senegal , West Africa to Israel , Iraq , and southern Russia . Hybrids were generated between all pairwise combinations tested , suggesting that barriers to mating between strains do not exist ( or at least are not widespread ) . The proportion of flies yielding at least one hybrid was roughly comparable amongst the 4 mating pairs tested , and consistent with the estimates of a relatively low frequency of hybrid formation ( <104 ) described previously [13] . Mating must be considered as a non-obligatory part of the infectious cycle , since the majority of co-infected flies failed to yield a hybrid , yet were still permissive to the development of mature metacyclic forms . The results involving these geographically diverse L . major stocks , selected without bias , confirm that even if sexual reproduction is not obligatory and relatively rare , the machinery for genetic exchange has been maintained across the species . Additionally , we demonstrated the capacity of a new world phlebotomine vector , Lu . longipalpis , to support genetic exchange in Leishmania . Since genetic exchange has not been observed for promastigotes growing in vitro , the findings reinforce the conclusion that the phlebotomine midgut provides a unique environment for sex involving the insect stages of the parasite to occur . For all of the hybrid progeny selected for double antibiotic resistance , the inheritance of both parental selectable markers was confirmed by PCR tests . Analysis of SNPs present on 4–6 additional chromosomes not linked to the selectable markers revealed that at 97% of the loci tested , progeny arising from crosses in the natural vector P . duboscqi inherited both parental alleles . However , 3% ( 4/150 ) loci showed loss of heterozygosity ( LOH ) , inheriting only one allele from either parent ( Tables 3–6 ) . Curiously , taking into account both our current and prior studies involving the pairing of parental lines Fn/Sat and Lv39c5/Hyg in P . duboscqi , we found no examples of LOH ( 0/160 loci tested ) , whereas crosses of these same parents in the unnatural vector Lu longipalpis reported here , the frequency of LOH was 10% ( 27/275 ) . These data suggest that the frequency of LOH may vary in individual crosses depending on both the parental lines and sand fly host . There are a number of mechanisms that could account for LOH – for example , meiotic or mitotic crossing over , or chromosomal segregation following mating associated with the rapidity by which Leishmania shows alterations in aneuploidy in culture [11] , [21] . However , the overwhelming fraction of loci/hybrids show biparental inheritance . Correspondingly , , preliminary next gen sequencing data of the hybrids ( J . Shaik , N . Akopyants , D . Dobson , P . Lawyer , D . Elnaiem , D . Sacks , and S . Beverley , in preparation ) , suggest that all hybrids recovered thus far arise biparentally across the entire genome ( Fig . 3 , S4 ) . . By contrast , maxicircle kDNA was again inherited in a uniparental manner , with a clear bias towards inheritance of kDNA from the Fn parent in crosses involving that parent . We have so far not been able to explore the possibility , documented for T . brucei [22] , that recovery of hybrids during the earliest stage of their generation in the midgut might reveal biparental inheritance of kDNA , with subsequent loss during mitotic division . Analysis of total DNA again revealed that while the majority of hybrids were ‘2n’ , a number of ‘3n’ progeny were observed . We also recovered for the first time a single ‘4n’ hybrid ( arising from a Fn/Sat×Sd/Hyg cross ) . Interestingly , while maintained stably as ‘4n’ in culture , this line reverted to ‘2n’ following recovery from an infected mouse . Tetraploid progeny had previously been seen in Leishmania , arising from attempts to inactivate essential genes [21] . In that study only one of 5 tetraploids tested remained infective to mice , however , following recovery this line remained tetraploid . Given the limited number of tetraploids obtained and examined thus far , further work will be required to systematically assess the stability and virulence of tetraploids in Leishmania . In the studies here and previously reported [13] , DNA contents of intermediate contents were not observed , although changes involving aneuploidy of a small number of chromosomes would not have been revealed by the methods used . These data could imply that meiosis occurs prior to or after a fusion event involving ‘1n’ cells or nuclei , with ’3n’ progeny resulting from the absence or incomplete meiotic division of one of the diploid parents . The ‘4n’ hybrid is most readily explained by fusion of diploid cells , and may represent an intermediate sexual stage that undergoes reduction to the diploid state , similar to the process described for Candida albicans [23] . Polyploid hybrids have been recovered in African trypanosomes , and inferred for American trypanosomes assuming T . cruzi undergoes reduction to an aneuploid state via a parasexual process [24] . For T . brucei , Mendelian inheritance patterns are well-established [25] and an epimastigote life-cycle stage residing in the tsetse fly salivary glands has recently been identified expressing meiotic genes [26] . Importantly , the expression of meiosis-specific proteins in most cases occurred prior to cell fusion , establishing a precedent in excavate protists for conventional meiotic division , although it should be noted that haploid trypanosomes have still not been formally demonstrated in T . brucei . The meiosis specific orthologs that were used to detect meiotic forms in T . brucei are also present in the L . major genome [27] , and we are currently investigating their expression during development of midgut forms in vivo . The second major finding arising from our studies pertains to the timing and developmental stage associations of hybrid formation in P . duboscqi sand flies . In 3 independent mating experiments investigating the timing of hybrid generation , only 1 hybrid was recovered amongst the 100 co-infected midguts dissected prior to day 6 ( 1% ) . Ten hybrids were recovered amongst the 134 guts dissected on days 6–9 ( 7 . 5% ) , and 15 of 112 ( 13 . 4% ) on days 10–15 . The delay in hybrid recovery strongly suggests that the rapidly dividing procyclic forms predominating in the midgut during days 1–2 post-infection are unlikely to be mating competent , especially as these forms have completely transitioned into nectomonads by days 4–5 . The increased frequency of hybrid recovery on days 6–9 coincides with the colonization of the posterior midgut by primarily nectomonad forms and their initial migration to the thoracic midgut . The highest rate of hybrid recovery observed at later time points ( days 10–15 ) might implicate more mature stages , e . g . haptomonads or metacyclics , as mating competent forms , or simply reflect the greater opportunity that nectomonads that are maintained in variable numbers even at late time points might have to mate , especially as they become more densely packed in the anterior gut . Since the flies from which hybrids were recovered days 5–9 post-infection harbored few and in most cases no metacyclics , it can be concluded that metacyclogenesis is not a pre-requisite for mating to occur . The intensity of infections at the time of dissection was also not a good predictor of mating success , as the total number of promastigotes per midgut was no higher in flies yielding hybrids compared to those that did not , and in some cases hybrids were recovered from flies that had very light infections ( <2000 promastigotes/midgut ) . In summary , neither the earliest dividing procyclic forms , nor the late developing , resting metacyclic forms , appear to be associated , or uniquely so , with hybrid formation , consistent with the evidence against mating involving these analogous life cycle stages in T . brucei [26] , [28] , [29] . We favor a role for nectomonad-related forms in mating based not only on the timing of hybrid recovery , but also the fact that these forms are unique to L . major stage differentiation in vivo and are not observed amongst the pleopmorphic promastigotes that appear during growth in culture , which have so far remained mating incompetent . Nectomonads also share with T . brucei epimastigotes the ability of their flagellum to mediate attachment to the substratum , which involves adhesion to the midgut epithelium for Leishmania , and to the salivary gland epithelium for African trypanosomes . By analogy with the process of gamete activation and fusion in other flagellates , e . g . Chlamydomonas [30] , such junctional complexes may inititate the signaling events that are required for hybrid formation in Leishmania and African trypanosomes to occur [29] . Despite the fact that Lu . longipalpis , which is a natural vector of L . infantum in the new world , does not naturally transmit L . major , it is permissive to the full development of this species in the laboratory [19] . The current studies confirm that it also supports genetic exchange in L . major . Genotype analysis of 61 hybrids revealed that in each case they were full genomic hybrids , since they possessed both parental alleles at unlinked loci on 7 different chromosomes , but again with the caveat that heterozygosity appears to have been lost at a number of loci . The inheritance bias of Fn maxicircle kDNA observed in hybrids generated in P . duboscqi was for the L . longipalpis hybrids complete , with all 61 progeny clones inheriting maxicircle kDNA exclusively from the Fn parent . The mechanisms that might account for the elimination of one of the kinetoplast organelles pre- or post-fusion , and how this process can be so selective , are currently unknown . In an initial series of experiments involving flies dissected only at late time points , the efficiency of hybrid recovery was roughly comparable to that observed in P . duboscqi , with an average of 15% of flies with mature infections yielding at least one hybrid progeny . When a time course study was carried out in Lu . longipalpis in order to determine when hybrid formation might begin to occur in these flies , we were surprised to recover hybrids from a high proportion of flies ( 45% ) already by day 3 . This remarkable rate of hybrid recovery was maintained or increased at the later time points . In subsequent experiments involving different populations of released adults from the same colony , we failed to observe such a high rate of hybrid recovery ( data not shown ) . While we cannot explain what remains so far a unique experience involving these flies , we note that the released adults used in the successful time course experiment were unusual in the virtual absence of any midguts harboring commensal bacteria that overgrew the promastigote growth medium containing penicillin and streptomycin . The possible influence of the midgut microbiota on promoting or inhibiting hybrid formation is currently being investigated . Regarding the early appearance of hybrids by day 3 , we and others [19] , [20] have observed that the sequence of promastigote stage differentiation proceeds more rapidly in Lu . longipalpis compared with P . duboscqi or P . papatasi , with nectomonads the dominant stage already by day 3 , and fully mature metacyclics the predominant stage already by day 8 . Nonetheless , we cannot rule out the contribution of the earliest procylic promastigote stage to hybrid formation in Lu . longipalpis , which would be consistent with the timing and morphology of the putative L . donovani hybrid observed in the single Lu . longipalpis fly reported by Sadlova et al . [15] . In summary , our analysis of a large number of hybrid progeny generated between multiple strains of L . major substantiates that sex is a normal aspect of their reproductive strategy . The timing of hybrid recovery in a natural vector implicates the involvement of nectomonad forms in hybrid formation . Our finding that Leishmania crosses can be achieved experimentally in a proven vector of the genus Luztomyia , is highly relevant to a number of population studies showing the widespread occurrence of hybrid genotypes in New World Leishmania [3]–[8] .
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The protocol was approved by the Animal Care and Use Committee of the NIAID , NIH ( protocol number LPD 68E ) . Invertebrates are not covered under NIH guidelines . L . major Fn/Sat was derived from a strain ( MHOM/IL/80/Friedlin ) originally isolated from a patient with cutaneous Leishmaniasis acquired in the Jordan Valley , and contains a heterozygous nourseothricin–resistance ( SAT ) marker , integrated along with a linked firefly luciferase reporter gene into one allele of the ∼20 rRNA cistrons located on chromosome 27 [31] using constructs and methods described previously [13]; L . major LV39c5/Hyg was derived from a strain originally isolated from a reservoir rodent host in southern Russia ( MHOM/Sv/59/P ) [32] and is heterozygous for an allelic replacement of the LPG5A on chromosome 24 by a hygromycin B–resistance cassette [16]; L . major Sd/Hyg was derived from a strain isolated from a patient with cutaneous lesions acquired in Senegal ( MHOM/SN/74/SD ) [33] , and is also heterozygous for an allelic replacement of the LPG5A on chromosome 24 by a hygromycin B–resistance cassette [16]; L . major Sd/Bsd is derived from the same Sd strain from Senegal and is heterozygous for allelic replacement of the LPG5B on chromosome 18 by a blastocidin-resistance cassette [16]; L . major Ry/Sat was derived from a strain originally isolated from a lesion biopsy of a laboratory worker accidentally exposed to sand flies that were experimentally infected with a strain of L . major ( WR2885 ) originating in Iraq [34] , and also contains a heterozygous SAT–resistance marker , integrated into one allele of the rRNA cistrons located on chromosome 27 using constructs and methods described previously [13] . All parasites were grown at 26°C in complete medium 199 ( CM199 ) supplemented with 20% heat-inactivated FCS , 100 U/ml penicillin , 100 µg/ml streptomycin , 2 mM L-glutamine , 40 mM Hepes , 0 . 1 mM adenine ( in 50 mM Hepes ) , 5 mg/ml hemin ( in 50% triethanolamine ) , and 1 mg/ml 6-biotin , and containing either 25 ug/ml blasticidin S ( Invitrogen , Carlsbad , CA ) , 25 ug/ml hygromycin B ( EMD Biosciences , San Diego , CA ) or 100 ug/ml SAT ( Jena Bioscience , Germany ) , or combinations of these antibiotics as necessary . Three to five day old female P . duboscqi or L . longipalpis sand flies , obtained from colonies initiated from field specimens collected in Mali and Brazil , respectively , were fed through a chick skin membrane on heparinized mouse blood containing 1–4×106/ml logarithmic phase promastigotes of each parental line of L . major . At various days after the infective feed ( 3–18 days ) , the sand flies were anesthetized with CO2 , their midguts dissected and homogenized in 100 µl CM199 , and the number of viable promastigotes and the percentage of each promastigote developmental stage were determined by counting under a hemacytometer . Promastigote stages were identified based on the morphologic criteria previously described [35] . For isolation of hybrids , individual midgut homogenates from co-infected flies were transferred to a single well of a 96 well flat bottom plate , and incubated at 26°C over night . An equal volume of CM199 containing 2× of the appropriate antibiotics was added on the following day for the selection of hybrids . Doubly drug-resistant promastigotes were cloned by distribution in 96 well blood agar plates in 0 . 1 ml CM199 containing both antibiotics . Poisson analysis was used to determine the percentage probability of clonality , and was in each case >95% . Parasite DNA was extracted as follows: cell extracts were suspended in TELT ( 0 . 05 M Tris , 005 M EDTA , 0 . 25 M LiCl and Triton X-4 ) and supplemented with an equal Volume of Phenol Chloroform isoamyl alcohol ( 25∶24∶1 ) . Cell suspensions were centrifuged ( 15′ 16 , 000 g 4°C ) and the aqueous phase was transferred to a new tube . DNA was precipitated in 96% ethanol and washed with 70% ethanol . DNA was then eluted with 100 ul of nuclease free water . For hybrids genotyping , a set of marker genes ( Table S1 ) was amplified by polymerase chain reaction ( PCR ) using Applied Biosystems 2× geneAmpPCR mix , 25 pmol of each primer and 10 ng of the DNA . PCR products were cleaned with ExoSAP-IT kit ( USB ) , and sequences confirmed with forward and/or reverse reads by Rocky Mountain Laboratory Genomics Unit DNA Sequencing Center , Division of Intramural Research , Hamilton , Montana . The sequences were analyzed using “Lasergene” software . SNP-CAPS analysis was performed as previously described [9] and was based on SNPs identified in the sequence analysis . Glucose phosphate dehydrogenase ( G6PD-LmjF34 . 0080 ) was PCR-amplified and cleaned using gel and PCR clean up system ( Promega ) . The clean PCR product ( 500 ng ) was digested with 15 units of SACII ( Thermo scientific ) for 16 hours . Cleaved DNA was then loaded on ethidium bromide gel . Band intensity quantification was done using image J software ( http://rsbweb . nih . gov/ij , National Institutes of Health , Bethesda , MD , USA ) . DNA content was determined using flow cytometry as previously described [21] . Briefly , 2 . 5×106 log-phase promastigotes were permeabilized with 4% paraformaldehide solution for 1 minute . Cells were centrifuged , re-suspended in PBS and fixed with absolute methanol for 15 minutes on ice . Cells were washed and resuspended in PBS at room temperature for 10 minutes followed by RNAseA treatment ( 200 µg/ml RNase A ) and propidium iodide ( 20 µg/ml ) staining for 30 minutes at room temperature . Data were acquired on a FacsCANTO 2 flow cytometer ( BD Bioscience ) , counting at least 10 , 000 cells per sample , and analyzed using FlowJo 9 . 5 . 2 software . The full methods and analysis of the parental and hybrid L . major lines studied here will be presented elsewhere ( J . Shaik , N . Akopyants , D . Dobson , P . Lawyer , D . Elnaiem , D . Sacks , and S . Beverley , in preparation ) . Briefly , Leishmania nuclear DNA was isolated and subjected to deep-sequencing using Illumina Hi-seq 2000 machine , yielding approximately 60× coverage . Reads were aligned to the L . major reference genome ( REF ) using novoalign MPI version 2 . 07 . 07 ( Novocraft , 2012 ) ; “NovoAlign ( http://www . novocraft . com ) . ” ) and SNPs were identified using the Samtools pileup function using SOAP SNP consensus model [36] . Custom scripts were developed for subsequent analysis and display . SNPs were filtered to remove those falling within regions of low coverage ( read depth less than 10 ) and to remove SNPs heterozygous within either parent . For the remaining homozygous SNPs , an ‘allelic’ density at each homozygous SNP position was estimated , assuming an average ploidy of 2n across the L . major genome . While not strictly true for any given parent or hybrid line ( all of which vary from each other ) , the error introduced by this assumption is small . The results of these analyses are displayed as ‘bottlebrush’ plots ( Fig . 3 , Fig . S4 ) where at each SNP position the inferred allelic density is displayed on the Y-axis . To test if there is a change over time in the proportion of hybrids recovered within each cross , we used an exact two-sided Cochran-Armitage test for trend using the day as the score combining across experiments and assuming no experiment effects . To compare the number of promastigotes per gut between flies yielding hybrids versus those not yielding hybrids , we used a stratified ( by day ) Wilcoxon-Mann-Whitney test . Tests were performed using SAS Version 9 . 3 ( Cochran-Armitage ) or the coin package in Hothorn et al . [37] . URL http://www . jstatsoft . org/v28/i08/ . ] .
|
Leishmania are pathogenic protozoa characterized by substantial diversity in the sand fly species that can transmit them , in the mammalian species that can serve as their reservoir hosts , and in the disease forms and severity of the clinical outcomes they can produce in humans . The possibility that this diversity has arisen , at least in part , by a process involving genetic exchange was recently given experimental support by the recovery of hybrid parasites from sand flies co-infected with two strains of Leishmania major . Here , we demonstrate the sexual competency of L . major strains originating across the full geographic range of this parasite species , and in both natural and unnatural sand fly vectors . Our genotype analyses of a large number of hybrid clones confirmed that they inherited both parental alleles at the majority of chromosomal marker loci analyzed , consistent with a meiotic process , while kinetoplast DNA was inherited from only one parent . Surprisingly , a few nuclear loci were sometimes inherited from only one parent , suggesting loss of heterozygosity . The early timing of hybrid recovery suggests that nectomonad promastigotes are the most likely mating competent stage of the parasite . These studies provide the strongest evidence to date that sex is a component of the natural reproductive strategy of L . major .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"biology"
] |
2013
|
The Mating Competence of Geographically Diverse Leishmania major Strains in Their Natural and Unnatural Sand Fly Vectors
|
Stem cells reside in a particular microenvironment known as a niche . The interaction between extrinsic cues originating from the niche and intrinsic factors in stem cells determines their identity and activity . Maintenance of stem cell identity and stem cell self-renewal are known to be controlled by chromatin factors . Herein , we use the Drosophila adult testis which has two adult stem cell lineages , the germline stem cell ( GSC ) lineage and the cyst stem cell ( CySC ) lineage , to study how chromatin factors regulate stem cell differentiation . We find that the chromatin factor Enhancer of Polycomb [E ( Pc ) ] acts in the CySC lineage to negatively control transcription of genes associated with multiple signaling pathways , including JAK-STAT and EGF , to promote cellular differentiation in the CySC lineage . E ( Pc ) also has a non-cell-autonomous role in regulating GSC lineage differentiation . When E ( Pc ) is specifically inactivated in the CySC lineage , defects occur in both germ cell differentiation and maintenance of germline identity . Furthermore , compromising Tip60 histone acetyltransferase activity in the CySC lineage recapitulates loss-of-function phenotypes of E ( Pc ) , suggesting that Tip60 and E ( Pc ) act together , consistent with published biochemical data . In summary , our results demonstrate that E ( Pc ) plays a central role in coordinating differentiation between the two adult stem cell lineages in Drosophila testes .
In physiological condition , adult stem cells are surrounded by other cell types and extracellular matrix . Recent studies have demonstrated a dynamic crosstalk between stem cells and their surrounding microenvironment termed as the stem cell niche [1] . Signaling molecules emanating from this niche contribute to the balance between self-renewal and differentiation of adult stem cells , which is essential for the maintenance of tissue homeostasis and regeneration in response to injury . Therefore a fundamental question in stem cell biology is how extrinsic cues and intrinsic factors cooperate to determine and maintain stem cell identity and activity . Two adult stem cell lineages reside in the Drosophila adult testis: the germline stem cell ( GSC ) lineage and the cyst stem cell ( CySC ) lineage ( Fig 1A ) . Both GSCs and CySCs attach to a group of post-mitotic somatic cells called hub cells and serve as a niche for each other [2] . Both GSCs and CySCs undergo asymmetric cell divisions to produce one self-renewed stem cell and one differentiated daughter cell in each lineage [3 , 4] . The differentiated daughter cell in the GSC lineage is called a gonialblast ( GB ) , which subsequently undergoes a transit-amplifying stage with exactly four rounds of mitosis . After exiting the mitotic expansion , germ cells enter the meiotic stage with an elongated G2 phase as spermatocytes , in which a robust gene expression program is initiated to prepare them for meiotic divisions and spermatid differentiation . On the other hand , the differentiated daughter cell in the CySC lineage becomes a cyst cell , which never divides again . Two cyst cells encapsulate synchronously dividing and differentiating germ cells and form a distinct germ cell cyst . Ectopic niche formation may result in an expanded stem cell population and lead to tumor formation [5] . Conversely , dysfunction of stem cells from an impaired niche is associated with compromised injury recovery , degenerative disease and aging [6] . Studies using Drosophila gonads have improved our understanding of the regulatory mechanisms within the stem cell niche [2 , 7] . Drosophila testis has provided an excellent model system by which to study the crosstalk among different stem cell lineages . For example , it has been shown that the JAK-STAT and TGF-β signaling pathways are important for male GSC maintenance through interactions with CySCs [8 , 9 , 10 , 11] . The JAK-STAT signaling pathway ligand Unpaired ( Upd ) is secreted by hub cells to activate the downstream transcription factor Stat92E in both CySCs and GSCs for their maintenance [8 , 9 , 11 , 12 , 13 , 14 , 15] . In addition , the EGF signaling pathway has been shown to control cyst cells to encapsulate germ cells and allow for their proper differentiation [16 , 17 , 18 , 19 , 20] . A protease called Stet acts in germ cells to cleave the Spitz ( Spi ) ligand to stimulate EGF signaling in cyst cells [18] . Activation of EGF signaling ensures encapsulation of germ cells by the cyst cell and promotes germ cell differentiation [16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24] . Most studies on germline and soma communication have focused on signaling pathways , while most work on chromatin regulators mainly addressed their cell-autonomous functions . However , recent studies have demonstrated their cooperation ( reviewed by [25 , 26] ) . For example , JAK-STAT signaling in both GSCs and CySCs is positively regulated by the nucleosome remodeling factor ( NURF ) [27] . On the other hand , the Socs36E gene encodes an inhibitor of the JAK-STAT signaling pathway , which is critical for maintaining balance between GSCs and CySCs at the niche [12 , 15 , 23 , 24] . Our previous studies showed that an H3K27me3-specific histone demethylase , dUTX , acts upstream of and negatively regulates the JAK-STAT signaling pathway through maintaining active Socs36E transcription [28] . Moreover , genes of the EGF signaling pathway might be directly regulated by the H3K27me3 methyltransferase Enhancer of Zeste [E ( z ) ] in cyst cells to promote germ cell differentiation [29] . However , identification of more crosstalk between signaling pathways and chromatin factors in the CySC lineage has been hampered by the limited number of cyst cells for experimental methods , such as Chromatin immunoprecipitation ( ChIP ) and protein co-immunoprecipitation ( Co-IP ) . Thus , regulation of CySC differentiation and the coordination of CySC differentiation with neighboring germ cells remain to be fully addressed . The enhancer of Polycomb [E ( Pc ) ] gene is known as a putative Polycomb group ( PcG ) gene which is conserved from yeast to mammals , suggesting its crucial roles in regulating chromatin structure across species . The yeast homolog of E ( Pc ) was identified as a component of the NuA4 ( nucleosome acetyltransferase of H4 ) histone acetyltransferase ( HAT ) complex [30 , 31 , 32] , which has been shown to contribute to the hyperacetylation state of both H4 and H2A to stimulate transcription [33 , 34 , 35 , 36 , 37] . Abnormal activity of the human E ( Pc ) homolog called EPC1 has been shown to cause T-cell leukemia/lymphoma [38] . However , the molecular and cellular mechanisms of in vivo functions of E ( Pc ) have been elusive . Here we use the Drosophila adult testis as a model system to study functions of E ( Pc ) in endogenous adult stem cell lineages . We find that E ( Pc ) promotes cyst cell differentiation by downregulating CySC-enriched transcription factors Zfh-1 and Yan . Loss of such repression by E ( Pc ) in CySC lineage blocks both cyst and germ cell differentiation , which causes both CySC-like and early-stage germline tumors , including GSC-like tumor and spermatogonial tumor . In addition , when E ( Pc ) is specifically knocked down in CySC lineage , some germ cells ectopically turn on expression of cyst cell markers such as Zfh-1 and Yan . When ChIP followed by high-throughput sequencing ( ChIP-seq ) is performed specifically in cyst cells , many components of key signaling pathways are identified as direct targets of E ( Pc ) , suggesting its central role in coordinating the crosstalk between CySC and GSC lineages . Finally , inactivation of Tip60 HAT activity in the CySC lineage leads to defects resembling loss-of-function phenotypes of E ( Pc ) , suggesting that they act together in vivo . Thus , E ( Pc ) establishes proper chromatin state in cyst cells to provide an instructive microenvironment to guide germ cell differentiation and protect germline identity .
To understand the in vivo functions of E ( Pc ) in the Drosophila adult testis , we first characterized the E ( Pc ) expression pattern . Using a GFP-tagged genomic rescue transgene ( Materials and Methods ) , the nuclear E ( Pc ) gDNA-GFP signal was detected in both germ cells and cyst cells throughout the adult testis ( S1A–S1A” and S1B–S1B” Fig ) . The nuclear localization of E ( Pc ) is consistent with the prediction that E ( Pc ) is a chromatin regulator [39] . Because E ( Pc ) is required for early development and the null allele is lethal at embryonic or early larval stage [40] , we first studied the functions of E ( Pc ) in adult testes using the RNAi knockdown strategy [41] . When E ( Pc ) was knocked down in cyst cells using a cyst cell-specific Tj-Gal4 driver [42] paired with two independent RNAi lines [E ( Pc ) dsRNA or E ( Pc ) shRNA , when phenotypes from these two lines are indistinguishable we called them E ( Pc ) RNAi , see Materials and Methods] , the GFP signal representing the E ( Pc ) gDNA-GFP fusion protein level was greatly reduced in cyst cells compared to the neighboring germ cells ( S1C–S1C” Fig ) , suggesting efficient knockdown . In the CySC lineage , two transcription factors are known to express in a spatiotemporally specific manner . The first is zinc-finger homeodomain protein 1 ( Zfh-1 ) , a transcription repressor with multiple zinc finger domains and a homeodomain . It is highly expressed in CySCs and early cyst cells , and it is required for CySC maintenance [10] . The second is Eyes absent ( Eya ) , which is expressed in later stage cyst cells . It is required for cyst cell differentiation [43] . Immunostaining experiments showed very few cyst cells ( 6 . 9±2 . 5 ) with overlapping Zfh-1 and Eya signals in the control testis ( N = 44 ) ( Fig 1B–1B” , S2A Fig ) . On the other hand , the percentage of testes with cyst cells carrying both Zfh-1 and Eya immunostaining signals was significantly increased in both Tj>E ( Pc ) dsRNA ( N = 32 ) and Tj>E ( Pc ) shRNA ( N = 35 ) testes ( S2A Fig ) , most likely the result of overpopulation of CySC-like Zfh-1-expressing cells ( Fig 1C–1C” ) . It has been shown that Zfh-1 overexpression in CySC lineage leads to overpopulation of both CySCs and GSCs [10] . Based on microarray analysis [44] and RNA-seq data [45] , a transcription repressor , anterior open , often termed as Yan , is highly expressed in stem cell-enriched samples . Yan is an ETS domain-containing transcriptional repressor antagonizing the EGF signaling pathway [46] , and it inhibits cellular differentiation [47] . Immunostaining experiments showed enriched Yan protein in CySCs and possibly their immediate daughter cells in the control testes ( N = 22 ) ( Fig 1D ) . By way of contrast , the number of Yan-positive cells increased in 72% of Tj>E ( Pc ) dsRNA ( N = 18 ) and 74% of Tj>E ( Pc ) shRNA ( N = 35 ) testes , ( Fig 1E ) . Furthermore , immunostaining against the pan cyst cell marker Traffic jam ( Tj ) [48] and the later stage cyst cell marker Eya [43] both showed significantly increased Tj-positive and Eya-positive cells in Tj>E ( Pc ) shRNA testes ( N = 31 ) ( S2B–S2D Fig ) . These data suggest two major phenotypes in the CySC lineage upon knocking down E ( Pc ) : first , the normal spatiotemporally specific expression pattern of CySC-lineage markers was not preserved . Second , there were excess cyst cells including both CySC-like cells and later stage cyst cells . In addition to knock down E ( Pc ) in the entire CySC lineage , the hs-FLP; Actin-FRT-stop-FRT-Gal4 , UAS-GFP; UAS- E ( Pc ) shRNA fly strain ( Materials and Methods ) was used to induce E ( Pc ) knockdown in a subset of cells in CySC lineage . When GFP-positive cells ( arrows in Fig 1F–1F” and 1G–1G” ) were compared with neighboring GFP-negative wild-type cells ( arrowheads in Fig 1F–1F” and 1G–1G” ) in the same testis under the same experimental condition , ectopic expression of Zfh-1 ( arrow in Fig 1F” , N = 13 cells ) and Yan ( arrow in Fig 1G” , N = 11 cells ) was detected exclusively in GFP-positive cells , consistent with the entire CySC lineage knockdown phenotype . Because all E ( Pc ) knockdown experiments primarily used Tj-Gal4 driver , histone H3-GFP was used as a reporter in both Tj-Gal4>UAS-H3-GFP ( N = 45 ) and Tj-Gal4>UAS-H3-GFP , UAS-E ( Pc ) shRNA ( N = 44 ) testes ( S2E–S2F”’ Fig ) . Although GFP-positive cells increased in E ( Pc ) knockdown testes ( S2E and S2F Fig ) , consistent with the overall increase of Tj-positive cells ( S2D Fig ) , GFP signal was detected exclusively in the CySC lineage with no overlap with Vasa-positive germ cells ( S2E”’ and S2F”’ Fig ) , suggesting that the cell type specificity of the Tj-Gal4 driver is unaffected . The CySC lineage has been thought to play a supportive role for germ cell differentiation by enclosing germ cells and providing instructive signals for germline differentiation and survival [49 , 50 , 51] . We found that knockdown of E ( Pc ) in cyst cells using Tj-Gal4 led to excess early-stage germ cells in 43% of Tj>E ( Pc ) dsRNA testes ( N = 40 ) . Further reduction of E ( Pc ) levels , using a loss-of-function mutant E ( Pc ) 1 [52] as heterozygotes , significantly enhanced the excess early germ cell phenotype to 70% of Tj>E ( Pc ) dsRNA testes ( N = 20 ) . Early stage germ cells visualized by bright DAPI staining [10 , 17] were restricted to the apical tip region in the control testes ( Fig 2A and 2A” ) , but became expanded in the E ( Pc ) knockdown testes ( Fig 2B and 2B” ) . Another early-stage germ cell marker , Notch [16] , showed a confined immunostaining signal in the control testes ( Fig 2A’ and 2A” ) , but significantly increased signal in the E ( Pc ) knockdown testes ( Fig 2B’ and 2B” ) . The expansion of Notch-positive cells with DAPI bright nuclei is often associated with germline defects in the mitosis-to-meiosis transition , as shown previously [10 , 16 , 17] . We further analyzed the cellular properties of the excess germ cells in both Tj>E ( Pc ) dsRNA ( Fig 2 ) and Tj>E ( Pc ) shRNA ( S3 Fig ) testes . In 12 . 5% of Tj>E ( Pc ) dsRNA testes ( N = 40 , Fig 2C ) and 30% of Tj>E ( Pc ) shRNA testes ( N = 37 , S3A Fig ) , excess germ cells resembled GSC- or GB-tumor with round spectrosome structure intermingled with Zfh-1-positive CySC-like cells . In 30% of Tj>E ( Pc ) dsRNA testes ( N = 40 , white dotted outline in Fig 2D–2D” ) and 40% of Tj>E ( Pc ) shRNA testes ( N = 37 , S3B and S3B’ Fig ) , excess germ cells were more like spermatogonial tumors with more than 16 Vasa-positive cells within one cyst , as visualized by anti-Armadillo staining delineating the encapsulating cyst cells ( Fig 2D , S3B Fig ) . However , different from the continuous fusome structure in spermatogonial cysts in wild-type or control testes [53 , 54 , 55] , germ cells within one cyst showed both dotted spectrosome ( yellow arrowheads in Fig 2D , S3B and S3B’ Fig ) and branched fusome ( yellow arrows in Fig 2D , S3B and S3B’ Fig ) structures , suggesting that these cells were not undergoing cell cycle in synchrony . This asynchrony was further confirmed when EdU ( 5-ethynyl-2′-deoxyuridine ) incorporation assay was performed to label S-phase cells [56] . Only a subset of excess germ cells was labeled by EdU ( arrows in Fig 2D’ ) in 54% of single cysts ( N = 26 ) in E ( Pc ) knockdown testes . Bag-of-marbles ( Bam ) is an important differentiation factor detectable in 4- to 16-cell spermatogonia in wild-type [57 , 58] and control testes ( S4A and S4A’ Fig ) . In all testes with spermatogonial tumor ( 45% of Tj>E ( Pc ) dsRNA testes , N = 20; 57% of Tj>E ( Pc ) shRNA testes , N = 21 ) , Bam expression was detected in excess germ cells ( S4B–S4C’ Fig ) . It has been reported that in bam mutant testes , the transition from mitotic spermatogonia to meiotic spermatocyte is abolished , and the testes are enriched with synchronously dividing spermatogonial cells [57 , 58] . Here the presence of Bam ( S4B–S4C’ Fig ) and the absence of continuous fusome structure ( Fig 2D , S3B and S3B’ Fig ) suggest that these excess germ cells in E ( Pc ) somatic knockdown testes had different cellular properties compared to the bam mutant . Even though these excess cells were all positively stained with the germ cell marker Vasa in both Tj>E ( Pc ) dsRNA testes ( Fig 2C , 2E and 2E” ) and Tj>E ( Pc ) shRNA testes ( S3A , S3B’ , S3C and S3C” Fig ) , the early-stage cyst marker Zfh-1 was also detectable in these cells in 12 . 8% of Tj>E ( Pc ) dsRNA testes ( N = 86 , Fig 2E’–2E” ) and 10% of Tj>E ( Pc ) shRNA testes ( N = 79 , S3C’–S3C” Fig ) . To rule out the possibility that these Zfh-1 signals in Vasa-positive cells resulted from background staining of Zfh-1 antibody , Tj>H2Av-mRFP control testes ( H2Av-mRFP used as a marker ) and Tj>E ( Pc ) shRNA testes were co-immunostained and imaged using the same microscopic parameters . Vasa-positive and Zfh-1-positive cells were found in 11% of Tj>E ( Pc ) shRNA testes ( N = 45 ) , but not in any of the Tj-Gal4/H2Av-mRFP control testes ( N = 26 ) . Moreover , germ cells co-stained with Zfh-1 and Vasa were not found in other known germline tumors , such as GSC-like tumor in nos>upd testes [8 , 9] ( N = 48 , S5A–S5B” Fig ) and spermatogonial tumor in bam1/bam114 testes [45 , 57 , 58 , 59] ( N = 40 , S5C–S5D” Fig ) . In addition , cells co-stained with Vasa and Yan , another early-stage cyst cell marker , were also found in 13% of Tj>E ( Pc ) dsRNA testes ( N = 45 , Fig 2F–2F” ) and 12% of Tj>E ( Pc ) shRNA testes ( N = 43 , S3D–S3D” Fig ) . Our previous studies demonstrate that E ( z ) , a key PcG protein , is required in cyst cells to prevent germ cells from expressing Zfh-1 , suggesting a non-cell-autonomous role of E ( z ) in antagonizing somatic cell fate in the germline [29] . Interestingly , in the present study , compromising E ( Pc ) function showed phenotypes similar to those resulting from inactivation of E ( z ) in CySC lineage . Furthermore , when E ( z ) levels were reduced by either an E ( z ) 731 null allele [29 , 60] or a deficiency chromosome that uncovers the E ( z ) genomic region at the Tj>E ( Pc ) shRNA background , a more severe excess early-stage germ cell phenotype was observed ( S6A–S6D Fig ) . The similarity of loss-of-function phenotypes and the genetic interactions between E ( Pc ) and E ( z ) are consistent with the previous report that E ( Pc ) acts as an enhancer of PcG mutant [40] . Moreover , although the Tj-Gal4 driver knocks down E ( Pc ) in all somatic gonadal cells , including hub cells ( Fig 1A ) , knockdown of E ( Pc ) using a hub cell-specific upd-Gal4 driver [61] did not lead to any detectable cellular defect ( N = 28 , S7 Fig ) . Notably , these negative data could result from the strength of the upd-Gal4 driver or protein perdurance in the post-mitotic hub cells , which would reduce the efficiency of knockdown effect . Taken together , our data suggest that E ( Pc ) is required in the CySC lineage to promote germ cell differentiation and antagonize somatic identity in the germline . In addition to the knockdown strategy , we used the Mosaic Analysis with a Repressible Cell Marker , or MARCM , system [62] to generate E ( Pc ) mutant clones positively labeled by GFP . In control wild-type clones , Zfh-1 was undetectable in 74% ( yellow arrows in S8A–S8A”’ Fig ) and had diminished signal ( yellow arrows in S8B–S8B”’ Fig ) in 26% of Eya-positive cyst cells ( N = 38 ) . By contrast , Zfh-1 was detectable in all E ( Pc ) mutant cyst cells co-labeled with Eya ( N = 49 , yellow arrows in S8C–S8C”’ Fig ) , consistent with overlapping Zfh-1 and Eya expression in the E ( Pc ) knockdown cyst cells shown previously ( Fig 1C” and 1F” ) . Moreover , in 12 . 5% of testes ( N = 64 ) with E ( Pc ) mutant clones , extra DAPI bright cells were found to intermingle with Zfh-1-positive cells ( S8D , S8D’ , S8E and S8E’ Fig ) , resembling the excess early germ cell phenotype observed in E ( Pc ) knockdown testes ( Fig 2 , S3 Fig ) . The lower penetrance of the germ cell phenotype using the E ( Pc ) MARCM clone compared to E ( Pc ) knockdown in the entire CySC lineage likely results from the technical difficulty in ensuring that both cyst cells that encapsulate germ cells are E ( Pc ) mutants . Indeed , in 63% of testes ( N = 102 ) with E ( Pc ) mutant clones , GFP-negative wild-type cyst cells were detectable . In order to fully understand the molecular mechanisms underlying E ( Pc ) function in cyst cells responsible for promoting cellular differentiation , a chromatin immunoprecipitation followed by high-throughput sequencing ( ChIP-seq ) strategy was developed to profile the direct targets of E ( Pc ) specifically in the CySC lineage . In Tj>E ( Pc ) shRNA testes , a GFP-tagged E ( Pc ) cDNA transgene was expressed using the same Tj-Gal4 driver . Not only were the E ( Pc ) knockdown phenotypes ( Figs 1 and 2 , S3 Fig ) fully rescued in Tj>E ( Pc ) cDNA-GFP , E ( Pc ) shRNA testes ( N = 137 ) , but the E ( Pc ) cDNA-GFP fusion protein was also exclusively detected in the CySC lineage ( Fig 3A ) . Of note , even though the E ( Pc ) cDNA-GFP transgene signal was reduced in Tj>E ( Pc ) shRNA testes [S9 Fig , Tj>E ( Pc ) cDNA-GFP , E ( Pc ) shRNA ( N = 50 ) vs . Tj>E ( Pc ) cDNA-GFP ( N = 22 ) ] , suggesting knockdown effects , we reason that the residual E ( Pc ) cDNA-GFP is sufficient to rescue the E ( Pc ) knockdown phenotypes . Therefore , this genetic background provided a unique opportunity to immunoprecipitate E ( Pc ) -bound chromatin in the CySC lineage using a ChIP-grade GFP antibody [63] . We next analyzed our ChIP-seq data to identify the direct targets of E ( Pc ) in the CySC lineage . When all target genes were plotted over a -1-kb to +4-kb region with respect to the transcription start sites ( TSSs ) , enrichment of E ( Pc ) could be detected within a 600-bp region upstream of TSSs ( Fig 3B ) , which agrees with the prediction that E ( Pc ) is a chromatin factor regulating transcription of target genes . Using MACS2 with default setting and the P-value cutoff of 1e-5 , we identified 4 , 698 E ( Pc ) -bound genes in cyst cells from adult testes . Using the GO term enrichment test [64 , 65] to further analyze the direct target genes of E ( Pc ) , we found that signaling pathway components , genes responsible for DNA damage checkpoint , and genes encoding histone modifying enzymes represent the top three categories of E ( Pc ) target genes ( Fig 3C ) . In the signaling pathway category , genes associated with epidermal growth factor ( Egf ) , JAK-STAT , Wnt and Notch signaling pathways are all significantly enriched ( P<0 . 001 ) . We also performed RNA-seq to compare transcriptomes between Tj>E ( Pc ) shRNA testes and Tj-Gal4 control testes ( Fig 3D ) . We then interrogated the E ( Pc ) target genes retrieved from the ChIP-seq dataset with the RNA-seq dataset ( S1–S3 Tables ) . Most of the overlapping genes were upregulated in Tj>E ( Pc ) shRNA testes ( 1 , 507 genes ) compared with Tj-Gal4 testes ( Fig 3E ) , suggesting that the normal function of E ( Pc ) is to suppress transcription . Noticeably , ChIP-seq experiments were performed specifically in cyst cells . At the same time , however , it is extremely challenging from a technical point of view to isolate cyst cells to perform cell type-specific RNA-seq analysis as a result of the tight association between cyst cells and germ cells . Our RNA-seq experiments were therefore performed using the whole testes which reflected transcriptome changes in both germ cells and cyst cells . Notwithstanding , because we could not pinpoint the cyst cell-specific genes that are bound by E ( Pc ) and have transcriptional change upon knockdown of E ( Pc ) , we focused on a few known genes expressed in cyst cells for further analyses . The zfh-1 gene was among the 1 , 507 upregulated genes ( 1 . 64-fold upregulation , P< 0 . 01 , Fig 4A ) . This is consistent with the excess of Zfh-1-positive cells , as detected by immunostaining in Tj>E ( Pc ) RNAi testes ( Fig 1C and 1F” ) . Enrichment of E ( Pc ) was found at the endogenous zfh-1 gene locus ( Fig 4B ) , suggesting that E ( Pc ) directly binds to and downregulates zfh-1 expression in cyst cells . By way of contrast , no change in eya mRNA level was detected ( Fig 4A ) , in agreement with the immunostaining results showing no change of Eya protein level in Tj>E ( Pc ) RNAi testes ( Fig 1C’ and 1F’ ) . Congruent results showed that no E ( Pc ) enrichment was detected at the endogenous eya gene locus ( Fig 4C ) . It may be recalled that Yan , the other early-stage cyst cell marker , showed ectopic expression in Tj>E ( Pc ) RNAi testes ( Fig 1D and 1E ) . Here , E ( Pc ) binding at the endogenous yan locus ( Fig 4D ) did not pass the cutoff using a peak calling algorithm [66] . To further analyze the potentially weak binding of E ( Pc ) at the yan genomic locus ( Fig 4D ) , ChIPed DNA was analyzed using quantitative PCR ( qPCR ) with a series of primer sets ( Fig 4E ) spanning over a 1 . 8-kb genomic region around the TSS region of the yan gene . Compared to the more upstream and downstream sequences , enrichment of E ( Pc ) could be detected near TSS ( Fig 4E ) . Thus , it is possible that yan is a weaker E ( Pc ) target gene compared to zfh-1 . This speculation is supported by the slight increase of yan detected in Tj>E ( Pc ) shRNA testes ( 1 . 28-fold upregulation , P = 0 . 08 , Fig 4A ) . If E ( Pc ) acts as a transcriptional repressor to downregulate zfh-1 expression , we then reasoned that overexpression of E ( Pc ) could lead to decreased Zfh-1 levels . Because Zfh-1 is required for CySC self-renewal and GSC maintenance [11] , reduction of Zfh-1 might result in loss of both CySCs and GSCs . Indeed , when E ( Pc ) was overexpressed in the CySC lineage using Tj>E ( Pc ) cDNA , Zfh-1-positive CySCs and early cyst cells were significantly reduced ( S10A–S10C Fig ) . A reduced number of Zfh-1 cells may lead , in turn , to reduction of the cells in the CySC lineage , as shown by decreased Tj-positive cells ( S10E–S10G Fig ) . GSCs also showed a significant decrease ( S10A , S10B and S10D Fig ) , corroborating a previous study reporting that Zfh-1 regulates GSC self-renewal non-cell-autonomously [10] . Collectively , our results showed that E ( Pc ) is both necessary and sufficient to repress zfh-1 expression in the CySC lineage . Signaling pathway genes comprise the top ontological category of E ( Pc ) targets ( Fig 3C ) , suggesting their important roles in coordinating crosstalk between somatic and germline lineages . A previous RNAi screen using Drosophila S2R+ cells has identified E ( Pc ) as a positive regulator of the receptor tyrosine kinase and ERK signaling pathway [67] . Our data showed that yan is a potential target of E ( Pc ) ( Fig 4D and 4E ) , which is normally repressed by E ( Pc ) ( Fig 1E and 1G” ) . Because Yan functions as an antagonist of EGF signaling [46 , 47] , E ( Pc ) could be a positive regulator of EGF signaling in the CySC lineage . We next studied the potential synergistic activities between E ( Pc ) and the EGF signaling pathway . The EGF signaling pathway has previously been shown to control the encapsulation of germ cells by cyst cells and then regulate their proper differentiation [16 , 17 , 18 , 19 , 20 , 22 , 68 , 69] . Consistent with their synergistic activities , knockdown of E ( Pc ) in cyst cells resulted in phenotypes resembling those caused by loss-of-function of EGF signaling pathway components . For example , when EGF signaling is compromised , it has been reported that germ cells have differentiation defects [16 , 17 , 18 , 23 , 24] and divide asynchronously [22] , similar to those germline phenotypes in the Tj>E ( Pc ) RNAi testes ( Fig 2D–2D” , S3B and S3B’ Fig ) . In addition , using the Vein-LacZ reporter as a readout of EGF signaling activity [16 , 24 , 70 , 71 , 72] , expression of this reporter was absent in early-stage Zfh-1-positive cyst cells ( yellow arrowhead in Fig 5A–5A” ) , but expression was robust in differentiated cyst cells ( yellow arrows in Fig 5A–5A” ) , suggesting increased EGF signaling activity during normal cyst cell differentiation , as reported previously [22] . However , in Tj>E ( Pc ) shRNA testes , Vein-LacZ expression was almost undetectable in later stage cyst cells ( yellow arrows in Fig 5B–5B” ) , suggesting compromised EGF signaling activity by E ( Pc ) knockdown . Quantification of the intensity of Vein-LacZ signal in later stage cyst cells ( yellow arrows in Fig 5A–5A” , 5B-B” ) showed significant difference between Tj-Gal4 and Tj>E ( Pc ) shRNA testes ( Fig 5C , Materials and Methods ) . Moreover , consistent with the synergistic activities between E ( Pc ) and the EGF signaling , halving the level of EGFR using a Egfrf2 null allele as heterozygotes enhanced the germline phenotype in Tj>E ( Pc ) dsRNA testes ( Fig 5D ) . On the other hand , a constitutively active form of Yan ( YanCA ) , when expressed in cyst cells using the Tj-Gal4 driver , resulted in phenotypes similar to those observed in Tj>E ( Pc ) RNAi testes ( Figs 1C , 2C and 2D , S3A , S3B and S3B’ Fig ) . First , Zfh-1-positive cells were overpopulated ( Fig 5E’ ) in all Tj>YanCA testes ( N = 21 ) . Second , excess germ cells were detected as GSC- or GB-like tumors ( yellow outline , Fig 5E–5E” ) in all Tj>YanCA testes ( N = 21 ) and spermatogonial tumors ( white outline , Fig 5E–5E” ) in 90% of Tj>YanCA testes ( N = 21 ) . Third , a null allele yanIP [73] acted as a strong suppressor of the germline differentiation defects in Tj>E ( Pc ) shRNA testes ( Fig 5F ) , supporting the hypothesis that part of the E ( Pc ) knockdown phenotype results from upregulated expression of Yan . Activated EGF signaling has been shown to induce the entry of phosphorylated active MAP kinase ( dpERK ) to the nucleus in order to regulate the transcription of target genes [16 , 68 , 74] . Therefore , to further understand how E ( Pc ) regulates EGF signaling , we characterized the expression level and localization of dpERK in E ( Pc ) knockdown cyst cells . We induced E ( Pc ) knockdown and GFP-positive cells using the strategy discussed above ( Fig 1F–1G” , Materials and Methods ) . As a control , GFP-positive wild-type cells were also induced using the same method . In neither case was the level , or subcellular localization , of dpERK distinguishable between GFP+ and GFP- cyst cells ( Fig 5G and 5G’ and Fig 5H and 5H’ ) , suggesting that E ( Pc ) may act in parallel with , or downstream of , dpERK to regulate the chromatin state of target genes responsive to the EGF signaling . The JAK-STAT signaling pathway has been shown to play prominent roles in regulating self-renewal of both CySCs and GSCs [8 , 9 , 10 , 11 , 12 , 15 , 23 , 24 , 75 , 76 , 77 , 78 , 79] . The Upd ligand is secreted locally from the hub cells and acts through the Domeless receptor to activate the Janus kinase Hopscotch and phosphorylate the STAT92E transcription factor , which is subsequently translocated to the nucleus to activate target gene transcription [80 , 81] . Our ChIP-seq data identified significant enrichment of E ( Pc ) at the genomic loci of multiple JAK-STAT pathway genes , including domeless , hopscotch and stat92E ( Fig 6A ) , suggesting that E ( Pc ) might directly regulate the activity of the JAK-STAT signaling pathway . We then used a 2X STAT-GFP reporter [28 , 82 , 83 , 84] , having two copies of the STAT92E DNA binding sites from a known JAK-STAT target Socs36E upstream of the GFP sequences , as a readout of JAK-STAT signaling activity in cyst cells . In the control testes ( N = 27 ) , the GFP signal was only detectable in CySCs localized to a diameter of one cell away from the hub region ( arrowheads in Fig 6B–6B” ) , but not in differentiated cyst cells labeled with Eya ( arrows in Fig 6B–6B” ) . By contrast , in 77% of Tj>E ( Pc ) shRNA testes ( N = 52 ) , Eya-positive cells showed a robust GFP signal ( arrows in Fig 6C–6C” ) , indicating active JAK-STAT signaling in these later stage cyst cells . Ectopic JAK-STAT activity in the cyst cells with compromised E ( Pc ) supports the hypothesis that E ( Pc ) represses JAK-STAT signaling . As shown previously , zfh-1 , another JAK-STAT signaling target gene [10] , also showed ectopic expression in later stage cyst cells in Tj>E ( Pc ) RNAi testes ( Fig 1C and 1F” ) . In summary , these data support the idea that E ( Pc ) directly represses JAK-STAT signaling activity in the CySC lineage . Drosophila E ( Pc ) was identified as a component of the Tip60 HAT complex in S2 cells [85] . Biochemical experiments demonstrate that Tip60 acetylates H4 and H2A and that such activity is conserved from yeast [31] to human [86] . In order to examine how Tip60 and E ( Pc ) cooperate in the CySC lineage , we first examined loss-of-function phenotypes of Tip60 using two strategies: a Tip60 RNAi transgene [87] ( Tj>Tip60 RNAi ) and a Tip60 dominant negative form ( Tj>Tip60E431Q ) [88] , both driven by the same Tj-Gal4 as that used in E ( Pc ) knockdown experiments ( Figs 1 and 2 ) . We found that both strategies led to defects similar to the phenotypes characterized in Tj>E ( Pc ) RNAi testes . First , excess Zfh-1-expressing cells could be detected in 74% of Tj>Tip60 RNAi ( N = 70 , Fig 7A’ ) and 77% of Tj>Tip60E431Q ( N = 52 , Fig 7B’ ) testes , leading to the co-expression of Zfh-1 and Eya in the cyst cells of both Tj>Tip60 RNAi ( yellow arrows , Fig 7A’–7A”’ ) and Tj>Tip60E431Q ( yellow arrows , Fig 7B’–7B”‘ ) testes . Second , in 46% of Tj>Tip60 RNAi ( N = 70 , Fig 7A ) and 50% of Tj>Tip60E431Q ( N = 52 , Fig 7B ) testes , expansion of germ cells with DAPI bright nuclei was detected . Further characterization of the excess germ cells showed early-stage germline tumor ( Fig 7C ) in 8% and spermatogonial tumor ( white dotted outline , Fig 7D ) in 28% of Tj>Tip60 RNAi testes ( N = 60 ) , respectively . Similar early-stage germline tumor ( Fig 7E ) and spermatogonial tumor ( white dotted outline , Fig 7F ) were also found in 8% and 47% of Tj>Tip60E431Q testes ( N = 38 ) , respectively . Third , cells with both germline marker Vasa and early cyst cell marker Zfh-1 could be detected in 18% of Tj>Tip60 RNAi ( N = 60 , Fig 7G–7G” ) and 19% of Tj>Tip60E431Q ( N = 21 , Fig 7I and 7I’ ) testes , respectively . Cells co-expressing Vasa and Yan , another early-stage cyst cell marker , were also observed in 10% of Tj>Tip60 RNAi ( N = 60 , Fig 7H–7H” ) and 29% of Tj>Tip60E431Q ( N = 38 , Fig 7J–7J” ) testes , respectively . Because the mutation in the Tip60E431Q transgene abolishes the HAT activity of Tip60 [89 , 90] , similar phenotypes between Tj>Tip60 RNAi and Tj>Tip60E431Q testes demonstrate that the function of Tip60 in the CySC lineage relies on its HAT enzymatic activity . In summary , both cyst cell and germline defects in either Tj>Tip60 RNAi or Tj>Tip60E431Q testes were similar to those found in Tj>E ( Pc ) RNAi testes ( Fig 1B–1E and Fig 2 and S3 Fig ) , suggesting that E ( Pc ) and Tip60 act together to regulate cyst cell differentiation cell-autonomously , as well as coordinate germ cell differentiation and maintain germline fate non-cell-autonomously . To further explore the potential synergistic activities between E ( Pc ) and Tip60 , we tested their genetic interactions . Because knockdown efficiency using the Gal4: UAS system depends on temperature [91 , 92 , 93] , flies raised at 25°C instead of 29°C showed less severe phenotypes and with lower penetrance . For example , germline tumor was detected in 0% of Tj>Tip60 RNAi ( N = 36 ) and 13% of Tj>E ( Pc ) shRNA ( N = 30 ) testes , respectively ( S11A Fig ) . By contrast , under the same condition , 31% of Tj>Tip60 RNAi , E ( Pc ) shRNA testes ( N = 39 ) showed germline tumor phenotype ( S11A Fig ) . In the same way , an E ( Pc ) w3 mutant [94 , 95] , used as heterozygotes , enhanced the germline phenotype in Tj>Tip60E431Q testes ( S11B Fig ) . These data suggest that E ( Pc ) and Tip60 act together in the CySC lineage to regulate germ cell differentiation . It was also notable that overexpression of Tip60 led to significant reduction of Zfh-1-positive cells ( S11C Fig ) and GSCs ( S11D Fig ) , similar to the effects caused by overexpression of E ( Pc ) ( S10C and S10D Fig ) . In summary , Tip60 resembles E ( Pc ) in its necessary and sufficient roles in repressing zfh-1 expression in the CySC lineage . Furthermore , if E ( Pc ) acts with Tip60 whose functions depend on its HAT activity , it is possible that E ( Pc ) regulates the histone H4 acetylation ( H4 ace ) state of its target genes . To examine this possibility , anti-H4 ace [96 , 97] ChIP-ed DNA from both Tj-Gal4 and Tj >E ( Pc ) RNAi testes were analyzed using qPCR with two series of primers spanning over the genomic loci of zfh-1 and yan , respectively . We found decreased H4 ace at both zfh-1 ( S12A Fig ) and yan ( S12B Fig ) genomic regions in E ( Pc ) knockdown testes compared with the control testes , suggesting that the activity of E ( Pc ) is also required for histone acetylation state at target genes .
Drosophila E ( Pc ) has been shown to be a component of the Tip60 HAT complex [85] . However , the functional relationship between E ( Pc ) and Tip60 in vivo has not been elucidated . We found that either knockdown of Tip60 or compromise of HAT activity of Tip60 resulted in phenotypes similar to those caused by E ( Pc ) loss-of-function ( Fig 7 ) . Moreover , enhancement of the E ( Pc ) phenotypes by Tip60 mutations ( S11A and S11B Fig ) suggests that E ( Pc ) acts with Tip60 . Finally , the levels of H4 acetylation at zfh-1 and yan genomic regions decrease upon E ( Pc ) knockdown ( S12 Fig ) , suggesting that E ( Pc ) is required for the HAT activity of Tip60 . Generally , histone acetylation has been linked to gene activation . However , both zfh-1 and yan are upregulated ( Fig 4A; Fig 1C” and 1F”; Fig 1E and 1G” ) with decreased H4 acetylation levels ( S12 Fig ) upon inactivation of E ( Pc ) . These data suggest that histone acetylation may repress gene expression . In this regard , it is noteworthy that the antibody against H4 acetylation used for ChIP-qPCR assay is not specific for a particular Lys residue . We speculate that the acetylation of the Lys12 of H4 ( H4K12ac ) might contribute to this phenomenon . It has been reported that H4K12ac is not associated with the active transcription region in early spermatocytes [99] and is enriched at the chromocentric heterochromatin region at polytene in salivary gland cells [100] in Drosophila . Moreover , Tip60 was reported to repress expression of differentiation genes to maintain pluripotency of mouse embryonic stem cells [101] , indicating histone acetylation as a contributor to gene silencing . In fact , histone acetylation at other Lys residues was also reported to have repressive roles of gene expression . For example , H3K56ac was reported to repress transcription of newly replicated DNA in budding yeast [102] . Another example is H4K20ac , which was found to be enriched with transcriptional repressors at silenced genes in human cells [103] . Taken together , these data indicate that histone acetylation is not always associated with gene activation , but that it could contribute to gene silencing . Even though knockdown of E ( Pc ) in CySC lineage leads to CySC differentiation defects , the most prominent phenotypes were detected in the germline . Germ cells in Tj>E ( Pc ) RNAi testes have interesting new phenotypes . First , excess germ cells divide asynchronously , a phenotype different from previously identified spermatogonial tumors in bam and benign gonial cell neoplasm ( bgcn ) mutant testes , in which the transition from spermatogonia to spermatocyte is abolished , and the testes are enriched with synchronously dividing spermatogonia [57 , 58] . However , the presence of Bam ( S4B–S4C’ Fig ) and the absence of continuous fusome structure ( Fig 2D , S3B and S3B’ Fig ) both suggest that these excess germ cells in Tj>E ( Pc ) RNAi testes have different cellular properties . In addition , the excess germ cells do not resemble expanded early-stage germ cells upon hyperactivation of the JAK-STAT signaling pathway in Drosophila testis [8 , 9 , 10] . In fact , the excess germ cells in Tj>E ( Pc ) RNAi testes have features resembling both spermatogonial tumors and GSC- or GB-like tumors ( Fig 2 and S3 and S4 Figs ) . Second , germ cells in Tj>E ( Pc ) RNAi testes ectopically turn on early-stage cyst cell markers , such as Zfh-1 and Yan , indicating that E ( Pc ) acts in cyst cells to prevent germ cells from taking somatic cell fate . The dichotomy between germline and soma represents the earliest lineage specification among many metazoan organisms . In multiple model organisms , including C . elegans and Drosophila , germ cell identity is determined by maternally loaded germ granules [104 , 105 , 106 , 107 , 108 , 109] . After specification , protection of germline fate requires both proper chromatin state and specific cytoplasmic factors in the germ cells [110 , 111 , 112 , 113 , 114 , 115 , 116 , 117] . Previous study identified PcG component E ( z ) as a non-cell autonomous factor in repressing the somatic fate of germ cells in adult Drosophila testis [29] . Here we identified that both E ( Pc ) and Tip60 play similar roles in cyst cells to maintain the germline identity in a non-cell autonomous manner , indicating that they might act with E ( z ) in regulating a critical signaling pathway ( or pathways ) to keep germline identity throughout adulthood . These results also emphasize the important roles of the somatic gonadal cells in protecting germline from taking somatic cell fate to ensure proper differentiation into functional gametes . Even though previous studies suggest that somatic gonadal cells control germ cell differentiation and maintain germline identity through multiple signaling pathways [8 , 9 , 11 , 16 , 17 , 29 , 49 , 76 , 98 , 118 , 119 , 120] , it is unclear how these signaling pathways themselves are regulated . Here our ChIP-seq results reveal that E ( Pc ) is enriched at key components of multiple signaling pathways known to be important in regulating germ cell function . For example , the EGF signaling pathway has been shown to regulate cyst cells in encapsulating germ cells and promoting their proper differentiation in Drosophila testis [16 , 17 , 18 , 19 , 20 , 22 , 68 , 69] . We found that the EGF antagonist Yan is highly enriched in CySCs , but decreased dramatically in later stage cyst cells repressed by E ( Pc ) during CySC differentiation . In line with this , the expression of Vein , which is downstream of the EGF signaling , is compromised in Tj>E ( Pc ) RNAi testes ( Fig 5B–5B” and 5C ) . Recently , decreased EGF signaling has been shown to induce extra germ cell division out of synchrony [22] . Similar asynchronous division of germ cells upon compromising either E ( Pc ) function or EGF signaling suggests that E ( Pc ) acts in synergy with the EGF signaling pathway , probably through regulation of the chromatin state at the endogenous yan locus . Similarly , E ( Pc ) was found to be enriched at multiple JAK-STAT pathway components . However , different from increased EGF signaling activity during CySC differentiation , high JAK-STAT signaling activity is only detectable in early-stage cells of both CySC and GSC lineages [12 , 14 , 77 , 84 , 121 , 122] . Hyperactivation of JAK-STAT signaling in either CySC or GSC lineage is sufficient to block cellular differentiation and results in tumors with CySC- and GSC-like features [8 , 9 , 10] . In this scenario , E ( Pc ) might downregulate JAK-STAT signaling to promote CySC differentiation by directly repressing the expression of key JAK-STAT components . Consistent with this finding , we observed that the 2X STAT-GFP reporter showed prolonged expression in later stage cyst cells when E ( Pc ) was inactivated ( Fig 6C–6C” ) . However , this reporter uses the upstream regulatory sequences from the Socs36E gene [82] , which itself acts as a repressor of JAK-STAT signaling [123 , 124] . This negative feedback regulation of JAK-STAT signaling might explain why removal of one copy of stat gene , using null allele stat06346 , or zfh-1 gene , using either a mutant allele zfh-175 . 26 or a deficiency chromosome that uncovers the zfh-1 gene region [10 , 125] ) , did not efficiently suppress the Tj>E ( Pc ) RNAi phenotype: the percentages of testes with medium and severe excess germ cells were 35% and 37% for Tj>E ( Pc ) shRNA testes ( N = 156 ) , 47% and 53% for Tj>E ( Pc ) shRNA , Stat92E06346/+ testes ( N = 75 ) , 34% and 43% for Tj>E ( Pc ) shRNA , Zfh-175 . 26/+ testes ( N = 155 ) , and 47% and 29% for Tj>E ( Pc ) shRNA , Df[Zfh-1]/+ testes ( N = 150 ) , respectively . Furthermore , studies in the Drosophila optic lobe identified E ( Pc ) as one JAK-STAT target positively regulated by JAK-STAT signaling [126] , suggesting mutual regulation between E ( Pc ) and the JAK-STAT signaling . Collectively , then , we found that E ( Pc ) regulates multiple signaling pathways and may act as a master regulator for the communications between the somatic and germline lineages in the Drosophila adult testis . The ultimate readout in E ( Pc ) mutants depends on the particular E ( Pc ) targets in the signaling pathway ( s ) and is complicated by the dual roles of E ( Pc ) in either activating or repressing gene expression . However , this complicated feature of E ( Pc ) regulation might be necessary to fine tune activities of different signaling pathways . In summary , we demonstrate that a chromatin factor E ( Pc ) acts in cyst cells and is responsible for germline differentiation and germ cell fate maintenance . These results emphasize the importance of the microenvironment where germ cells reside in antagonizing somatic identity and promoting germ cell differentiation . Similar to Drosophila testis , many mammalian stem cell niches support multiple stem cells . For example , both hair follicle stem cells and melanocyte stem cells co-occupy the hair follicle bulge [127 , 128] . The hair follicle stem cells have been shown to function as a niche for melanocyte stem cells through the TGF-β signaling [129] . Similarly , mesenchymal stem cells and hematopoietic stem cells co-exist in the bone marrow , and mesenchymal stem cells constitute the hematopoietic stem cell niche [130] . Understanding the coordination between two stem cell lineages during differentiation may shed light on other complex niches that support multiple stem cell populations .
Flies were raised under standard yeast/molasses medium at 25°C unless stated otherwise . The following flies were used: E ( Pc ) 1 ( Bloomington Drosophila Stock Center , BL3056 ) , E ( Pc ) w3 ( BL9396 ) , UAS-E ( Pc ) dsRNA ( BL28686 ) , UAS-E ( Pc ) shRNA ( BL35271 ) , upd-Gal4 ( from D . Harrison , University of Kentucky , Lexington , KY , USA ) , Tj-Gal4 ( Kyoto stock center , DGRC#104055 ) , Egfrf2 ( BL2768 ) , yanIP ( BL3101 ) , E ( z ) 731 ( BL24470 ) , Df [E ( z] ( BL29023 ) , UAS-yan . ACT ( BL5789 ) , Vein-lacZ ( BL11749 ) , 2X STAT-GFP [82] , Stat92E06346 ( from N . Perrimon , Harvard Medical School , Boston , MA , USA ) , UAS-Tip60 dsRNA ( BL28563 ) , UAS-dTip60E431Q , UAS-Tip60 ( from Felice Elefant , Drexel University , Philadelphia , Pennsylvania , USA ) , Bam-HA [131] , hs-FLP; Act5c . FRT-CD2-FRT . Gal4; UAS-GFP flies ( from Allan Spradling , Carnegie Institution for Science , Department of Embryology , Baltimore , Maryland , USA ) [132] , UAS-GFP hs-FLP; FRT42D , Tub-Gal80; Tub-Gal4 ( from Duojia Pan , Johns Hopkins Medical Institution , Baltimore , Maryland , USA ) , P{neoFRT}42D ( BL1802 ) , zfh-175 . 26 ( from Ruth Lehmann , NYU school of medicine , New York , USA ) , Df ( zfh-1 ) ( BL7917 ) , UAS-H3GFP , UAS-Upd ( from Stephen DiNardo , Perelman School of Medicine at the University of Pennsylvania , Department of Cell and Developmental Biology , Philadelphia , PA , USA ) , bam114/TM6B ( from Margaret T . Fuller , Developmental Biology and Genetics , Stanford University School of Medicine , Stanford , CA ) , bam1/TM3 ( from Allan Spradling , Carnegie Institution for Science , Department of Embryology , Baltimore , Maryland , USA ) . To study function of E ( Pc ) in cyst cells , two independent RNAi lines UAS-E ( Pc ) dsRNA and UAS-E ( Pc ) shRNA were crossed with different drivers upd-Gal4 and Tj-Gal4 at 25°C , respectively . Newly enclosed progenies were shifted to 29°C and maintained for 8–10 days ( D ) before dissection . For Tip60 function study , RNAi line UAS-Tip60 dsRNA and dominant negative HAT deficient line UAS-Tip60E431Q were crossed with Tj-Gal4 at 25°C and then adult progenies were shifted to 29°C and maintained for 8–10 D before dissection . To identify if E ( Pc ) genetic interacts with Egfr , yan , E ( z ) , Stat92E , alleles Egfrf2 , yanIP , E ( z ) 731 , Stat92E06346 , zfh-175 . 26 and deficiency lines Df [E ( z ) ] , Df ( zfh-1 ) were used . Flies with the following genotypes: Tj-Gal4/ Egfrf2; UAS-E ( Pc ) dsRNA/+ , Tj-Gal4/+; UAS-E ( Pc ) dsRNA/+ were shifted to 29°C for 3D before analysis . Flies with the following genotypes: Tj-Gal4/yanIP; UAS-E ( Pc ) shRNA/+ , Tj-Gal4/+; E ( z ) 731 / UAS-E ( Pc ) shRNA , Tj-Gal4/+; Df [E ( z ) ]/UAS-E ( Pc ) shRNA , Tj-Gal4/+; UAS-E ( Pc ) shRNA/ Stat92E06346 , Tj-Gal4/+; UAS-E ( Pc ) shRNA/ zfh-175 . 26 , Tj-Gal4/+; UAS-E ( Pc ) shRNA/ Df ( zfh-1 ) were shifted to 29°C for 5D before dissection . To study if expression of E ( Pc ) cDNA-GFP in cyst cells is sufficient to rescue Tj>E ( Pc ) RNAi phenotype , flies with the following genotype: Tj-Gal4/ UAS-E ( Pc ) cDNA-GFP; UAS-E ( Pc ) shRNA/+ , Tj-Gal4/ UAS-E ( Pc ) cDNA-GFP; UAS-E ( Pc ) dsRNA/+ were dissected at 5D after shifting from 25°C to 29°C . To test potential defects by overexpression of E ( Pc ) or Tip60 , testes from Tj-Gal4/ UAS-E ( Pc ) cDNA , Tj-Gal4/ +; UAS-Tip60 cDNA/+ males 10-11D after shifting from 25°C to 29°C were analyzed . To analyze function of E ( Pc ) in individual cyst cells , flies with the following genotype: hs-FLP; Act5c . FRT-CD2-FRT . Gal4/+; UAS-GFP/UAS-E ( Pc ) shRNA and hs-FLP; Act5c . FRT-CD2-FRT . Gal4/+;UAS-GFP/+ ( hs: heatshock promoter , Actin: actin promoter ) were heat shocked at pupal stages for two days with two hours on each day . Enclosed flies were aged for 5-6D after heat shock and used for dissection and immunostaining . To generate MARCM clones , E ( Pc ) 1 null allele was recombined with FRT42D to generate FRT42D , E ( Pc ) 1/Cyo flies . Adult flies with following genotype: UAS-GFP hs-FLP; FRT42D , Tub-Gal80/ FRT42D , E ( Pc ) 1; Tub-Gal4 and control flies UAS-GFP hs-FLP; FRT42D , Tub-Gal80/ FRT42D; Tub-Gal4 were aged for one day , then heat shocked for 2 hours and aged until dissection . For transgenic fly UASp-E ( Pc ) cDNA and UASp-E ( Pc ) cDNA-GFP , E ( Pc ) cDNA was amplified using cDNA prepared from wild-type testis as the template . The 5’ half of E ( Pc ) cDNA was amplified as a KpnI and NotI flanked fragment with E ( Pc ) F1 and R1 primers . The 3’ half of E ( Pc ) cDNA was amplified as an NotI and XbaI flanked fragment using E ( Pc ) F2 and R2 primers . These two fragments were then ligated into pGEM-T-easy vector ( Promega ) followed by sequencing . To insert the GFP sequences at the 3’-end of E ( Pc ) cDNA , a Pml I site was generated right upstream of the stop codon of E ( Pc ) within R2 primer . GFP fragment was amplified as a Pml I and Xba I flanked fragment with Primer 5’ GFP and 3’ GFP , followed by ligation into pGEM-T-E ( Pc ) 3’ half cDNA opened with Pml I and XbaI restriction enzyme digestion . Finally , the 5’ half E ( Pc ) cDNA in a KpnI to NotI fragment and the 3’ half with and without GFP in a Not I to Xba I fragment were ligated into pBlueScript vector ( Agilent Technologies ) cut with Kpn I and Xba I in a 3-way ligation to generate a KpnI and XbaI flanked full-length E ( Pc ) cDNA tagged with GFP . Then the E ( Pc ) full cDNA tagged with GFP was cut with Kpn I and Xba I and ligated into UASp vector cut using same two enzymes . To generate E ( Pc ) genomic plasmid tagged with GFP , a 21 kb P[acman] BAC clone ( CH322-140G22 ) covering the entire E ( Pc ) genomic region was ordered from BACPAC Resources Center ( BPRC ) . Zra I is one unique enzyme site close to the stop codon of E ( Pc ) genomic region . Pac I is another unique enzyme site within the 3’UTR region of E ( Pc ) . Using primers 3’UTR F and 3’UTR R ended with Asc I and Pac I , an approximate 3 Kb fragment was amplified using the BAC clone as template and ligated into pGEM-T easy vector . Using primers GFP F and GFP R ended with Zra I and Asc I GFP sequence was amplified . Then , the GFP in a Zra I to Asc I fragment was ligated into pGEM-T 3’UTR vector , cut with AscI and PacI to generate a GFP-3’ UTR fragment flanked by ZraI and PacI . Then GFP-3’ UTR cut with ZraI and PacI was ligated into P[acman] , opened with ZraI and PacI to generate E ( Pc ) genomic plasmid tagged with GFP . Transgenic fly lines were generated by Bestgene Inc ( Chino Hills , CA ) . More than three independent transgenic lines were generated for each transgene . Primers: Testes were dissected in 1X PBS and then fixed in 4% formaldehyde in 1X PBS for 30 min at room temperature ( RT ) . Then testes were washed twice with 20min each time using 1X PBST ( 0 . 1% triton ) at RT . Testes were incubated with Primary antibodies on nutator at 4°C overnight . After twice wash with 1X PBST , testes were incubated with secondary antibodies in darkness at RT for 2 hours . After twice wash with 1X PBST , testes were mounted using Vectashield ( Vector H-1200 ) . Primary antibodies used are: Vasa ( Rabbit , Santa Cruz , sc-30210 ) , Vasa ( Rat , 1:100 , developed by Spradling , A . C . / Williams , D . obtained from DSHB ) , Zfh-1 ( Rabbit , 1:5000 , from R . Lehmann ) , Fas III ( Mouse , 1:100 , DSHB , 7G10 ) , Armadillo ( Mouse , 1:200 , DSHB , N2 7A1 ) , Eya ( Mouse , 1:25 , DSHB , 10H6 ) , Tj ( Guinea pig , 1:1000 , from M . Van Doren ) , Yan ( Mouse , 1:25 after pre-absorption against Drosophila embryos , DSHB , 8B12H9 ) , GFP ( Chicken , 1:1000 , Abcam , ab13970 ) , dpERK ( Rabbit , 1:100 , Cell signaling , #4370 ) , HA ( Rat , 1:50 , Roche , 3F10 ) , β-Galactosidase ( Mouse , 1:200 , Sigma , G4644 ) . For dpERK staining , testes were dissected in 10 mM Tris-Cl pH 6 . 8 , 180 mM KCl , 50 mM NaF , 1 mM Na3VO4 , 10 mM b-glycerophosphate as described before [16] . Secondary antibodies were all Alexa Fluor series ( 1:200 , Molecular Probes ) . Images were taken with Zeiss LSM 510 META or LSM 700 . Images were processed using Adobe Photoshop . EdU incorporation was performed with Click-iT EdU Alexa Fluor 488 imaging kit ( Invitrogen C10083 ) . Dissected testes were incubated with EdU solution for 30min , followed by fixation and immunostaining as described . To compare Vein intensity between the Tj-Gal4 control and Tj>E ( Pc ) shRNA testes , H2Av-mRFP ( BL34498 ) transgene was used as a marker to distinguish the two genotyped fly testes . Testes dissected from Tj-Gal4/H2Av-mRFP; Vein-LacZ/+ males were compared with Tj-Gal4/+; Vein-LacZ/UAS-E ( Pc ) shRNA testes , which were immunostained together and imaged using the same parameters . Control testes could be identified based on the H2Av-mRFP marker . Vein-LacZ fluorescence intensity was measured for each Z stack across the entire nucleus using Image J software and summed up . Data were analyzed and presented using GraphPad Prism software . Flies with following genotype: Tj-Gal4/ UAS-E ( Pc ) cDNA-GFP; UAS-E ( Pc ) shRNA/+ were collected as newly eclosed males and aged for 5D at 29°C after shift from 25°C . Approximately 2 , 000 pairs of testis were dissected and grouped into two batches which were used as two replicates for ChIP experiments , which were performed using ChIP-IT high sensitivity kit ( #53040 , Active motif ) following the manufacturer’s instruction . A ChIP-grade GFP antibody ( Abcam , ab290 ) was applied . Sonication of fixed testes was performed using Bioruptor sonicator ( UCD-200 , diagenode ) using the following setting: 0 . 5min ON , 1min OFF repetitively for a total of 25min . The size of DNA associated with sonicated chromatin was checked which was approximate 400–500 bp . Similar protocol was used for ChIP with H4 tetra acetylation antibody ( A gift from Keji Zhao , NHLBI , NIH ) and 100 pairs of flies with following genotype: Tj-Gal4/+;E ( Pc ) shRNA/+ or Tj-Gal4/+ . Libraries were generated using reagents provided in the Illumina TruSeq ChIP Sample Preparation Kit ( IP-202-1012 ) . The Illumina compatible libraries were sequenced with Mi-seq desktop sequencer ( Mi-Seq , Illumina ) . Then 75 bp single-end read sequencing was performed . FASTQ raw data files were filtered with quality control software Fastqc ( www . bioinformatics . babraham . ac . uk/projects/fastqc/ ) . BOWTIE program [version 0 . 12 . 7 , [133]] was utilized to align reads to Drosophila genome ( dm3 ) , with the running parameters ( bowtie -p 8 -t -a—phred33 -quals -n 2 -e 70 -l 48 -m 1—best—strata ) . Single-end reads were treated as separate single reads . At each chromosome position , only one read was retained to get a non-redundant read count data . SAM formatted alignment files will be uploaded onto NIH GEO database upon paper acceptance . Enrichment of reads across the genome was analyzed by MACS2 [66] for peak calling . The peak calling was performed with paired experiment ( GFP ChIP ) and control genome input under default parameter settings . UCSC genome browser customized visualization tools were also applied in the analysis [134] . SAMtools [135] software suite was utilized to convert between related read formats . Go term analysis was performed using Gorilla [64 , 65] . The qPCR experiments were performed as previously described [136] . Two independent biological replicates were used . Each PCR reaction was performed in duplicates and averaged Ct values were used . Primers used for qPCR are listed in S4 Table . One pairs of Tj-Gal4 or Tj-Gal4/+; UAS-E ( Pc ) shRNA/+ testes were dissected in PBS , respectively as one replicate . Two replicates were generated for each genotype . Total RNA was purified following the manufacturer’s instruction of PicoPure RNA isolation kit ( KIT0204 , Life technologies ) . Then both libraries were generated using reagents provided in Illumina TruSeq RNA Sample Preparation Kit ( RS-122-2001 ) . The Illumina compatible libraries were sequenced with Illumina Hiseq2500 sequencer in the high-throughput sequencing core facility at Johns Hopkins University Bayview with 50 bp single-end reads . For the alignment to fly genome and gene mapping , sequencing reads were examined by fastqc quality control software ( http://www . bioinformatics . babraham . ac . uk/projects/fastqc/ ) . The reads which passed quality filter were mapped back to Drosophila genome ( dm3 ) ( Flybase dmel_r5 . 43 , as of Jan 2012 , ftp://ftp . flybase . net/releases/FB2012_01/dmel_r5 . 43/ ) . Bowtie aligner ( version 0 . 12 . 7 ) [133] was utilized with the following configuration ( -a—phred33-quals -n 2 -e 70 -l 28 -m 1—best–strata ) which allows two mismatches and only one alignment site . We then assigned each read into gene regions . The annotation for protein coding genes were retrieved from Flybase database ( as of Jan 2012 , ftp://ftp . flybase . net/releases/FB2012_01/dmel_r5 . 43/ ) . The exons from different alternative splicing isoforms were merged to find the maximum genome coverage regions per gene . When a read is mapped to a region with more than one gene , i . e . , one merged exon region overlaps with a non-coding gene , the count is split as equal possibilities into these two genes with half count for each . A matrix file with the number of reads assigned into each gene per sample was prepared for the following data analysis . To identify differentially expressed genes , we utilized the edgeR software package [137] in R to find the normalization factors for each sample with various sizes [by the TMM ( Trimmed Mean of M value ) and upper quantile normalization methods] . The edgeR method models short reads into negative binomial distribution and estimates the biological replicate variance ( dispersion ) . Tag-wise dispersion estimation was performed in “Tj-Gal4” , “Tj-Gal4/+; E ( Pc ) shRNA/+” two groupings of read count profiles . We introduced quantity term “corrected RPKM ( cRPKM ) ” by the formula: pseudo . alt * 1e+09 / ( length of merged transcripts ) / ( common . lib . size ) . The common . lib . size was calculated from the calcNormFactors function of edgeR , which performs TMM and upper quantile normalization methods and set a reference library . The pseudo . alt contains read counts after normalization across the input multiple profiles . The pseudo . alt was calculated by edgeR using quantile normalization and maximum likelihood method . The pseudo . alt contains pseudo read counts after correcting the library size and composition differences . After cRPKM calculation , gene expression levels per sample were pair-wisely compared with spearman correlation ( correlation coefficient rho ) . A pair-wise inter-profile distance was defined as ( 1-rho ) and set up a distance matrix . A dimension reduction method , multidimensional scaling in R ( http://stat . ethz . ch/R-manual/R-devel/library/stats/html/cmdscale . html ) , was utilized to visualize the global similarity relationship of 4 samples . GEO accession number for ChIP-seq and RNA-seq data is GSE93828 .
|
Tissue maintenance and repair rely on adult stem cells , which can divide to generate new stem cells as well as cells committed for becoming specific cell types . Stem cell activity needs to be tightly controlled because insufficient or unlimited stem cell division may lead to tissue degeneration or tumorigenesis . This control depends not only on stem cells themselves , but also on the microenvironment where stem cells reside . The chromatin structure of stem cells is crucial to determine their activities . The signaling pathways connecting stem cells with their microenvironment is also important . Here we ask how chromatin factors interact with signaling pathways in determining stem cell activity . We use Drosophila adult testis as a model system , in which two types of stem cells co-exist and interact: germline stem cells and somatic stem cells . We find that a chromatin regulator called Enhancer of Polycomb [E ( Pc ) ] acts in somatic cells to promote germ cell differentiation and maintain germ cell fate . This regulation is mediated by several signaling pathways , such as EGF and JAK-STAT pathways . E ( Pc ) also works with another chromatin regulator , the histone acetyltransferase Tip60 , in somatic cells . Insufficient activity of the E ( Pc ) homolog in human leads to cancers . Our studies of E ( Pc ) may help understanding its roles as a tumor suppressor .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"rna",
"interference",
"animals",
"cell",
"differentiation",
"germ",
"cells",
"animal",
"models",
"developmental",
"biology",
"drosophila",
"melanogaster",
"model",
"organisms",
"stem",
"cells",
"experimental",
"organism",
"systems",
"epigenetics",
"drosophila",
"research",
"and",
"analysis",
"methods",
"animal",
"cells",
"genetic",
"interference",
"stem",
"cell",
"niche",
"gene",
"expression",
"insects",
"jak-stat",
"signaling",
"cascade",
"arthropoda",
"testes",
"biochemistry",
"rna",
"signal",
"transduction",
"anatomy",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"cell",
"signaling",
"organisms",
"signaling",
"cascades",
"genital",
"anatomy"
] |
2017
|
Enhancer of polycomb coordinates multiple signaling pathways to promote both cyst and germline stem cell differentiation in the Drosophila adult testis
|
Chromosomal structural change triggers carcinogenesis and the formation of other genetic diseases . The breakpoint junctions of these rearrangements often contain small overlapping sequences called “microhomology , ” yet the genetic pathway ( s ) responsible have yet to be defined . We report a simple genetic system to detect microhomology-mediated repair ( MHMR ) events after a DNA double-strand break ( DSB ) in budding yeast cells . MHMR using >15 bp operates as a single-strand annealing variant , requiring the non-essential DNA polymerase subunit Pol32 . MHMR is inhibited by sequence mismatches , but independent of extensive DNA synthesis like break-induced replication . However , MHMR using less than 14 bp is genetically distinct from that using longer microhomology and far less efficient for the repair of distant DSBs . MHMR catalyzes chromosomal translocation almost as efficiently as intra-chromosomal repair . The results suggest that the intrinsic annealing propensity between microhomology sequences efficiently leads to chromosomal rearrangements .
Chromosome structural variations such as deletions , duplications , inversions and chromosomal translocations contribute to evolution and genetic diseases [1] , [2] , [3] . A chromosomal translocation is a fusion between two non-homologous chromosomes . It contributes to cancer by forming a chimeric fusion protein or joining the regulatory region of one gene to the translated region of another gene , causing dysregulated gene expression [4] , [5] . Altered expression of oncogenes or tumor suppressor genes then contributes to the development and progression of tumors [6] . Clinically , the detection of specific chromosomal translocations in patients can help in the diagnosis , treatment selection , and prognosis of the disease [4] , [7] , [8] . Substantial effort is therefore underway to characterize chromosomal translocation breakpoint junctions and the associated genetic changes . Puzzlingly however , much about how chromosomal translocations arise remains poorly understood . Such knowledge could constitute the first step to preventing their occurrence and curbing the chromosomal instability common in cancer cells . DNA damage , in particular the DNA double strand break ( DSB ) , is the inciting event in many chromosomal rearrangements [4] , [5] , [9] , [10] , [11] . Successful repair of DSBs avoids the persistence of toxic DNA lesions and maintains chromosomal integrity [12] . Accordingly , cells and organisms with a compromised DNA repair capacity demonstrate an elevated frequency of chromosomal translocations and chromosomal instability [5] , [9] . Thus , all eukaryotic cells have two main pathways for repairing DSBs: non-homologous end-joining ( NHEJ ) and homologous recombination ( HR ) . NHEJ joins two free DNA ends after a break by direct re-ligation whereas HR uses a homologous template for repair , most typically a sister chromatid [13] , [14] . Nevertheless , neither canonical pathway is fully responsible for the formation of chromosomal translocations , as their repair products do not exhibit some of the key features of chromosomal translocation breakpoints described below [1] , [15] . Recent technological advancement has allowed for the recovery and analysis of many breakpoint junctions from chromosomal rearrangements at the nucleotide level [2] , [3] , [16] , [17] . These studies revealed that the breakpoint junctions often contain a few base pairs ( 2–20 bp ) of overlapping sequences between joining chromosomal ends , and these small overlapping sequences are broadly called “microhomology” [18] , [19] . The frequent presence of microhomology at breakpoints of chromosomal translocations could provide insight into the repair mechanisms used to form these chromosomal aberrations [16] , [17] , [18] , [20] . Nevertheless , the precise mechanisms of microhomology-mediated DSB repair and its role in chromosomal translocations is not yet defined . Recently , multiple microhomology-mediated repair ( MHMR ) pathways have been proposed to explain the usage of microhomology to repair DNA breaks . Microhomology-mediated end-joining ( MMEJ ) represents the Ku-independent end-joining repair process that anneals microhomologous sequences near the broken DNA ends [15] , [21] . Synthesis-dependent-MMEJ ( SD-MMEJ ) creates de novo microhomology by transient templated synthesis at the DNA end [22] . Microhomology also facilitates distinct HR events such as microhomology-mediated synthesis-dependent strand annealing ( MM-SDSA ) that requires non-processive DNA synthesis and a POL32-independent “template switch” mechanism [23] . Microhomology-mediated break-induced replication ( MM-BIR ) operates as the RecA/Rad51-independent BIR to remove a collapsed replication fork , thereby producing the complex rearrangements found in copy number variations [1] . Microhomology/microsatellite-induced replication ( MMIR ) is genetically different from HR , NHEJ and MMEJ , producing segmental duplications from replication based DNA breakage [24] . These results suggest that the usage of microhomology is not a result of one defined repair pathway , but rather , the imprudent repair of breaks or collapsed replication forks relying on the intrinsic stability of annealed microhomology . Yet , the biological principles of microhomology-dependent repair , including how a particular pathway is selected among multiple repair options , are undefined . Previously , we reported that MMEJ is particularly effective at repairing DSBs with non-complementary DNA ends . However , we realized that the previous assay system accidentally introduced a 12 bp imperfect microhomology sequence into the strain with non-complementary DNA ends but not to that with complementary ends , granting an unforeseen advantage [25] . These findings prompted a re-evaluation of the usage of microhomology flanking the DSB for repair . By systematically altering several key parameters , we determined the influence microhomology exerts on DNA DSB repair processes , and its contributions to the formation of reciprocal chromosomal translocations in a budding yeast model . We found that MHMR relies heavily on multiple factors that affect the stability of strand annealing between flanking microhomologous sequences . The stability of strand annealing dictates the efficiency of repair and the genetic factors , or pathway , involved . We also found that microhomology located on non-homologous chromosomes promotes chromosomal translocation . These results may provide the molecular basis for how DNA breaks lead to chromosomal rearrangements .
In order to study microhomology-mediated DSB repair , we devised a simple genetic system to efficiently score the frequency of MHMR events , distinguishing them from non-homologous end joining ( NHEJ ) . Direct repeats of microhomology ( 6–18 bp ) were placed on either side of an HO endonuclease recognition sequence at the MATa locus , flanking the 2 kb hygromycin B phosphotransferase ( HPH ) gene that confers resistance to hygromycin B ( Figure 1A ) . The microhomology sequence on the centromeric side of the DSB was only 2 bp from the HO cut-site . The strains lacked other HO recognition sites typically present in HML and HMR , and the HO gene was expressed under a galactose-inducible promoter ( see Strains List in Table S1 ) . Upon galactose-driven induction of HO , the recognition sequence was efficiently cleaved , creating a DNA DSB ( data not shown ) . We then measured the survival frequency and the hygromycin resistance of the surviving colonies . If the break was repaired by NHEJ , the surviving colony was still hygromycin-resistant , but repair using the direct microhomology repeats led to hygromycin-sensitive survivors ( Figure 1A ) . The types of repair events were further validated by recovering repair junctions by PCR and sequencing them ( Table S2 ) . Most ( 23 out of 24 ) hygromycin resistant events exhibited small base pair additions or deletions at the repair junction , typical of NHEJ events [26] , whereas all hygromycin sensitive events resulted from a deletion of one of the repeats and the inter-repeat DNA , utilizing the microhomology for repair ( Table S2 ) . Deletion of YKU70 or DNL4 dramatically reduced hygromycin resistant survival , further validating the role of NHEJ in hygromycin resistant repair events ( Table 1 ) . The successful development of a MHMR assay prompted us to characterize the necessary features of microhomology flanking a DSB that enable MHMR events . We first tested the effect of microhomology size on the DSB repair efficiency by changing the length of microhomology from 6 to 18 bps . As predicted , the length of microhomology exerted no effect on the NHEJ frequency; the hygromycin resistant ( HygR ) survival frequency was near constant ( ∼0 . 1% ) for all yeast strains tested ( Figure 1B ) . In contrast to NHEJ , the repair of the break using microhomology ( and thus HygS survival ) increased as the length of microhomology increased ( Figure 1B ) . The MHMR frequency of yeast with a DSB flanked by 17 bp microhomology corresponds to ∼10% , while those flanked by 6 or 12 bp is only ∼0 . 00001% . MHMR efficiency increased approximately 10-fold for every additional nucleotide of microhomology between 12 and 17 bp ( Figure 1B ) . Since the length of homology was a critical parameter for MHMR , we determined whether MHMR depends on the homology annealing factor Rad52 , thus corresponding to a HR pathway variant . Surprisingly , deletion of RAD52 led to distinctly different outcomes in MHMR frequency according to the length of microhomology . At the longer lengths of microhomology ( 15–18 bp ) , deletion of RAD52 reduced MHMR frequency 3–10 fold , and a rad52Δ rad59Δ double gene deletion nearly abrogated all MHMR events ( Figure 1C and Table 1 ) . The rad59Δ alone also reduced microhomology-mediated repair , albeit more modestly than the rad52Δ ( a 2 . 5-fold decrease versus a 10-fold decrease ) ( Table 1 ) . However at shorter lengths ( 12–13 bps ) , Rad52 inhibited the usage of microhomology ( Figure 1C ) . The results suggest that the size of microhomology is an important parameter for MHMR , and DSB repair using 15–18 bp microhomology falls under a Rad52-dependent repair mechanism . To initiate the MHMR as seen in Figure 1C , we assumed that one or both microhomologies flanking a DSB should become single-stranded , and thus , DNA end resection is likely required for the repair process . Predictably , both sgs1Δexo1Δ and sgs1Δexo1Δmre11-H125N , that are deficient in end resection [27] , [28] , [29] , demonstrated 10-fold decreases in the MHMR efficiency using the microhomology for repair ( Table 1 ) . Resection deficiency nevertheless increased NHEJ frequency dramatically , and thus , the total survival efficiency in these mutants was reduced only moderately ( Table 1 ) . We then positioned the telomere-proximal-side microhomology of 12 or 18 bp at two different locations from the break ( 60 bp or 2 kb ) and tested whether the location of microhomology from the break affected MHMR frequency ( Figure 2A ) . Since hygromycin resistance cannot discern the NHEJ events from the MHMR events in the strain carrying the microhomology located 60 bp from the break , we relied on sequencing to distinguish MHMR from NHEJ ( Figure 2B ) . We found that the MHMR frequency of a DSB flanked by 18 bp of microhomology 60 bp apart was approximately 2 . 8-fold higher than 18 bp 2 kb apart , while the MHMR frequency of a DSB flanked by 12 bp of microhomology 60 bp apart was approximately 280-fold higher than 12 bp 2 kb apart ( Figure 2B ) . These results suggest that the distance of the microhomology from the DNA break strongly influences the repair efficiency , especially when the size of microhomology is small . The DNA between the microhomology and the break constitutes a 3′ flap upon annealing of microhomology , and therefore , the location of the microhomology from the break dictates the 3′ flap size . Possibly , long non-homologous 3′ flaps destabilize annealing between microhomology and thereby reduce MHMR frequency . We thus examined whether the number and length of 3′ flaps affected MHMR by positioning one of the two microhomologies at several locations: immediately next to the DSB , or 5 , 10 , 15 , 20 , 25 , and 50 bp away from the break , while the other microhomology on the telomere-proximal-side was fixed at 2 kb distal from the break . We found that a second flap size of 5 bp or longer strongly inhibited the MHMR process ( Figure 2C ) . Deletion of RAD1 , a single-strand DNA endonuclease that forms a complex with Rad10 and cleaves 3′ flap DNA [30] , [31] , reduced MHMR frequency 4-fold even in the strain having the single HPH-containing 3′ flap only ( Figure 2C ) . MHMR in the absence of Rad1 was unaffected by the size of the 3′ flap next to the microhomology and occurred with similar efficiency in all of the strains . The results suggest that non-homologous 3′ flap removal is an important step of efficient MHMR and strongly influences survival frequency . The presence of mismatches reduces the efficiency of the HR pathway [32] , [33] , [34] , [35] , [36] , [37] . The presence of mismatches within the microhomology would also likely influence MHMR frequency by destabilizing the annealing process [31] . We examined whether mismatched sequences affect the frequency of MHMR events in our system . We replaced the 18 bp microhomology located 2 kb from the break with that carrying one , two , or three mismatched nucleotides at various positions and then measured the frequency of MHMR upon induction of HO endonuclease ( Figure 3A ) . We found that mismatches effectively suppressed MHMR in both the wild type and the yku70Δrad52Δ mutant , but the inhibition was substantially greater in wild type . As the number of mismatches increased , the yku70Δrad52Δ mutant repaired the DNA break flanked by mismatched microhomology equally well , or even better than wild type , indicating that highly mismatched microhomology can be used to repair a DSB , albeit at a lower efficiency than perfectly matched homology ( Figure 3B ) . The majority of the HygS repair events in yku70Δrad52Δ mutants use the given mismatched microhomology , as confirmed by sequencing of the repair junctions . A few breakpoint junctions were not recovered , likely due to large deletions flanking the DSB using other endogenous microhomology similar to DSB repair events in yku70Δ mutants ( Table S2 ) [25] . These results suggest that mismatches within the microhomology reduce MHMR efficiency . NHEJ has been shown to inhibit repair processes involving microhomology , both in yeast and mammals [21] , [31] , [38] . To test whether NHEJ also inhibits the MHMR process shown here , we expressed HO endonuclease for short intervals in the yeast strain carrying 17 bp of microhomology , and the MHMR frequency was compared in wild type , yku70Δ and dnl4Δ mutants , the last two of which are defective for NHEJ [39] . If NHEJ was inhibiting MHMR during pulsed induction , we expected an increase in the MHMR frequency in the NHEJ mutants . However , the wild type and both NHEJ mutants displayed similar frequencies of MHMR events ( Figure 3C ) . The efficiency of repair was also similar to the repair during continuous HO endonuclease cleavage ( compare Figure 1B and Figure 3C ) . We conclude that NHEJ does not inhibit MHMR in our experimental setting . Previously , we reported that DSB repair by microhomology-mediated end joining occurred preferentially in a strain with two contemporaneous DNA breaks , which created broken DNA ends with no complementary base pairing potential [21] , [25] . However , the previous assay system fortuitously introduced a 12 bp imperfect microhomology sequence in the strain with non-complementary DNA ends , facilitating MMEJ repair [21] . The efficient MHMR in our current studies , using a strain with a single HO cut-site and complementary ends , further challenges our earlier premise that MMEJ is restricted to specific end configurations . To resolve this issue , we constructed several yeast strains carrying one or two HO cut-sites , HO cut-sites from different MAT genes , and with complementary and non-complementary ends ( Figure S1A ) . We found that microhomology could support DSB repair regardless of the end configuration of the induced DNA breaks ( Figure S1B ) . These results broaden the utility of MHMR for the repair of different types of DSBs , regardless of end configuration . Mechanistically , microhomology could mediate DSB repair via mechanisms similar to break-induced replication ( BIR ) or single strand annealing ( SSA ) . Rad51 is central to strand invasion for gene conversion and BIR , but inhibitory for SSA [40] , [41] , [42] , [43] , [44] , [45] . In contrast , Rad59 is essential for SSA and for only a subset of gene conversion/BIR events [42] , [45] , [46] , [47] . Therefore , the dependence of recombination events on Rad51 or Rad59 could offer mechanistic insights into the repair process . As shown in Table 1 , we found that deletion of RAD59 reduced recombination involving 17 bp microhomology nearly 3-fold . In contrast , deletion of RAD51 improved MHMR frequency almost 4-fold . Deletion of both RAD51 and RAD59 rendered MHMR still more efficient than wild type . Deletion of RAD51 also improved the MHMR frequency of rad52Δ rad59Δ to the level commensurate with that of rad52Δ mutant . The results suggest that MHMR may operate similarly to SSA albeit with unique redundancy between Rad52 and Rad59 [46] . We also examined the effect of POL32 deletion on the MHMR frequency . Pol32 is an accessory protein for DNA polymerase δ , and it is dispensable for normal replication but essential for BIR and a subset of gene conversion pathways [48] , [49] , [50] , [51] . Provided that MHMR is a SSA variant , we anticipated that Pol32 should be dispensable for MHMR . Surprisingly , deletion of POL32 severely reduced MHMR frequency , decreasing it more than 12-fold ( Table 1 ) . The role of Pol32 in MHMR is not likely to recruit translesion polymerases because deletion of REV3 and/or RAD30 did not impact MHMR frequency ( Table 1 ) [21] . The results suggest that MHMR is distinct from the established HR pathways as the genetic requirement is not consistent with either SSA or BIR . The inability to discern the type of repair events used for MHMR in our system by genetic tests prompted us to employ another assay listed below . In BIR , mutagenic DNA synthesis proceeds toward the end of the chromosome , yielding a high level of frameshift mutations at a lys2::Ins ( A4 ) gene integrated 36 kb distal from the break [52] . In contrast , SSA likely does not involve repair synthesis at such a distant location and therefore should not lead to an increase in the LYS2 frameshift mutation frequency ( Figure 4A ) . Measuring the LYS2 frameshift mutation frequency could thus help to discern if the repair mechanism involved SSA or BIR . We placed the lys2::Ins ( A4 ) gene 36 kb from the HO cut-site flanked by 17 bp microhomology and measured the frequency of LYS2 reversion ( Figure 4A ) [52] . As predicted , HO expression elevated ( ∼100 fold ) the LYS2 frameshift mutation frequency in the strain AM1291 that repaired a DSB by BIR [52] . However , the LYS2 frameshift mutation did not increase after HO expression in the strain bearing microhomology flanking the break site ( Figure 4B ) . The results suggest that MHMR operates differently than BIR , and likely resembles SSA . Previous studies suggest that BIR and a subset of gene conversion repair processes are the only HR events dependent on Pol32 [50] , [51] . We were puzzled that Pol32 played an important role in the MHMR events resembling SSA , and we further investigated why MHMR in our system relied heavily on Pol32 . In our strain , the microhomology was located two nucleotides away from the HO cut , creating a very short ( 2 bp ) non-homologous tail on the centromeric side of the DNA break . Such a short non-homologous tail could be removed by the proofreading activity of DNA polymerase δ [53] . We reasoned that Pol32 could catalyze MHMR by removing short 3′ flaps as part of the proofreading activity of polymerase δ . To test this idea , we measured the survival of YDV1 . 18 . 0 , which bears a single long ( 2 kb long ) telomeric 3′ flap but lacks a 2 bp 3′ flap on the centromeric side in the pol32 mutant ( Figure S2A ) . We found that microhomology-dependent repair in YDV1 . 18 . 0 was still dependent on Pol32 , suggesting that Pol32 is required for step ( s ) other than short 3′ flap removal ( Figure S2B ) . Alternatively , we hypothesized that Pol32 may be involved in the repair synthesis from the annealed microhomology , further stabilizing the interaction of the annealed DNA duplex . In this model , the shorter the length of homology , the more important Pol32 would be for repair . We examined the survival frequency of the pol32-deleted strains carrying 205 bp ( EAY1141 ) or 1 . 3-kb ( YMV80 ) direct repeat sequences ( Figure S2A and S2B ) . We found that deletion of POL32 decreased survival frequency moderately ( 0 . 64-fold ) in EAY1141 but not at all in YMV80 ( Figure S2B ) . These results suggest that Pol32 is important for recombination using short stretches of homology . The frequent presence of microhomology at breakpoint junctions in chromosomal rearrangements prompted us to test whether MHMR could promote chromosomal translocations [16] , [17] , [18] , [20] . To measure the frequency of MHMR between two non-homologous chromosomes , we placed the 17 bp microhomology sequence on the centromeric side of chromosome III and the telomeric side of chromosome V , flanking HO cut-sites on both chromosomes ( Figure 5A ) . We expected three ways to repair these two breaks: ( 1 ) NHEJ without translocation , ( 2 ) chromosomal translocation by NHEJ , and ( 3 ) chromosomal translocation by MHMR of one junction and NHEJ of the other junction ( Figure 5A ) . With continuous induction of HO endonuclease , the incidence of chromosomal translocation increased from an average of 10 . 7% of survivors suffering an NHEJ-mediated translocation to an additional 43 . 7% of survivors suffering a microhomology-mediated translocation , bringing to total translocation frequency to 54 . 4% of all survivors . Therefore , the presence of the microhomology on the non-homologous chromosome increased the total frequency of chromosomal translocation ( Figure 5B ) . In SSA , inter-chromosomal repair occurs as efficiently as intra-chromosomal repair [54] . We thus tested whether MHMR could catalyze chromosomal translocations as efficiently as intra-chromosomal events . To test this idea , we constructed a yeast strain that could repair both breaks by MHMR intra-chromosomally and inter-chromosomally ( Figure 5C ) . 99 . 3% of survivors lost both HPH and URA3 marker genes and repaired the breaks by MHMR . Importantly , of those survivors , 40 . 8% survived by reciprocal chromosomal translocation ( Figure 5D ) . The MHMR-mediated chromosomal translocation was dependent on end resection because deletion of SGS1 EXO1 severely reduced the frequency of chromosomal translocation ( Table S3 ) . The results indicate that intra- and inter-chromosomal MHMR occur with almost equal frequency , similar to SSA . These results may partially explain why microhomology is so often found at the breakpoint junctions of chromosomal rearrangements .
By inserting various sizes of microhomology with or without mismatches at multiple locations flanking a DSB , we systematically addressed the role of microhomology in the repair of a DNA break and the formation of chromosomal translocations . The results demonstrate that more than one mechanism exists for catalyzing microhomology-mediated repair . Involvement of different mechanisms , carried out by different genetic factors , depends on the location and the length of microhomology . We also demonstrated that MHMR lacks preference for intra-chromosomal repair and promotes high levels of chromosomal translocations . Our results uncovered the surprising complexity of MHMR processes and mutagenic potential . Successful MHMR depends on the size and location of the microhomology and the presence of mismatches , because these parameters strongly affect the frequency and type of MHMR ( Figure 1 , Figure 2 , and Figure 3 ) . The genetic requirements for MHMR are also radically different between those for shorter or longer microhomology; repair of a DSB using microhomology 15 bp and longer is Rad52 dependent whereas repair involving less than 15 bp microhomology is inhibited by Ku and Rad52 ( Table 1 ) . Furthermore , the repair events mediated by 17 or 18 bp of microhomology do not fully conform to the genetic requirements for typical SSA , as they become heavily dependent on Pol32 and either Rad52 or Rad59 . MHMR thus resembles MMIR that repairs broken replication forks and produces Pol32-dependent segmental duplications [24] . These results , along with evidence from several other studies [1] , [15] , [22] , [50] suggest that microhomology directs multiple different repair events , many of which are the variants of established repair mechanisms but with distinct genetic and thus , mechanistic differences . All of these pathways may exploit the stability of annealed microhomology to strengthen the association between broken DNA ends . The size , the degree of homology , and the location with respect to the break all contribute to the thermodynamic energy of strand annealing and dictate the repair mechanism and the repair outcomes , utilizing the biochemical activities of various DNA repair enzymes ( Table S4 ) [15] , [21] . Thus , the most distinguishing feature of this process may be the genetic and mechanistic complexity of MHMR , and we speculate that this complexity may have evolved to deal with a wide range of DNA lesions induced by toxic chemicals and metabolites . However , this adaptability poses a significant challenge for the establishment of its genetic attributes , and similar conclusions were proposed to account for the genetic plasticity of NHEJ [55] . Additionally , these results raise a concern about the validity of some of the earlier results pertaining to MHMR , because these studies did not consider the possibility that MHMR may consist of multiple pathways encompassing widely different genetic and mechanistic requirements . Realization of this complexity challenges the generalization of all repair events that occur in the absence of a certain gene and use various lengths of microhomology as mechanistically common MHMR events . Despite the unique challenges associated with analyzing MHMR events , we demonstrated that MHMR using longer than 15 bp of microhomology in our experimental system operates as the SSA but not the BIR variant . Our conclusion is based on the fact that the mutagenic BIR mechanism would need to replicate the DNA all the way to the end of the chromosome [52] , and our data indicated that there was no such low-fidelity replication occurring 36 kb from the break-site in the lys2::Ins ( A4 ) fluctuation assay . According to the sum of the functional assays and mutant analysis , we have compiled a model of the microhomology-mediated pathway ( Figure 6 ) . The model proposes that resection induces the onset of MHMR , allowing for the single-stranded microhomology to anneal . Rad52 facilitates annealing between microhomology if the microhomology length is 15 bp or longer . However , if the microhomology length is shorter than 15 bp , Rad52 prevents annealing of microhomology and thereby inhibits mutagenic deletions in the DNA . This inhibition of MHMR between short lengths of microhomology may be attributed to the minimum size of ssDNA catalyzed by Rad52 for annealing [56] , yet no such information is available . We also found that Rad59 contributes to MHMR especially when Rad52 is absent . Surprisingly , the deletion of RAD51 offset the role of Rad59 in MHMR , bringing survival frequency of the rad52 rad59 rad51 triple mutant to the equivalent of a single rad52 deletion ( Table 1 ) . Rad59 could thus facilitate MHMR by neutralizing the inhibitory activity of Rad51 , as has been previously reported [57] . Upon successful annealing between microhomology , Rad1/Rad10 cleaves 3′ flap DNA , and Pol32 stabilizes the annealing intermediate between single-strand DNA to allow for Polδ to extend the annealed homologous sequence and complete the repair process ( Figure 6 ) . We suggest that Pol32 likely functions similarly for BIR and a subset of gene conversion events to maximize the stability of strand pairing . Previous work studying the SSA pathway found that repeats of 29 bp were used 0 . 2% of the time to repair a DSB [46] . Yet , our work found that 17 bp repeats efficiently repair the DSB with ∼10% survival efficiency . We hypothesized that this high efficiency of repair was due to the microhomology location at the very end of the DNA , and that 3′ flap removal may hinder this repair process . Accordingly , the DNA sequence at the end of the broken DNA is much more catalytic for microhomology-mediated deletions and translocations , as a 5 bp nucleotide flap inhibits repair 10-fold ( Figure 2C ) . We speculate that the instability of the annealed intermediate in MHMR can be offset by the lack of a second tail or by the extension of annealed sequences via repair synthesis initiated from DNA end without further processing . Our results also showed that MHMR can process a second flap independently of Rad1 , further emphasizing the difference between SSA and microhomology-mediated SSA . The presence of a Rad1-independent mechanism to remove 3′ flaps has been proposed before but the genetics and the mechanisms of such pathway ( s ) are not identified yet . We propose that such pathway ( s ) can efficiently catalyze MHMR with two flaps in the absence of Rad1/10 endonuclease . Most importantly , we showed that the presence of microhomology across the break between different chromosomes dramatically promotes the formation of chromosomal translocations ( Figure 5 ) . The results may explain why women with an increased familial risk of breast cancer and breast cancer patients themselves have a higher frequency of MHMR and SSA repair pathways in their white blood cells [58] . The increased frequency of microhomology-mediated chromosomal translocations is partly due to the lack of bias to intra-chromosomal repair as seen in the repair events by SSA . The results are the direct opposite of NHEJ , which shows a strong bias to intra-chromosomal repair and suppression of chromosomal translocations [59] . Evidence has emerged that ATM-dependent end-tethering suppresses inter-chromosomal end joining and thereby suppresses break-induced chromosomal translocations [59] , [60] . We propose that either end-tethering does not inhibit inter-chromosomal SSA or becomes nonfunctional at the time of the SSA or MHMR process . Regardless , these results shed light on how MHMR induces chromosomal translocations . Evidence indicates that breakpoint junctions of chromosomal rearrangements in humans contain microhomology of 2–20 bp and such events are markedly elevated in NHEJ deficient cells [2] , [15] . The findings that MHMR efficiently catalyzes chromosomal translocations support its contributions to chromosomal translocation formation in humans . However , most breakpoint junctions of chromosomal translocations in human studies show 1–6 bp microhomology , which is much shorter than that used for MHMR events described in this study . To account for this discrepancy , we surmise that the usage of microhomology is dictated not only by the efficiency of such sequence to catalyze MHMR but also the frequency of available flanking microhomology . The low efficiency of MHMR using shorter microhomology can be offset by the high availability of shorter microhomology and thus their frequent appearance at breakpoints in chromosomal aberrations in humans . Additional difference in end processing and/or repair protein activity such as Rad52 or resection enzyme ( s ) among species may disproportionally favor the usage of certain size microhomology in MHMR . Further study into understanding MHMR pathways and their regulation could lead to the etiology of chromosomal aberrations in patients with a higher baseline of mutagenic DNA repair processes .
All yeast strains are derived from JKM139 or JKM179 [26] , [61] . The genotype of JKM139 is hoΔ MATa hmlΔ::ADE1 hmrΔ::ADE1 ade1-100 leu2-3 , 112 lys5 trp1::hisG ura3-52 ade3::GAL-HO . The genotype for JKM179 is hoΔ MATα hmlΔ::ADE1 hmrΔ::ADE1 ade1-100 leu2-3 , 112 lys5 trp1::hisG ura3-52 ade3::GAL-HO ( Table S1 ) . YDV strains and their derivatives were made by amplification of the HPH gene from pAG26 with 90-bp oligonucleotides , containing 20-bp of homology to HPH , various sizes of microhomology sequence , and homology to the Z1 region of MATa on chromosome III . Gene deletion mutants were constructed by PCR-based one step gene deletion technique using oligonucleotides flanked by terminal sequences homologous to the open reading frame of target genes [62] . Logarithmically growing yeast cells were incubated in YEP-Glycerol for 16 hours , and serial dilutions were plated onto YEPD and YEP-galactose plates . Galactose induces HO endonuclease expression [21] . To induce HO expression for shorter duration , 2% ( w/v ) galactose was added to logarithmically growing yeast cells in YEP-glycerol medium , and after the indicated time of incubation , aliquots of culture were removed and plated onto YEPD to inhibit further HO endonuclease expression [59] . Survival frequency was calculated by dividing the number of colonies surviving on YEP-galactose by the number of colonies surviving on YEPD plate . The plates were replica-plated on hygromycin-containing or uracil-deficient plates to determine whether they retained the HPH or URA3 genes , respectively . Logarithmically growing yeast cells in YEP-glycerol media were harvested by centrifugation and re-suspended to a concentration of 5×108 cells/ml . Cells ( 2×108 cells ) were plated on 150 mm lysine drop-out plates containing dextrose or galactose . The median of five strains was taken for each experiment . Each experiment was repeated three times with five strains each , and an average median value was calculated . Spontaneous LYS+ reversion frequency was calculated from the number of colonies on the lysine− plates , and BIR-induced LYS+ reversion frequency was calculated from the number of colonies on the lysine− galactose-containing plates [52] . The entire experiment ( five single colonies per strain ) was repeated two more times for triplicate values . Unless otherwise stated , all experiments were conducted in triplicate , so that an average and standard deviation were calculated . P-values were calculated for mutants as compared to the respective wild type strain using a two-tailed paired t-test .
|
Cancer results from an accumulation of mutations that transform a normal cell into one that proliferates uncontrollably . DNA double-strand breaks ( DSBs ) can lead to genetic mutations and chromosome rearrangements , underscoring the importance of functional DNA DSB repair pathways in the maintenance of chromosome integrity and tumor suppression . Ample evidence suggests that cells possess multiple DSB repair mechanisms with distinct mutational potentials , and one or more of these pathways is likely responsible for the formation of chromosomal translocations . Importantly , at the junctions of many rearrangements , small ( 2–20 bp in length ) overlapping sequences from each of the original sequences , termed “microhomology , ” are found , and they may provide a clue as to how these rearrangements form . Here , we describe our genetic investigation into how flanking microhomology influences the type and frequency of DSB repair . We also show that microhomology-mediated repair ( MHMR ) efficiently induces chromosomal translocations . This research provides a basic understanding of the mechanisms that utilize microhomology for mutagenic repair .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genetics",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
Microhomology Directs Diverse DNA Break Repair Pathways and Chromosomal Translocations
|
Simulation of biomolecular networks is now indispensable for studying biological systems , from small reaction networks to large ensembles of cells . Here we present a novel approach for stochastic simulation of networks embedded in the dynamic environment of the cell and its surroundings . We thus sample trajectories of the stochastic process described by the chemical master equation with time-varying propensities . A comparative analysis shows that existing approaches can either fail dramatically , or else can impose impractical computational burdens due to numerical integration of reaction propensities , especially when cell ensembles are studied . Here we introduce the Extrande method which , given a simulated time course of dynamic network inputs , provides a conditionally exact and several orders-of-magnitude faster simulation solution . The new approach makes it feasible to demonstrate—using decision-making by a large population of quorum sensing bacteria—that robustness to fluctuations from upstream signaling places strong constraints on the design of networks determining cell fate . Our approach has the potential to significantly advance both understanding of molecular systems biology and design of synthetic circuits .
Dynamic simulation is an essential and widespread approach for studying biomolecular networks in cell biology [1] . However , the computational resources required can quickly become limiting for several reasons . Cellular networks are complex , containing many biomolecular species and reactions . The effects of biochemical stochasticity can be pervasive at the single-cell level [2 , 3] , implying that stochastic simulation approaches are often needed . And cells do not live in isolation , which requires simulation on multiple scales , ranging from the single cell to large ensembles of communicating cells [4 , 5] . In these circumstances , parsimonious models of intracellular networks offer dimension reduction [6–8] and significant advantages [9] . However , such models often only provide accurate descriptions when they include the effects of interactions with other fluctuating processes in the cell and of signals arising extracellularly [10–12] . While it is straightforward to write a Chemical Master Equation describing the stochastic dynamics of these models , it is usually impenetrable to analysis and one needs to make use of simulation methods . The stochastic simulation algorithm ( SSA ) [13 , 14] allows only the random timing of reactions in the network model to be taken into account ( often known as intrinsic noise ) , but cannot be used when other processes interacting with the network cause its propensities to fluctuate between reaction occurrences . The SSA assumes constant propensities between reactions ( and hence exponentially distributed waiting times ) . Here we present a new approach relaxing this assumption , called Extrande , for stochastic simulation of a biomolecular network of interest embedded in the dynamic , fluctuating environment of the cell and its surroundings . An extensible implementation of Extrande for general reaction networks with multiple inputs is given in the S1 File . Biological processes that interact with the network or model of interest are sometimes called extrinsic processes [15] . They often significantly change the stochastic behaviour and dynamics of the network [16 , 17] . We briefly give two illustrations of the biological importance of extrinsic processes as motivation for the development of our approach , the first well-established , and the second considered here . First , although intrinsic noise is an important contributor , extrinsic processes are known to be a substantial and sometimes dominant source of variation in gene expression levels across cells and over time [18–21] . We are now beginning to understand the underlying biological sources [22] , which include effects related to circadian oscillations , temperature , chromatin remodelling , the cell-cycle and pulsatile transcription factors [23 , 24] . To understand gene expression , it is therefore essential to move beyond the SSA , which can only account for intrinsic noise , and to include other sources of variation . Second , fluctuations in the expression , degradation and recycling of proteins inevitably affect the way networks containing those proteins function and the extent of stochasticity in the input they provide to other networks . Fluctuations in the component proteins of signal transduction networks limit information transfer [25] , affect transduction network ‘design’ [26] and , although often overlooked , are inevitably conveyed ( as extrinsic inputs ) to the networks regulated by signaling . Here , the computational advantages of Extrande will allow us to demonstrate how fluctuations in the protein componentry of signal transduction networks are conveyed to signaling outputs and place strong constraints on the design of networks determining cell fate , thus influencing the distribution of phenotypes at the population level . Without the ability to simulate biomolecular networks that are exposed to fluctuating inputs , the ability to address such questions is severely restricted . There are two existing approaches to stochastic simulation of reaction networks subject to dynamic , fluctuating inputs . The first class of algorithms [5 , 13 , 27] implements the SSA , under the approximation that the input remains constant between the occurrences of any two reactions . However , this approximation can give spurious results even when dynamic inputs to the network are changing relatively slowly . We term these collectively the Slow Input Approximation method ( SIA ) . The second class of algorithms [28–30] involves step-wise numerical integration of reaction propensities until a target value for the integral is reached . Algorithms in this class would be ( conditionally ) exact , if it were not for the presence of numerical error in integration , but can impose large and impractical computational burdens , especially when cell ensembles are studied . We term these collectively the integral method ( distinguishing next and direct integral approaches below ) . We perform a comparative analysis of both methods with Extrande and demonstrate that our method offers an accurate and computationally efficient alternative approach . Extrande involves no analytical or numerical integration but instead relies on ‘thinning’ techniques [31 , 32] . Other approaches using rejection methods have also recently been proposed as a means to tackle systems with time-dependent propensities [12 , 33] .
The stochastic simulation algorithm ( SSA ) [13 , 14] allows simulation of biomolecular reaction networks taking into account the discreteness of these systems as well as the intrinsic randomness in the timing of reaction events . The SSA assumes that the propensity of each reaction channel to fire , hence the probability of the reaction to occur over a small time interval , remains constant between reaction events . This naturally restrains the use of SSA to simulate networks embedded in dynamic , fluctuating environments because the reaction propensities then become time-varying quantities under the influence of extrinsic processes . Extrande ( Box 1 ) —or Extra Reaction Algorithm for Networks in Dynamic Environments—allows exact stochastic simulation of any downstream reaction network , conditional upon a time course of the dynamic inputs that is simulated up-front . The method involves no analytical or numerical integration , though we give a connection to the direct integral method below , and instead makes use of point process ‘thinning’ techniques [31 , 32] , where some simulated events are discarded . The only error incurred is any error associated with the input pre-simulation , typically an approximate simulation of a stochastic differential equation ( Box 1 ) . The Extrande approach can be understood as introducing an extra , ‘virtual’ reaction channel into the system ( whose occurrence does not change molecule numbers ) . The propensity of the extra channel is designed to fluctuate over time so that ( when added to the sum of all other reaction propensities ) the total propensity in the augmented system becomes constant between events and equal to an upper bound on the sum of the propensities in the original system . To accomplish this , the method exploits the exogeneity of the dynamic inputs—the assumption of negligible retroactivity [35] from network to inputs . In particular , their exogeneity means that Extrande is able to make use of the ‘future’ trajectory of the inputs to find an upper bound , B , on the total propensity , which is valid over a certain time interval L ( see Step 3 , Box 1 ) . Simulation of the augmented system is feasible by means of an SSA-like algorithm . The method uses the bound on the total propensity to generate a putative reaction reaction time τ ( Step 4 ) . If the reaction time exceeds the time horizon L , it is rejected; the system time advances by L ( Step 6 ) , and the procedure restarts by determining a new bound . Otherwise , time advances by τ and a reaction is chosen based on the updated reaction propensities ( at time t+τ ) ( Steps 8–15 ) . The reaction events of the virtual channel are discarded , leaving those of the other channels—because the simulated timing and types of the biochemical reaction channels are unaffected by the behaviour of the extra channel , the result is a trajectory of the original system ( see Methods ) . We study the decision to enter competence ( for uptake of extracellular DNA ) by the model organism Bacillus subtilis . It is well established [39–41] that the source of differentiation of 10–20% of the cell population under stress conditions is fluctuations in transcription of the master competence regulator , ComK . The ComS-MecA-ComK competence module is regulated by the activated transcription factor pComA , the output of the transduction mechanism relaying extracellular , quorum sensing signals ( CSF and ComX ) , see Fig 3A . We study the effect of this upstream signaling on differentiation into the competent phenotype . A useful approach to understanding the structure-function relationship in systems biology is to rewire networks found in nature and compare function with the wild-type , which can then shed light on why apparently similar network structures were not adopted naturally [42] . In the wild-type , upstream signaling acts via activation of the ComS promoter by pComA binding ( Fig 3A , thick black arrow ) . We compare the behaviour of wild-type cells to those with a Synthetic Decision-Making network ( SynDM ) which is regulated , in addition , via activation of the ComK promoter by pComA binding ( red dashed arrow ) . We model ComK-driven progress and entry into functional competence , and write Progress ( t ) = k ∫ 0 t ComK ( s ) d s , where k is an effective rate of ComK-driven differentiation . A cell is taken to enter ( functional ) competence at the time when Progress ( t ) = 1 . The value of the parameter k is set so that the wild-type and SynDM networks give equal fractions of competent cells with a constant level of pComA ( 1000 molecules ) . We tune rate parameters associated with the ComK promoter of the SynDM network so that the fraction of SynDM cells entering competence ( 0 . 18 ) is the same as for wild-type cells , in the absence of fluctuations in pComA levels ( see S1 Text ) . A table listing all reactions and parameter values used in our models of the competence module of wild-type B . subtilis and the SynDM networks is given in the S1 Text . We use the linear noise approximation ( LNA ) [43] to model the the upstream signaling ( with CSF and ComX fixed at steady-state levels ) , giving a mean for pComA of 1000 molecules throughout . Importantly , we include in the model gene expression and degradation of the proteins comprising the signal transduction mechanism because it is now understood that the resultant variation has important effects on signaling and information transfer [26] . A single Ornstein-Uhlenbeck ( OU ) process is sufficient to closely match the mean , variance and autocorrelation function of pComA given by the LNA ( see S1 Text ) . We therefore use a single OU process for the pComA input in what follows . A range of protein lifetimes is considered , consistent with the broad range of cell-cycle periods observed for bacteria under different growth conditions [44] , where nutrient limitation can result in periods in excess of 10h . Our baseline LNA model of the upstream signaling module gives a lifetime and CV of pComA fluctuations equal to 5h and 0 . 35 . We take the pComA input to be exogenous to the ComS-MecA-ComK competence module since it is in high abundance relative to the 2 promoters it binds ( the only interaction between the two modules ) . The importance in determining cell fate of the time taken for the cell to complete different differentiation programs ( to the point of irreversible commitment ) has recently been emphasised [45] . The SynDM network creates a differentiated sub-population by activating the differentiation program in most or all of the cell population ( Fig 3C & 3D ) , with entry to competence the outcome of a ‘race’ to differentiate over the relevant time window . In the SynDM network , binding of pComA to the ComK promoter results more often in periods of non-zero ComK expression than in the wild-type population , but when such periods occur , they are less sustained ( see Fig 3B–3D , and Fig . E in S1 Text ) . The typical rate of progress of a SynDM cell to competence is increased by a higher level of pComA ( see Fig . E in S1 Text ) , and extrinsic fluctuations in the pComA level therefore affect the fraction of cells entering competence ( Fig 3C & 3D ) . In contrast , the wild-type activates the differentiation program in a smaller sub-population , the size of which is under modest regulation by pComA ( Fig 3F ) —a high proportion of the active wild-type cells then enter competence because , once activated , ComK expression rarely deactivates in the wild-type ( see Fig 3B , and Fig . E in S1 Text ) . We find two important advantages of the wild-type design ( in addition to the implied reduction in the metabolic cost of gene expression ) . First , the fraction of cells entering competence is considerably more robust to the fluctuations from upstream signaling in pComA ( Fig 3E ) . For example , with the baseline model of upstream signaling , the SynDM network has a competent fraction ( 40% ) which is more than 2 . 25 times the competent fraction when pComA is held constant at its mean level , whereas the competent fraction of wild-type cells ( 17% cf 18% ) has changed very little . The difference in robustness is explained by the sensitivity of the probability of competence for a SynDM cell as a function of the time average of the signal , 〈pComA〉 , which switches quite rapidly from zero to one ( Fig 3F ) . Since the fraction of competent cells is equal to the average of Prob[Competence|〈pComA〉] over the distribution of 〈pComA〉 ( which is approximately the distribution of pComA for longer lifetimes ) , the competent fraction increases in the presence of extrinsic fluctuations for SynDM ( recall the mean of pComA is 1000 molecules ) . In contrast , Prob[Competence|〈pComA〉] is approximately linear for the wild-type network , which implies that the competent fraction depends largely on the mean of pComA alone . Such plots ( Fig 3F ) should prove a useful diagnostic tool for the design of synthetic decision-making networks . The second advantage of the wild-type design is that the fraction of cells entering competence is also considerably more robust than SynDM to heterogeneity across the cell population in the rate at which ComK-driven differentation proceeds ( Fig 3G ) . The reason is evident from the progress to competence trajectories in Fig 3B–3D . We note that fluctuations from upstream signaling in pComA can also cause decreases in the fraction of competent SynDM cells , as seen for higher rates of differentiation ( Fig 3G ) . Heterogeneity in the rate at which differentiation programs proceed is inevitable where cellular decisions are executed by large gene expression networks and involve substantial physiological changes [46] . These in silico experiments ( Fig 3 ) , made computationally feasible by Extrande , cast light on the wild-type network design in which quorum signaling input to the competence decision-making network ( ComS-MecA-ComK ) by the transcription factor pComA exerts its effect only at the promoter of ComS and not at the promoter of ComK . The experiments reveal exquisite robustness of the wild-type design to fluctuations from upstream signaling and to heterogeneity in downstream processes , and demonstrate the computational potential of Extrande for in silico network design .
Stochastic simulation of biomolecular networks is now indispensable for studying biological systems , from small reaction networks to large ensembles of cells . The effects of stochasticity can be pervasive at the single-cell level , determining the distribution of phenotypes in a population and thus potentially affecting evolutionary outcomes . However , studying such phenomena requires stochastic simulation of a large ensemble of cells that can take into account both intrinsic and extrinsic sources of cellular variation . This can be hugely costly in terms of CPU time , placing important in silico experiments out of reach . Here we provide the new Extrande approach—for stochastic simulation of a biomolecular network embedded in the dynamic environment of the cell and its surroundings—which substantially increases the computational feasibility of such experiments without compromising accuracy . We show that previous approaches to this problem either can fail dramatically , even when inputs vary relatively slowly , or impose impractical computational burdens due to costly numerical integration of reaction propensities . Given a simulated trajectory of fluctuating network inputs , the Extrande approach provides a conditionally exact solution that can speed up simulation by several orders of magnitude compared to integral methods . In practice , we find that integral methods suffer from the high cost of propensity evaluations during numerical integration . Extrande bypasses numerical integration by introducing an extra reaction channel—one designed to keep the total propensity of the ‘augmented’ system constant between events—hence making the problem of finding the time to the next event analytically tractable . Importantly , our numerical results demonstrate that the overhead costs induced by the Extrande method—for example , due to thinning and rejection events , and due to obtaining the ceiling of the input process when a global ceiling is not available–are significantly lower than the cost of accurate numerical integration . In practice , we observe speed-ups by a factor as great as 2 . 5×104 ( Fig 2C ) . Recent work [12] proposes to handle fluctuating environments in a different manner , by deriving a network model for the biochemistry that takes account of the dynamic input and follows the correct ( marginal ) probability law . Explicit simulation of the input is bypassed . The resultant ‘uncoupled’ network model has time-varying reaction propensities and can then be simulated using integral or thinning methods . However , analytical derivation of the uncoupled network model is not always possible , particularly when there are multiple inputs . The accuracy of the method then depends on finding suitable approximation schemes . There are two main limitations of modelling using the Extrande method . The first is that Extrande , being a method of obtaining trajectories of the chemical master equation ( with time-dependent propensities ) , has the same applicability limitations as the master equation; namely there is an implicit assumption that the system is dilute ( point particles ) and well-mixed , conditions which are not met when molecular crowding is significant [47 , 48] . The second limitation is that Extrande assumes that the inputs influence the system of interest but the latter does not influence the inputs ( which implies the inputs can be pre-simulated ) . Hence the method is useful , for example , to understand how certain external stimuli such as light and temperature can affect the stochastic dynamics of a system . For the case of a chemical stimulus , the method can provide an accurate description of the stochastic dynamics if the system and its output do not significantly feedback to adjust the original chemical stimulus , for example by a regulatory mechanism . We exploit the benefits of the proposed Extrande simulation method here to study the decision-making behaviour of a quorum sensing population of bacterial cells . The in silico experiments presented ( Fig 3 ) took approximately two computing months using Extrande ( and an Intel Xeon , 3 . 3GHz quad-core processor with 32GB of RAM ) , but would have been prohibitive using the integral method due to the approximate 70-fold slow down needed to ensure even modest accuracy ( see Fig . D in S1 Text ) . The results elucidate the costs and benefits of alternative network designs for the probabilistic differentiation of a sub-population of cells in response to upstream signaling . Our findings argue for the biological significance of fluctuations in signaling inputs that arise from synthesis and degradation of the protein componentry of signal transduction networks , and show that these fluctuations have important consequences for downstream networks such as those deciding cell fate . We expect the accuracy and reductions in CPU time made possible by Extrande to help open up the landscape of computationally feasible simulation of biomolecular networks and cell ensembles . Extrande thus has the potential to accelerate both understanding of molecular systems biology and the design of synthetic networks .
The Extrande approach relies on augmenting the reaction network with an extra , ‘virtual’ channel ( giving the augmented system , Z ) , so as to make simulation of the augmented system feasible , while ensuring that the simulated timings and types of biochemical reactions are unaffected by the firings of the extra channel . In the Extrande method , the conditional propensity of the extra channel depends on the history of the extra channel ( as well as on the history of the original system , H t X ) , and so does the upper bound . A related Proposition in [32] does not allow for this dependence ( see S1 Text ) . We therefore provide the new proof below . To see the dependence on the extra channel , note that the bound is in general updated in Step 3 of the Extrande algorithm ( Box 1 ) after each firing of the extra channel . The reaction network to be simulated ( Box 1 ) has the number of molecules of each species at time t given by X ( t ) = X ( 0 ) + S R ( t ) , where R ( t ) = {R1 ( t ) , … , RM ( t ) } is the vector of processes counting the number of times each biochemical reaction channel fires during the time interval [0 , t] , and S = {v1 , … , vM} is the stoichiometric matrix . The ‘Poisson’ or random time-change representation [49] expresses R ( t ) in terms of M independent , unit rate Poisson processes , Y ( t ) = {Y1 ( t ) , … , YM ( t ) } , and so can be written here as X ( t ) = X ( 0 ) + S Y 1 ∫ 0 t a 1 [ X ( s ) , I ( s ) ] d s , . . . , Y M ∫ 0 t a M [ X ( s ) , I ( s ) ] d s T , ( 1 ) where I is the possibly multivariate input , superscript T denotes transpose of a vector , and aj[X ( s ) , I ( s ) ] is the propensity of the jth reaction , for j = 1 , … , M , conditional on { H s X , I } . We denote by I ( the σ-field generated by ) the entire trajectory of the input . We introduce as a simulation device the extra , virtual reaction RM+1: ∅ → ∅ , to form the augmented system Z ( t ) = X ( t ) R M + 1 ( t ) = X ( 0 ) 0 + S 0 0 1 R ( t ) R M + 1 ( t ) . The random time-change representation of the augmented system is in terms of ( M+1 ) independent , unit rate Poisson processes , Y ( t ) = {Y1 ( t ) , … , YM+1 ( t ) } Z ( t ) = Z ( 0 ) + ( S 0 0 1 ) × ( [ … , Y j ( ∫ 0 t a j [ X ( s ) , I ( s ) ] d s ) , … ] , Y M + 1 ( ∫ 0 t a M + 1 ( s ) d s ) ) T ( 2 ) where aM+1 ( s ) is the propensity of the extra reaction channel ( conditional on { H s Z , I } ) , and where we set aj[X ( s ) , I ( s ) ] , for j = 1 , … , M , as the propensity of the jth reaction conditional on { H s Z , I } , which now includes the history of the extra channel , RM+1 . Notice that Eq 2 is identical to Eq 1 in its expression of the original system , X ( t ) , or equivalently of R ( t ) . Therefore , if the propensity aM+1 is chosen to somehow make simulation of [R ( t ) , RM+1 ( t ) ] straightforward , we are able to simulate our target , R ( t ) , by simulating the augmented system in Eq 2 and then ignoring RM+1 ( t ) . To do this , let B ( t ) be an ( H t Z , I ) -measurable random variable satisfying ( with probability 1 ) that a 0 ( t ) = ∑ j = 1 M a j [ X ( t ) , I ( t ) ] ≤ B ( t ) , t ≥ 0 , so that B ( t ) is a stochastic upper bound for the total biochemical reaction propensity . Now define the propensity of the extra channel ( conditional on { H t Z , I } ) as: a M + 1 ( t ) = B ( t ) - a 0 ( t ) . The ground process ( see S1 Text ) of [R ( t ) , RM+1 ( t ) ] has propensity ( conditional on { H t Z , I } ) given by ∑ j = 1 M + 1 a j ( t ) = B ( t ) , by construction . The Extrande method chooses the stochastic bound , B ( t ) , so that it is constant between firings of the augmented system ( see Box 1 ) , which makes straightforward the simulation of the ground process of [R ( t ) , RM+1 ( t ) ] . We write the ith occurrence time of the ground process of [R ( t ) , RM+1 ( t ) ] as Ti , i = 1 , 2 , … It is now the case that Prob { T i + 1 - T i ≤ t | T 1 , Z 1 , . . . , T i , Z i , I } = 1 - exp { - t B ( T i ) } , where Zi is the channel corresponding to the ith firing . The waiting time has an exponential distribution and the occurrence times {T1 , T2 , …} are therefore just those of a ( H t Z , I ) -Poisson process with propensity B ( t ) , and can be simulated analogously to the SSA as in Step 4 of Box 1 . What remains is to assign each firing time Ti to one of the ( M+1 ) channels of the augmented system . We do the allocation sequentially , using the result from counting process theory [50] that , for j = 1 , … , ( M+1 ) : Prob { Z i + 1 = j | T 1 , Z 1 , . . . , T i , Z i , T i + 1 , I } = a j [ X ˜ ( T i + 1 ) , I ˜ ( T i + 1 ) ] B ( T i ) , ( 3 ) where we have used the left-continuous versions ( X ˜ ( t ) , I ˜ ( t ) ) of ( X ( t ) , I ( t ) ) , and B ˜ ( T i + 1 ) = B ( T i ) . Eq 3 is implemented by Steps 9–15 in Box 1 . The intuition for Eq 3 uses Bayes’ theorem . Consider a small interval of time dt . The probability that the channel is the jth one given that some reaction fires at time Ti+1 , since the probability of more than one reaction can be neglected , is given by [ d t · a j ( X ˜ T i + 1 , I ˜ T i + 1 ) ] / [ d t · k = 1 M + 1 ∑ a k ( X ˜ T i + 1 , I ˜ T i + 1 ) ] . The target of the Extrande simulation , R ( t ) , is now obtained by ignoring all the firing times of the extra channel after simulation of the augmented system is complete . This completes the proof . ■ We note that the condition limt → ∞ Rj ( t ) = ∞ ( j = 1 , … , M ) is needed for the representation in Eq 1 , but is not needed for the validity of the Extrande method . The random time-change representation is used here to make the proof more accessible . The Extrande algorithm results in a probability law , P , under which the functions aj[X ( t ) , I ( t ) ] give the propensities of the biochemical reactions conditional upon ( H t Z , I ) . Because the aj[X ( t ) , I ( t ) ] are ( H t X , I ) -measurable , they also give the ( H t X , I ) -conditional propensities of the biochemical reactions under P , as required of the probability measure P resulting from the Extrande algorithm . Finally , we remark that a description equivalent to the random time-change representation , Eq 1 , is the Chemical Master Equation [49] . Specifically , for the conditional probability P ( n , t ) = Prob ( X ( t ) = n | X ( 0 ) = n 0 ; I ) one can write d P ( n , t ) d t = ∑ j = 1 M a j [ n - v j , I ( t ) ] P ( n - v j , t ) - a j [ n , I ( t ) ] P ( n , t ) , ( 4 ) whose propensities are time-varying , stochastic functions due to the dependence on the input process .
|
Simulation algorithms have become indispensable tools in modern quantitative biology , providing deep insight into many biochemical systems , including gene regulatory networks . However , current stochastic simulation approaches handle the effects of fluctuating extracellular signals and upstream processes poorly , either failing to give qualitatively reliable predictions or being very inefficient computationally . Here we introduce the Extrande method , a novel approach for simulation of biomolecular networks embedded in the dynamic environment of the cell and its surroundings . The method is accurate and computationally efficient , and hence fills an important gap in the field of stochastic simulation . In particular , we employ it to study a bacterial decision-making network and demonstrate that robustness to fluctuations from upstream signaling places strong constraints on the design of networks determining cell fate .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"signaling",
"networks",
"cell",
"differentiation",
"circadian",
"oscillators",
"simulation",
"and",
"modeling",
"developmental",
"biology",
"mathematics",
"network",
"analysis",
"chronobiology",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"gene",
"expression",
"biophysics",
"physics",
"biochemistry",
"biochemical",
"simulations",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"numerical",
"integration",
"numerical",
"analysis",
"biophysical",
"simulations"
] |
2016
|
Stochastic Simulation of Biomolecular Networks in Dynamic Environments
|
The Caenorhabditis elegans class A and B synthetic multivulva ( synMuv ) genes redundantly antagonize an EGF/Ras pathway to prevent ectopic vulval induction . We identify a class A synMuv mutation in the promoter of the lin-3 EGF gene , establishing that lin-3 is the key biological target of the class A synMuv genes in vulval development and that the repressive activities of the class A and B synMuv pathways are integrated at the level of lin-3 expression . Using FISH with single mRNA molecule resolution , we find that lin-3 EGF expression is tightly restricted to only a few tissues in wild-type animals , including the germline . In synMuv double mutants , lin-3 EGF is ectopically expressed at low levels throughout the animal . Our findings reveal that the widespread ectopic expression of a growth factor mRNA at concentrations much lower than that in the normal domain of expression can abnormally activate the Ras pathway and alter cell fates . These results suggest hypotheses for the mechanistic basis of the functional redundancy between the tumor-suppressor-like class A and B synMuv genes: the class A synMuv genes either directly or indirectly specifically repress ectopic lin-3 expression; while the class B synMuv genes might function similarly , but alternatively might act to repress lin-3 as a consequence of their role in preventing cells from adopting a germline-like fate . Analogous genes in mammals might function as tumor suppressors by preventing broad ectopic expression of EGF-like ligands .
Signaling by epidermal growth factor ( EGF ) family ligands and EGF receptor ( EGFR ) family tyrosine kinases controls many aspects of mammalian development and can drive cancers: EGFRs are commonly overexpressed or constitutively activated by mutations in tumor cells [1] , and EGF-family ligands can be misregulated in cancer . For example , the EGF-family ligands heparin-binding EGF-like growth factor , amphiregulin , and TGF-α are upregulated in cancer cells from many different cancer types [2] , [3] , and TGF-α overexpression causes widespread epithelial hyperplasia in mice [4] , [5] . Growth factors often signal through a Ras pathway , and approximately 20% of tumors carry a constitutively active Ras mutation [6] . In the nematode Caenorhabditis elegans the EGF-family ligand LIN-3 acts through the EGFR LET-23 and the Ras protein LET-60 to control many aspects of development , including the induction of the hermaphrodite vulva [7]–[10] . In wild-type animals , of a set of six equipotent cells , three ( P5 . p , P6 . p and P7 . p ) adopt vulval cell fates , while the other three ( P3 . 4 , P4 . p , and P8 . p ) adopt non-vulval fates [11] . The expression of vulval cell fates requires EGF/Ras signaling , and mutations that reduce EGF/Ras signaling cause a vulvaless ( Vul ) phenotype in which none of the six cells adopts vulval cell fates [7]–[10] . The anchor cell , located closest to P6 . p , is the only cell that both expresses LIN-3 EGF and is located near the six Pn . p cells [10] , and laser ablation of the anchor cell results in a Vul phenotype [12] like that seen in mutants defective in lin-3 EGF or let-23 EGFR . Overactivation of the EGF/Ras pathway , by overexpression of lin-3 EGF or by an activating mutation in either let-23 EGFR or let-60 Ras , causes a multivulva ( Muv ) phenotype in which all six Pn . p cells adopt vulval cell fates [8]–[10] , [13] . In vulval development , EGF/Ras signaling is antagonized by the synthetic multivulva ( synMuv ) genes . The synMuv genes define two classes , A and B [14] , [15] . In synMuv single mutants or in class A double mutants or class B double mutants , vulval development is mostly normal . By contrast , animals mutant in both a class A synMuv gene and a class B synMuv gene exhibit a strong Muv phenotype . Many class B synMuv genes have homologs that function in histone modification , chromatin remodeling , and transcriptional repression . For example , the class B synMuv genes encode a DP/E2F/Rb complex [16] , [17] , a nucleosome remodeling and deacteylase ( NuRD ) complex [18] , [19] , two histone methyltransferases [20] , [21] and a heterochromatin protein 1 homolog [22] . Of the three molecularly-characterized class A synMuv genes , two encode proteins with a zinc-finger-like THAP domain [23]–[25] . The expression patterns of three class A synMuv proteins have been studied , and all three are localized to the nucleus , suggesting that class A synMuv proteins regulate transcription [25] , [26] . The synMuv genes function at least in part by repressing expression of lin-3 EGF . Loss-of-function mutations in either let-23 EGFR or lin-3 EGF can suppress the synMuv phenotype [16] , [27] , [28] , indicating that the synMuv genes act upstream of or in parallel to lin-3 . Furthermore , lin-3 mRNA levels are increased in synMuv double mutants but not in synMuv single mutants [27] , and overexpression of lin-3 EGF causes a Muv phenotype [10] . Laser ablation of the anchor cell , the source of LIN-3 in wild-type vulval development , does not fully suppress the Muv phenotype of synMuv double mutants [28] , indicating that synMuv genes cannot simply prevent overexpression of lin-3 from the anchor cell . Mosaic analyses of the class B synMuv gene lin-37 and the lin-15 locus , which contains both a class A and a class B synMuv gene , did not identify a single site of action . Both experiments indicated that lin-15 and lin-37 do not act cell-autonomously in the Pn . P cells and suggested that lin-15 and lin-37 might function in the syncytial hypodermal cell hyp7 [29] , [30] . Heterologous expression experiments showed that the class B synMuv gene lin-35 functions in hyp7 to antagonize vulval cell fates , and tissue-specific RNAi of lin-3 in hyp7 can suppress the synMuv phenotype , indicating that repression of lin-3 in the hypoderm is an important function of the synMuv genes [27] , [31] . Another study using the same heterologous promoters found that the class B synMuv gene hpl-2 functions in both hyp7 and the Pn . p cells [32] . However , it is not known where lin-3 is overexpressed in synMuv mutants , how the synMuv genes control lin-3 expression , or if the synMuv genes control targets other than lin-3 important for vulval development . Here we report the identification of a lin-3 EGF promoter mutation that causes a dominant class A synMuv phenotype . The effect of this mutation reveals that the only major role of the class A synMuv genes in vulval development is to repress lin-3 . We find that lin-3 mRNA is ectopically expressed throughout the animal in synMuv mutants . Our results show that low levels of ectopic lin-3 expression outside the cells that normally produce and respond to lin-3 can adversely alter the development of C . elegans , and we propose that the class A and class B synMuv genes might prevent ectopic lin-3 expression by distinct mechanisms .
During a screen for new class A synMuv mutations , we identified a Muv animal in the F1 generation after ethyl methanesulfonate ( EMS ) mutagenesis of the class B synMuv mutant lin-52 ( n771 ) . We named the mutation that caused this defect n4441 . To seek additional mutations that like n4441 dominantly cause a class A synMuv phenotype , we screened approximately 492 , 000 F1 progeny of lin-52 ( n771 ) animals mutagenized by EMS and approximately 89 , 000 progeny of animals mutagenized by N-ethyl-N-nitrosourea ( ENU ) , but we did not identify any additional class A synMuv mutants . As a single mutant , n4441 animals are wild-type at 20°C and exhibit a low penetrance Muv defect at 25°C ( Table 1 ) , comparable to that of most class A synMuv mutants [15] . Double mutants between n4441 and the class B synMuv mutations lin-15B ( n744 ) , lin-52 ( n771 ) , or lin-61 ( n3447 ) exhibit a strong synMuv phenotype . n4441 causes a fully penetrant Muv defect as a heterozygote in the class B synMuv mutant background lin-15B ( n744 ) , indicating that n4441 dominantly causes a class A synMuv phenotype . n4441 causes a 97% penetrant synMuv defect in the weak class B synMuv mutant background lin-61 ( n3447 ) at 22 . 5°C , comparable to the previously reported phenotype of double mutants between lin-61 ( n3447 ) and the strong class A synMuv mutations lin-15A ( n767 ) or lin-38 ( n751 ) [15] . To determine how n4441 interacts with other class A synMuv mutations , we built double mutants between n4441 and an allele of each known class A synMuv gene . We used the putative null alleles lin-8 ( n2731 ) , lin-15A ( n767 ) , and lin-56 ( n2728 ) and the missense allele lin-38 ( n751 ) , since a null allele of lin-38 causes lethality ( A . M . S and H . R . H . , unpublished results ) . At 20°C and 25°C , the double mutants n4441; lin-15A ( n767 ) , lin-38 ( n751 ) ; n4441 , and lin-56 ( n2728 ) ; n4441 were enhanced for the Muv phenotype when compared to their respective single mutants ( Table 1 ) . The lin-8 ( n2731 ) ; n4441 double mutant was roughly comparable to n4441 alone when scored at 25°C and also exhibited a low penetrance Muv defect at 20°C , which neither n4441 or lin-8 ( n2731 ) did on their own ( Table 1 ) . Thus , mutations in all known class A synMuv genes can enhance the Muv phenotype of n4441 , but the enhancement is much weaker than the enhancement caused by class B synMuv mutations . Several members of a Tip60/NuA4 histone acetyltransferase complex were previously identified as class C synMuv genes [33] . Class C synMuv genes are strongly Muv in combination with class A synMuv mutations and weakly Muv in combination with class B synMuv mutations and can be considered a subset of the class B synMuv genes [15] . To test if n4441 might be a class C synMuv gene , we built a double mutant between n4441 and the partial loss-of-function class C synMuv mutation mys-1 ( n3681 ) , as null mutants of mys-1 cannot be maintained as homozygous strains [33] . The mys-1 ( n3681 ) ; n4441 double mutant exhibited a 56% penetrant Muv defect at 20°C , which is much stronger than the 5% penetrant Muv defect of the n4441; lin-15A ( n767 ) strain at 20°C , despite the fact that lin-15A ( n767 ) is a null mutation . We conclude that n4441 is not a class C synMuv mutation . By performing SNP mapping experiments using the CB4856 polymorphic strain of C . elegans , we mapped the n4441 mutation to a 661 kb region containing approximately 170 genes between SNPs dbP6 and uCE4-1148 ( Figure 1A ) . n4441 dominantly causes a synMuv phenotype and thus might well be a gain-of-function mutation , so we sought loss-of-function mutations in the gene affected by n4441 . n4441/nT1[qIs51]; lin-15B ( n744 ) animals , which display a fully penetrant Muv defect , were mutagenized with EMS . The nT1[qIs51] translocation causes inviability when homozygous and suppresses recombination across an interval that includes lin-3 [34] . Approximately 6 , 800 F1 progeny were screened , and two animals were identified that were non-Muv and produced only non-Muv progeny , indicating that they contained a suppressor mutation tightly linked to n4441 . We named these mutations n4929 and n4951 . n4441 n4929; lin-15B ( n744 ) animals were sterile and exhibited a very low penetrance Muv defect ( Figure 1D ) . n4441 n4951/nT1[qIs51] animals were superficially wild-type with no Muv defect , and n4441 n4951 homozygotes died as L1 larvae with a rod-like appearance ( Figure 1D ) . The rod-like lethal phenotype is characteristic of loss-of-function mutations in genes in the EGF/Ras pathway required for vulval induction [35] . The only known gene in the EGF/Ras pathway in the genetic interval containing n4441 is lin-3 , which encodes the EGF ligand . Strong loss-of-function alleles of lin-3 cause a rod-like lethal phenotype , and lin-3 mutations can also cause sterility [36] , [37] . The n4929 mutant carries a G-to-A transition in the first nucleotide of exon 8 of lin-3 and is predicted to mutate an arginine to a lysine at amino acid 347 of LIN-3 ( Figure 1B ) . The n4951 mutant carries a G-to-A transition that results in a nonsense mutation predicted to truncate LIN-3 after only 26 amino acids , before the EGF domain ( Figure 1B ) . The lin-3 ( n1059 ) nonsense mutation failed to complement the sterility caused by n4929 and the lethality caused by n4951 , proving that n4929 and n4951 are alleles of lin-3 . Since the lin-3 ( n4951 ) nonsense mutation suppressed the n4441 synMuv defect in cis , but the lin-3 ( n1059 ) nonsense mutation did not suppress the n4441 synMuv defect in trans ( Figure 1D ) , a lin-3 loss-of-function mutation is a cis dominant suppressor of n4441 , indicating that n4441 is a gain-of-function allele of lin-3 . We determined the sequences of all exons and introns of lin-3 and of approximately 11 kb of upstream DNA in lin-3 ( n4441 ) mutants . The only mutation was a G-to-A transition at nucleotide 30904 of cosmid F36H1 , approximately 200 bp upstream of the lin-3 transcript F36H1 . 4a ( http://www . wormbase . org , release WS200 , 20 Mar 2009 ) ( Figure 1C ) . To show that the F36H1 ( 30904 ) mutation is required for the class A synMuv phenotype caused by lin-3 ( n4441 ) , we sought recombinants between lin-3 ( n4441 ) and the lin-3 ( n4951 ) nonsense mutation , which is 5 . 3 kb downstream of F36H1 ( 30904 ) . We screened approximately 90 , 000 progeny from lin-3 ( n4441 n4951 ) /+; lin-15B ( n744 ) animals , identified five independent Muv animals and established homozygous lines . None of the five lines contained the lin-3 ( n4951 ) mutation , and all five carried the F36H1 ( 30904 ) G-to-A mutation . Thus , the lin-3 ( n4441 ) mutation that causes the class A synMuv phenotype must be to the left of lin-3 ( n4951 ) , because if it were to the right then the recombinants would not carry the F36H1 ( 30904 ) mutation . The 5 . 3 kb between F36H1 ( 30904 ) and lin-3 ( n4951 ) , as well as 10 . 8 kb of DNA upstream of F36H1 ( 30904 ) , carried no additional mutations in lin-3 ( n4441 ) animals . If the mutation that causes the lin-3 ( n4441 ) synMuv phenotype is not the F36H1 ( 30904 ) mutation , then the lin-3 ( n4441 ) mutation must be at least 10 . 8 kb to the left of the F36H1 ( 30904 ) mutation . However , in that case , assuming a constant recombination rate throughout the lin-3 interval , the likelihood that all five recombination events would have occurred between F36H1 ( 30904 ) and lin-3 ( n4951 ) is ( ( 5 . 3 ) / ( 5 . 3+10 . 8 ) ) 5 , or <0 . 004 . We conclude that the G-to-A mutation at nucleotide 30904 of cosmid F36F1 is necessary for the class A synMuv phenotype caused by lin-3 ( n4441 ) . However , we cannot rule out the possibility there is a second mutation more than 11 kb upstream of lin-3 that is also required along with the F36H1 ( 30904 ) mutation to cause a class A synMuv phenotype . There are no known consensus transcription factor binding sites that include the site of the lin-3 ( n4441 ) mutation ( Transfac database of known transcription binding sites; http://www . gene-regulation . com ) . The region surrounding the lin-3 ( n4441 ) mutation is moderately conserved in the related nematodes C . briggsae and C . remanei ( data not shown ) . lin-3 ( n4441 ) might be a class A synMuv specific allele of lin-3 . Alternatively , lin-3 ( n4441 ) might cause weak overexpression of lin-3 if weak overexpression of lin-3 behaves like a class A synMuv mutation . To differentiate between these alternatives , we overexpressed lin-3 weakly using the syIs12 integrated transgene . syIs12 expresses the EGF domain of lin-3 under the control of a heat-shock promoter [38] . At 20°C in the absence of heat-shock , syIs12 did not cause a Muv phenotype ( Table 2 ) . syIs12; lin-15B ( n744 ) animals were mostly wild-type , with only a 1% penetrant Muv defect , whereas syIs12; lin-15A ( n767 ) animals exhibited a Muv defect with 40% penetrance ( Table 2 ) . Thus , weak overexpression of lin-3 from the syIs12 transgene was enhanced by a class A synMuv mutation but not by a class B synMuv mutation . By contrast , lin-3 ( n4441 ) was enhanced much more strongly by class B synMuv mutations than by class A synMuv mutations ( Table 1 ) . We conclude that lin-3 ( n4441 ) is a class A synMuv specific allele of lin-3 and does not simply cause weak overexpression of lin-3 . The class A and B synMuv genes redundantly repress expression of lin-3 mRNA [27] . To test if the lin-3 ( n4441 ) mutation affects lin-3 mRNA levels similarly to other class A synMuv mutations , we assayed lin-3 mRNA levels using real-time RT-PCR . As previously reported , the class B synMuv mutant lin-15B ( n744 ) has wild-type lin-3 levels ( Figure 2B ) . The class A synMuv mutants lin-15A ( n767 ) and lin-3 ( n4441 ) both had slightly increased levels of lin-3 mRNA . The synMuv double mutants lin-15AB ( e1763 ) and lin-3 ( n4441 ) ; lin-15B ( n744 ) had substantially increased lin-3 mRNA levels ( Figure 2B ) . Therefore , the lin-3 ( n4441 ) mutation behaves as a class A synMuv mutation with respect to lin-3 mRNA repression . The lin-3 ( n4441 ) mutation is located 211 bp upstream of lin-3 and is also 465 bp upstream of F36H1 . 12 , which is upstream of lin-3 in the opposite orientation ( Figure 2A ) . To determine if lin-3 ( n4441 ) or other synMuv mutations also affect expression of F36H1 . 12 , we assayed F36H1 . 12 mRNA levels by real-time RT-PCR . F36H1 . 12 mRNA levels were roughly equivalent to those of the wild type in all possible single and double mutant combinations involving lin-15A ( n767 ) , lin-3 ( n4441 ) , and lin-15B ( n744 ) ( Figure 2C ) . Therefore , the synMuv proteins specifically repress lin-3 and do not establish a broad domain of repression . Although lin-3 is overexpressed in synMuv double mutants [27] ( also , Figure 2 ) , it is not known where this overexpression occurs . GFP- and LacZ-tagged lin-3 repetitive transgene arrays have been used as reporters for lin-3 expression [10] , [39] , [40] , but these reporters might not be appropriate for determining lin-3 expression in synMuv mutants: first , the level of ectopic lin-3 expression might be too low to visualize using a GFP reporter; second , many synMuv mutations affect the expression of repetitive transgene arrays , potentially confounding interpretation of the expression pattern of such reporters [41] . Instead , we assayed lin-3 expression using a fluorescence in situ hybridization ( FISH ) technique that has sufficient sensitivity to detect single mRNA molecules [42] . We used 48 non-overlapping probes against lin-3 ( Table S1 ) , each conjugated to a single fluorophore , to label individual mRNA molecules brightly enough to be visible as distinct fluorescent spots . Because there are 48 probes that bind independently to the target mRNA , any single probe that binds non-specifically should not cause a false-positive signal . The distribution of intensities of the spots in any given animal was unimodal , consistent with each spot's representing a single mRNA molecule ( Figure S1 ) . Furthermore , by comparing the spot intensities in different tissues and mutants , we found that the level of expression in a given cell or tissue was independent of the intensity of the spots in that cell or tissue , and if the number of spots in a cell was altered then the average intensity of spots in that cell was unchanged ( Figure S2 ) . If each spot represented multiple mRNA molecules , then as the expression level in a given cell increased the average number of mRNA molecules in each of those spots would also be expected to increase , leading to greater intensity . Because the intensity of each spot was independent of the level of expression , we conclude that each spot is likely to represent a single mRNA molecule . We also found that all tissues are accessible to FISH probes , as probes directed against ama-1 and eft-2 robustly detected mRNA in all cells ( data not shown ) . However , we cannot know if we are detecting every mRNA molecule; it is possible that some mRNA molecules are not accessible to the oligonucleotide probes or are not detected for some other reason . We first determined the expression pattern of lin-3 in wild-type animals at the late L2 to early L3 stage when vulval induction occurs . Previous studies found that at the early L3 stage lin-3 is expressed in the anchor cell and in the pharynx [10] , [40] . We indeed observed robust expression of lin-3 in the anchor cell and throughout the pharynx . We also saw expression of lin-3 in the germline ( Figure 3A and Figure S3 ) . In some wild-type animals we also observed a few copies of lin-3 mRNA in one or more cells in the tail , on the ventral side slightly anterior to the anus . In addition , a few copies of lin-3 mRNA were seen on the ventral side of the animal , slightly behind the posterior gonad arm . We imaged several animals that were slightly older , in the late L3 stage , and observed expression of several copies of lin-3 mRNA in the region where P6 . p and its descendants are located ( data not shown ) , consistent with previous reports of expression of lin-3 in the descendants of P6 . p by the L4 stage [39] . We did not consistently detect any lin-3 mRNA in other tissues , although in some animals we observed a single lin-3 mRNA molecule elsewhere . For example , in the animal shown in Figure 3A a single lin-3 mRNA molecule was observed in or near an intestinal cell close to the anchor cell . Overall , other than for those tissues that highly expressed lin-3 there was very tight repression of lin-3 . The numbers of copies of lin-3 mRNA we observed in each tissue in individual animals are listed in Table S2 and are summarized in Table 3 . The expression pattern we observed for lin-3 is consistent with that seen using GFP- and LacZ-tagged lin-3 reporters [10] , [39] , [40] and with functional studies of lin-3 [43] , indicating that most if not all of the mRNA spots identified by this technique are likely to represent actual lin-3 mRNA molecules . lin-15AB ( e1763 ) animals expressed lin-3 in the pharynx , germline , and anchor cell at levels grossly similar to those of wild-type animals ( Figure 3D and Table 3 ) . In addition there was widespread ectopic expression of lin-3 , with an average of approximately 1100 ectopic copies of lin-3 mRNA observed per animal ( Figure 3D and Table 3 ) . This ectopic expression was much weaker than the normal expression in the anchor cell; whereas an average of 29 copies of lin-3 mRNA was seen in the anchor cell in wild-type animals ( Table 3 ) , only one or a few copies of lin-3 mRNA were observed in most cells in lin-15AB ( e1763 ) mutants . Because we could not see cell boundaries , we could not determine if every cell ectopically expressed lin-3 , but there were no tissues that appeared to lack ectopic lin-3 mRNA ( Figure S4 ) . Cells around the perimeter of the animal expressed lin-3 in the lin-15AB ( e1763 ) mutant , consistent with ectopic expression in the hypodermis ( Figure 3D and Figure S4 ) . There were also many ectopic lin-3 mRNA copies that clearly were not in the hypodermis ( Figure S4 ) . We also determined lin-3 expression in lin-15A ( n767 ) and lin-15B ( n744 ) single mutants . lin-15B ( n744 ) animals had a lin-3 expression pattern similar to that of wild-type animals ( Figure 3C ) . In lin-15B ( n744 ) mutants there was an extremely low level of ectopic lin-3 expression , with an average of six ectopic lin-3 mRNA molecules detected per animal ( Table 3 ) , but lin-3 was still tightly repressed outside of the germline , anchor cell , and pharynx . lin-15A ( n767 ) animals exhibited broad ectopic expression of lin-3 , but at a much lower level than that of lin-15AB ( e1763 ) animals ( Figure 3B ) . An average of 64 copies of lin-3 mRNA were seen outside of the pharynx , germline , and anchor cell in lin-15A ( n767 ) animals ( Table 3 ) . Unlike in lin-15AB ( e1763 ) animals , in any given lin-15A ( n767 ) animal most cells did not display ectopic lin-3 expression . However , we observed no obvious cell or tissue specificity to the ectopic expression among several lin-15A ( n767 ) animals . Rather , it appeared that in lin-15A ( n767 ) animals lin-3 is globally derepressed , but at a very low level . The numerous class B synMuv genes have highly similar although not identical effects on vulval development [15] . However , the class B synMuv genes have widely differing effects on other aspects of growth and development . For example , PGL-1 , which is normally expressed in the germline , is misexpressed in the somatic cells of mutants of many class B synMuv genes , including lin-15B , but not in mutants of some other class B synMuv genes , including lin-36 , lin-52 , and lin-53 [44] , [45] . We therefore investigated the role of the class B synMuv genes lin-36 , lin-52 , and lin-53 in controlling lin-3 expression . We determined the expression pattern of lin-3 in lin-36 ( n766 ) ; lin-15A ( n767 ) , lin-52 ( n771 ) ; lin-15A ( n767 ) , and lin-53 ( n833 ) ; lin-15A ( n767 ) mutants . All three double mutants exhibited ubiquitous ectopic expression of lin-3 , with lin-3 mRNA observed in most if not all tissues ( Figure 4 ) . There were no obvious differences in the spatial pattern of lin-3 expression in lin-36 ( n766 ) ; lin-15A ( n767 ) , lin-52 ( n771 ) ; lin-15A ( n767 ) , and lin-53 ( n833 ) ; lin-15A ( n767 ) double mutants as compared to lin-15AB ( e1763 ) mutants . The lin-3 ( n4441 ) mutation could cause global derepression of lin-3 similarly to lin-15A ( n767 ) , or it could affect lin-3 expression in a subset of tissues . We examined the expression of lin-3 mRNA in lin-3 ( n4441 ) and lin-3 ( n4441 ) ; lin-15B ( n744 ) animals . lin-3 ( n4441 ) animals had widespread but weak ectopic expression of lin-3 , similar to lin-15A ( n767 ) animals ( Figure 5A and Table 3 ) . lin-3 ( n4441 ) ; lin-15B ( n744 ) animals exhibited ectopic lin-3 expression in most cells and were indistinguishable from lin-15AB ( e1763 ) animals ( Figure 5B and Table 3 ) .
Identifying the biologically relevant targets of transcriptional regulators that control development is a challenging problem . The synMuv genes encode putative transcriptional repressors that prevent ectopic vulval development . Mutating a synMuv binding site in a target gene might relieve repression of that target , and if that repression were essential to prevent ectopic vulval development could cause a dominant synMuv phenotype . We isolated a mutation in the lin-3 EGF gene that derepresses lin-3 transcription and causes a dominant class A synMuv phenotype . This finding establishes that lin-3 is a functionally important target of the class A synMuv genes , consistent with a previous report that lin-3 expression is repressed by the synMuv genes and that double-stranded RNA directed against lin-3 can suppress the synMuv phenotype [27] . Importantly , the lin-3 ( n4441 ) mutation fully recapitulates the class A synMuv phenotype with regard to vulval development and lin-3 expression and causes a class A synMuv phenotype equivalent to that caused by strong alleles of class A synMuv genes . If the class A synMuv genes repressed multiple targets to prevent ectopic vulval development , then a mutation that abolished class A synMuv-mediated repression of lin-3 would recapitulate only partially the class A synMuv phenotype . We conclude that lin-3 is likely to be the only key biologically relevant target of the class A synMuv genes in vulval development . The simplest interpretation of the effect of the lin-3 ( n4441 ) mutation is that this mutation abolishes a binding site for a transcriptional repressor consisting of or controlled by class A synMuv proteins . However , the effect of the lin-3 ( n4441 ) mutation is slightly enhanced by mutations in all other class A synMuv genes . If the lin-3 ( n4441 ) mutation completely inactivated a binding site that responds to only one of the known class A synMuv proteins , then mutation of that class A synMuv gene should not enhance the synMuv phenotype caused by lin-3 ( n4441 ) . One possibility is that a complex consisting of multiple class A synMuv proteins binds to the lin-3 promoter , the lin-3 ( n4441 ) mutation strongly reduces but does not completely eliminate that binding , and removing any one class A synMuv protein does not fully abrogate the ability of the complex to bind to the lin-3 locus and repress transcription . This model is consistent with the observation that most class A synMuv mutations , including lin-3 ( n4441 ) , are enhanced by class A synMuv mutations in other genes [15] . Alternatively , the class A synMuv genes might indirectly repress lin-3 by regulating the expression or activity of or by binding to another protein that binds to the lin-3 promoter to prevent ectopic transcription . Because a mutation in the lin-3 promoter can cause a class A synMuv phenotype , the class A and class B synMuv pathways must be integrated at the point of lin-3 repression , and hence it is unlikely that the class A and B synMuv genes redundantly control a transcriptional regulator which in turn controls lin-3 expression . lin-3 expression in the germline had not been previously observed , likely because the reporters used to assay lin-3 expression were either silenced in the germline [46] or lacked distant regulatory regions necessary to drive germline expression . Mutations in the FOG and FBF translational inhibitor RNA-binding proteins cause a germline-dependent Muv phenotype , and the FBF proteins can bind to the 3′ UTR of lin-3 in vitro , suggesting that germline lin-3 mRNA is translationally repressed during the larval stage when vulval induction occurs [43] . In many class B synMuv mutants , somatic cells express normally germline-specific genes [19] , [44] , [45] . Given our finding that lin-3 is normally expressed in the germline , one possibility is that the class B synMuv genes repress ectopic lin-3 expression in somatic cells as a consequence of their role in ensuring that somatic cells do not inappropriately adopt germline-like fates . The class B synMuv genes might all directly repress lin-3 in somatic cells . Alternatively , as there are a large number of class B synMuv genes and their effects on vulval development are not identical , perhaps at least some class B synMuv genes indirectly repress lin-3 by preventing the ectopic adoption of germline-like fates . In class B synMuv single mutants , the somatic cells adopt a more germline-like fate that would include lin-3 expression except that the class A synMuv genes still tightly repress lin-3 , mostly preventing ectopic lin-3 expression . In class A synMuv single mutants , lin-3 is not tightly repressed , but most somatic cells are not fated to express lin-3 , so there is only a low level of leaky ectopic lin-3 expression . However , in class AB synMuv double mutants , the somatic cells adopt a germline-like fate that includes lin-3 expression , and there is no class A synMuv mechanism that tightly represses lin-3 , resulting in widespread and substantial ectopic lin-3 expression . In short , we suggest that the synthetic Muv phenotype caused by mutations in the synMuv genes might be a consequence of two distinct functions of the class A and class B synMuv genes: the class A synMuv genes either directly or indirectly tightly repress ectopic lin-3 transcription , and the class B synMuv genes prevent somatic expression of germline-expressed genes , which include lin-3; only if both functions are lost will somatic cells ectopically express sufficient lin-3 mRNA to cause ectopic vulval induction . These findings raise the possibility that the development of some human tumors might require the loss of one tumor suppressor gene that prevents cells from adopting a fate that is permissive for the expression of a growth factor and the loss of a second tumor suppressor gene that specifically represses the expression of that growth factor . A subset of the class B synMuv genes is required to prevent the somatic misexpression of normally germline-restricted P-granule proteins such as PGL-1 [44] , [45] . We found that lin-15B mutants , which do exhibit somatic PGL-1 expression , and lin-36 , lin-52 , and lin-53 mutants , which do not exhibit somatic PGL-1 , all have highly similar effects on lin-3 expression . These results indicate that different germline genes are broadly repressed in the soma by different sets of transcriptional repressors . The class B synMuv genes define one such group of repressors and are classified together because they have comparable effects on the germline gene lin-3 , resulting in similar vulval phenotypes . Many such partially-overlapping groups of transcriptional repressors , including various subsets of the class B synMuv genes , are likely to be required for the repression in the soma of other germline-restricted genes . Whereas lin-3 expression in most cells is tightly repressed by the synMuv genes , the anchor cell and germline exhibit robust lin-3 expression that is not substantially affected by the synMuv genes . While it has not been reported whether or not any synMuv genes are expressed in the anchor cell , several synMuv genes are expressed in the germline [17] , [22] , [26] , [33] , and we are not aware of studies that have conclusively shown any synMuv genes not to be expressed in the germline . In most cells , the synMuv genes reduce lin-3 expression from an average of one to two copies per cell to nearly zero copies per cell . The synMuv genes clearly do not have a similar fold effect on lin-3 expression in the anchor cell and germline . However , it is possible that the synMuv genes repress a similar absolute number of leaky lin-3 mRNA molecules in all cells; given the animal-to-animal variability in lin-3 expression we likely would not have been able to detect such a small increase in the anchor cell or germline . Alternatively , the synMuv genes might not repress lin-3 in the anchor cell or germline . The strong activator ( s ) of lin-3 that drive expression in those tissues could override the activity of the synMuv genes , or one or more synMuv genes might not be expressed in those tissues , thereby compromising synMuv repression of lin-3 . In lin-15AB mutants , lin-3 is ectopically expressed throughout the animal in a broad range of cells and tissues . Site-of-action experiments have shown that the synMuv genes function at least in large part in the hyp7 hypodermal syncytium to prevent ectopic vulval development [31] . The expression pattern of lin-3 in synMuv mutants does not directly identify the site-of-action of synMuv genes in regulating vulval development but does show that the synMuv genes function throughout the animals to keep lin-3 very tightly repressed in numerous cells and tissues . lin-3 EGF regulates non-vulval cell fates in C . elegans development , and at least some of these fates , such as the P11/P12 fate , are also regulated by the synMuv genes in a manner analogous to that of vulval development [47] . In short , the synMuv genes act throughout the animal to prevent ectopic lin-3 expression , which can cause a variety of developmental abnormalities . Mutants with a displaced anchor cell show that lin-3 can act at a distance [48] , so a cell ectopically expressing lin-3 could affect fates in both nearby and distant cells . We suggest that for any given cell-fate decision , the site of action of the synMuv genes is likely to be spread across multiple cells and determined by the size and proximity of those cells to the cell being regulated by lin-3 . In the case of vulval development , hyp7 plays the major role , given its large size and close proximity to the Pn . p cells , with likely lesser contributions from many other cells . The site at which the synMuv genes repress lin-3 to ensure proper vulval development is therefore probably a combination of the Pn . p cells themselves and neighboring cells that do not normally either express or respond to lin-3 . This situation is similar to that in which both tumor cells and the microenvironment surrounding the tumor provide factors that drive tumor development [49] . We suggest that analogously to the synMuv genes some tumor suppressor genes function by repressing growth factor expression in both tumor cells and the surrounding microenvironment . In synMuv double mutants , lin-3 was ectopically expressed but at a much lower level than at its major normal site of function , the anchor cell . synMuv double mutants might ectopically express as few as one to two copies of lin-3 mRNA per cell . Thus , normal C . elegans development requires lin-3 to be exceedingly tightly repressed outside of a few cells , and only slight expression of lin-3 throughout the animal can cause abnormal cell-fate transformations . Such low levels of ectopic expression would likely be missed by most techniques used to assay gene expression . We suggest it could be important to examine the expression of EGF-family ligands in tumors using highly sensitive techniques with single-molecule resolution to determine if broad low-level misexpression of EGF-family ligands plays a role in oncogenic growth . In C . elegans , the tight repression of lin-3 EGF requires both the class A synMuv gene pathway and the class B gene synMuv pathway , which includes homologs of known tumor suppressor genes , such as lin-35 Rb . Therefore , some tumor suppressor genes in mammals might function by tightly repressing low-level ectopic expression of EGF-family ligands in many cells , possibly in both the tumor and the microenvironment surrounding the tumor .
C . elegans strains were cultured by standard methods on OP50 bacteria [50] . All animals were grown at 20°C , except where otherwise noted . The wild-type strain was N2 , except in SNP mapping experiments in which the polymorphic CB4856 strain was also used [51] . The following mutations were used in this study: LGI: dpy-5 ( e61 ) , lin-61 ( n3447 ) , lin-53 ( n833 ) LGII: lin-8 ( n2731 ) , lin-56 ( n2728 ) , lin-38 ( n751 ) , syIs12 LGIII: dpy-17 ( e164 ) , lin-36 ( n766 ) , unc-32 ( e189 ) , lin-52 ( n771 ) LGIV: lin-3 ( n4441 ) , lin-3 ( n4929 ) , lin-3 ( n4951 ) , lin-3 ( n1059 ) LGX: lin-15A ( n767 ) , lin-15B ( n744 ) , lin-15 ( e1763 ) The balancer strain nT1[qIs51] IV∶V [34] was used; qIs51 is a GFP-expressing transgene integrated onto the nT1 translocation . Table S3 lists all strains used in this study . Synchronized animals were harvested at or near the L2-to-L3 larval transition , when vulval induction occurs . N2 animals were harvested 33 hours after starved L1 larvae were placed on plates with food . Some mutants grew more slowly and were harvested after 39 hours . Quantitative RT-PCR for lin-3 was performed as previously described [15] . lin-3 was amplified using the primers CGCATTTCTCATTGTCATGC and CTGGTGGGCACATATGACTC . Animals were grown to the L2-to-L3 transition as in the quantitative RT-PCR experiments . Fixation and hybridization were performed as described previously [42] , except that worms were fixed for one hour instead of 45 minutes . The lin-3 probes ( Biosearch Technologies , Inc ) were conjugated to the fluorophore Cy5 using the Amersham Cy5 Mono-reactive Dye pack ( GE Healthcare ) . DNA was visualized using 4′ , 6-diamidino-2-phenylindole ( DAPI ) . The probe sequences used are shown in Table S1 . Figure 2 and Figure 3 are maximum intensity projections of a Z-stack of images processed with the Find Edges and Smooth operations in ImageJ . lin-3 mRNA spots were computationally identified with manually determined thresholds as previously described [42] . The number of molecules within each tissue were then manually counted . The anchor cell was identified based on position . lin-3 mRNA expression in the anchor cell appeared as a tight cluster of spots; molecules within that cluster were considered to be in the anchor cell . To determine which lin-3 mRNA molecules were in the pharynx , we noted the outline of the pharynx that was clearly visible as a dark boundary in the Cy5 channel of the image stacks ( see Figure S3 and Figure S4 ) . The boundaries of the germline were estimated from the positions of DAPI-labeled germline nuclei .
|
Extracellular signals that drive cells to divide must be carefully restricted so that only the correct cells receive those signals . Failure to properly control the expression of signaling molecules can lead to aberrant development and cancer . Studies of vulval development in the nematode Caenorhabditis elegans have helped define various multi-step signaling pathways involved in cancer . Here we report that two groups of proteins that control the EGF/Ras/MAP kinase pathway of vulval development act by tightly repressing the spatial expression of the gene lin-3 , which encodes an EGF-like signaling molecule . Using a technique that detects single mRNA molecules , we show that inactivation of these proteins causes a low ectopic expression of lin-3 in many cells . In response , the EGF/Ras/MAP kinase pathway is activated in cells normally not exposed to the lin-3 signal , and vulval development is abnormal . This process is analogous to the cancerous growth that occurs in humans when mutations cause both tumor cells and the microenvironment surrounding the tumor cells to ectopically express factors that drive cellular proliferation . We propose that mammalian genes analogous to those that repress lin-3 expression in C . elegans vulval development act as tumor suppressors by preventing broad ectopic expression of EGF-like ligands .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"animal",
"genetics",
"gene",
"regulation",
"cell",
"differentiation",
"animal",
"models",
"developmental",
"biology",
"caenorhabditis",
"elegans",
"model",
"organisms",
"organism",
"development",
"molecular",
"development",
"molecular",
"genetics",
"pattern",
"formation",
"gene",
"expression",
"organogenesis",
"biology",
"morphogens",
"signaling",
"genetics",
"genetics",
"and",
"genomics",
"cell",
"fate",
"determination"
] |
2011
|
The Caenorhabditis elegans Synthetic Multivulva Genes Prevent Ras Pathway Activation by Tightly Repressing Global Ectopic Expression of lin-3 EGF
|
In most sexually reproducing organisms , the fundamental process of meiosis is implemented concurrently with two differentiation programs that occur at different rates and generate distinct cell types , sperm and oocytes . However , little is known about how the meiotic program is influenced by such contrasting developmental programs . Here we present a detailed timeline of late meiotic prophase during spermatogenesis in Caenorhabditis elegans using cytological and molecular landmarks to interrelate changes in chromosome dynamics with germ cell cellularization , spindle formation , and cell cycle transitions . This analysis expands our understanding C . elegans spermatogenesis , as it identifies multiple spermatogenesis-specific features of the meiotic program and provides a framework for comparative studies . Post-pachytene chromatin of spermatocytes is distinct from that of oocytes in both composition and morphology . Strikingly , C . elegans spermatogenesis includes a previously undescribed karyosome stage , a common but poorly understood feature of meiosis in many organisms . We find that karyosome formation , in which chromosomes form a constricted mass within an intact nuclear envelope , follows desynapsis , involves a global down-regulation of transcription , and may support the sequential activation of multiple kinases that prepare spermatocytes for meiotic divisions . In spermatocytes , the presence of centrioles alters both the relative timing of meiotic spindle assembly and its ultimate structure . These microtubule differences are accompanied by differences in kinetochores , which connect microtubules to chromosomes . The sperm-specific features of meiosis revealed here illuminate how the underlying molecular machinery required for meiosis is differentially regulated in each sex .
During either sperm or oocyte production , meiotic chromosomes undergo a continuum of similar events that are tightly regulated by the cell cycle . Meiosis starts with an extended G2 phase called meiotic prophase in which chromosomes first shorten ( leptotene ) , then pair and assemble synaptonemal complexes ( SC ) ( zygotene ) before completing recombination ( pachytene ) . Chromosomes then disassemble their SC ( diplotene ) and fully condense their bivalents ( diakinesis ) . A subsequent transition from G2 to M is mediated by cell cycle kinases , including POLO and cdk-cyclin B , which drive nuclear envelope breakdown ( NEBD ) , meiotic spindle assembly , and chromosome remodeling . Lastly , during M phase , two rounds of chromosome segregation generate haploid gametes with homologs segregating during the first ‘reductive division’ and sister chromatids segregating during the second . Since kinetochores of sister chromatids must orient towards the same spindle pole during the reductive division , some level of cohesion must be maintained between sister chromatids . Ultimately , proper meiotic segregation necessitates the combined activities of several regulatory proteins , including the Aurora B kinase [1] , [2] . Concurrently , each sex executes the distinct developmental programs of spermatogenesis or oogenesis . Although there is still much to learn , comparative studies have documented several differences between meiosis of spermatogenesis and oogenesis . During meiotic prophase , germ cells interact with distinct gonadal environments [3]–[5] and are differentially regulated by apoptosis and cell cycle checkpoints [6]–[9] . For example , spermatocytes and oocytes vary in requiring an external signal to trigger the G2 to M transition , and many meiotic programs include a diapause at the end of meiotic prophase during which chromosomes aggregate to form a single , transcriptionally down-regulated mass called a karyosome [10] , [11] . Later , during meiotic divisions , spermatocyte chromosomes segregate on centrally positioned centriole-based spindles to form four equally sized haploid spermatids [12] while oocyte chromosomes segregate on tiny , asymmetrically-positioned , acentriolar spindles to generate a single haploid oocyte and 2–3 degenerate polar bodies [8] . Challenges specific to spermatogenesis include the segregation of unpaired and/or heteromorphic sex chromosomes [13] , [14] and the hypercompaction of the haploid sperm chromatin by systematic replacement of somatic histones with both histone variants and diverse protamine and protamine-like proteins [15] , [16] . Several features make Caenorhabditis elegans ideal for analyzing sex-specific differences in meiosis . Many key proteins required for meiosis are evolutionarily conserved from worms to mammals [17]–[19] . Cells progressing through meiosis can be followed in a linear array along the length of the tube-like gonad in either isolated gonads or through the transparent body wall [20] . In hermaphrodites , a common pool of germ cells can generate either sperm or oocytes [21] . Studies of C . elegans oogenesis have provided insights regarding homolog pairing , meiotic recombination , desynapsis , and preparing gametes for meiotic divisions; in addition , they have identified key molecular markers for each meiotic stage [22]–[28] . Studies of C . elegans spermatogenesis have demonstrated its many assets as a model system , including a simplified differentiation program that occurs in the absence of accessory somatic cells or an extended post-meiotic differentiation period . Spermatogenesis-specific mutants can be studied in either males or hermaphrodites , which produce 200–300 sperm before switching to oocyte production [21] , [29] , [30] . However since few studies of C . elegans spermatogenesis have focused on meiotic prophase , molecular studies of this period will expand our understanding of fundamental events of meiosis and sex-specific modifications required in each sex . The goal of this study was to explore how spermatogenesis-specific features coordinate with or modify the basic C . elegans meiotic program . In past studies , investigators have faced several challenges in linking underlying molecular events with cytological observations in late spermatogenesis . First , the rapid progression makes short-lived stages challenging to visualize in fixed preparations . Second , it is difficult to differentiate fine changes in the morphology of small meiotic chromosomes . To overcome these obstacles , we optimized preparation methods and identified molecular markers that differentiate specific stages of sperm meiosis . These markers define a broad set of cytological and molecular landmarks and enabled us to construct a detailed timeline of late meiotic prophase during C . elegans spermatogenesis . While this study identifies many aspects of meiosis that are common to both spermatogenesis and oogenesis , it also identifies multiple spermatogenesis-specific features . Our observations provide a foundation for understanding not only how cell-signaling pathways converge to control cell cycle progression and pace during meiosis but also how underlying molecular processes are differentially regulated between males and females .
In C . elegans , germ cells commit to oogenesis or spermatogenesis upon transition from mitosis to meiosis [31] but it was unknown when sex-specific differences in chromosome morphology could first be detected . To address this , we compared DAPI-stained nuclei in gonads isolated from adult males and hermaphrodites . Germ cells progress through early stages of meiosis while attached to a shared central core of cytoplasm known as the rachis [21] . Examination of nuclei undergoing DNA replication in the distal “mitotic region” , meiotic homolog alignment during leptotene/zygotene stages ( crescent-shaped nuclei in the transition zones ) , or synapsis during the pachytene stage ( basket-shaped nuclei ) failed to reveal any obvious sex-specific differences in either nuclear size or shape ( Figure 1A and 1B ) [32] . Following pachytene , oocytes undergo a sequence of events that lead to their maturation [25] , [27] . In late pachytene , many oocytes are culled by physiological germline apoptosis [33] . Surviving oocytes enlarge as they acquire large quantities of mRNA and protein from neighboring pachytene cells via cytoplasmic bridges [34] ( Figure 1A ) . These oocytes become positioned to one-side and proceed single-file through the proximal gonad . At the same time , the chromosomes further condense to form compact bivalents . Thus , we refer to the region that includes the diplotene and diakinesis stages during both oogenesis and spermatogenesis as the ‘condensation zone’ ( Figure 1 ) . During this period , oocytes detach from the rachis [35] then breakdown their nucleolus [25] , [28] , [36] . They also down-regulate global transcriptional activity , as suggested by a dramatic decrease of serine 2 phosphorylation on the heptad repeat of the C-terminal domain of the large subunit of RNA polymerase II ( pCTD-ser2 ) , [37] , [38] . As the nuclear envelope expands , oocyte chromosomes detach from the nuclear envelope as separate entities [21] . In response to a sperm-derived signal , the oocyte closest to the spermatheca , referred to as the -1 oocyte , undergoes NEBD then assembles its meiotic spindle [25] , [26] , [39] . In contrast , post-pachytene spermatocytes in the condensation zone undergo a distinct series of morphological changes . First , the lack of physiological germline apoptosis [33] and dramatic cell growth enables spermatocytes to proceed in several single file rows around the rachis ( Figure 1B and Figure 2A ) . During early chromosome condensation , chromosomes fail to separate into distinct entities . Instead they aggregate into a single mass within which individual chromosomes are not discernable , although DNA staining by DAPI often appears non-uniform ( Figure 2A and 2C ) . Spermatocytes with this aggregated chromosome morphology are the most prevalent cells within the condensation zone ( 12–24/gonad ) suggesting that spermatocytes exist in this state for an extended period ( Figure 2A ) . To distinguish where this aggregated chromosome stage fits within the meiotic program in C . elegans , we correlated its occurrence with cytologically observable events before and after its formation . First , we used both differential interference contrast ( DIC ) optics and epifluorescence to examine non-fixed , flattened male gonads stained with the DNA dye Hoechst 33258 ( Figure 2D and 2E ) . All spermatocytes in the condensation zone were attached to the rachis , while those in the adjacent division zone were detached . Proximal rachis-attached spermatocytes contained aggregated chromosomes within intact nuclear envelopes ( Figure 2D and 2E and Figure 3A ) . Immunolocalization using antibodies against nuclear envelope markers , including lamin and nuclear pore complex proteins ( Figure 4A and data not shown ) confirmed this observation . Notably , nuclear envelope volume and shape remained relatively constant throughout the condensation zone , suggesting that nuclear envelope reduction is not driving chromosome aggregation . In parallel immunolocalization experiments using anti-α-tubulin antibodies , extensive microtubule networks throughout the cytoplasm also distinguished rachis-attached ( Figure 3A ) from rachis-detached spermatocytes transitioning to M-phase , which possessed prominent microtubule asters ( Figure 3B ) . Similar meiotic structures called “karyosomes” or “karyospheres” , in which paired homologs aggregate during or after diplotene , have been described during sperm and oocyte formation in organisms ranging from Drosophila to humans [10] , [40] , [41] . While karyosome function remains unclear , karyosome formation is most commonly associated with oogenesis in other organisms and is hypothesized to facilitate pre-division chromosome remodeling and grouping prior to meiotic divisions [10] , [40] . Because both the morphology and timing of chromosome aggregation in C . elegans spermatogenesis correlates with karyosome formation in other organisms , we heretofore refer to this stage as the karyosome stage ( Figure 2 , Figure 3 , and Figure 4 ) . After karyosome formation , spermatocytes detach from the proximal end of the rachis and rapidly enter meiotic divisions . In most gonads , we detected 0–2 nuclei transitioning from the karyosome stage to metaphase in two distinct stages . We define spermatocytes in late diakinesis as those that have newly detached from the rachis but possess intact nuclear envelopes and a slight degree of separation between their individual bivalents ( Figure 2 and Figure 4A ) . Relative to chromosomes within oocytes at diakinesis ( late condensation zone in Figure 1A ) , chromosomes within these spermatocytes at late diakinesis remained tightly confined within the smaller nucleus ( Figure 2A–2C ) . We define prometaphase spermatocytes as those that have undergone partial or complete NEBD but whose chromosomes have not fully congressed to the metaphase plate ( Figure 3B ) . We also routinely observed one or more metaphase I spermatocytes in each gonad ( Figure 2 and Figure 3C ) . To better understand karyosomes , we characterized the molecular events leading up to their formation . Prior to karyosome formation C . elegans sperm nuclear basic proteins ( SNBPs ) are incorporated into late pachytene chromosomes ( [17] and data not shown ) . Immunolocalization studies using anti-fibrillarin ( FIB-1 ) antibody revealed nucleoli abruptly disappearing before karyosome formation in late diplotene ( Figure 4B ) . As in oocytes [28] , nucleolar breakdown in spermatocytes coincided with the initiation of histone H3 ( ser10 ) phosphorylation ( pHisH3-ser10 ) ( Figure 4C ) , a histone modification associated with pre-M-phase chromosome condensation [42] . pHisH3-ser10 levels were low in late diplotene but increased rapidly such that karyosome nuclei were the brightest staining within the gonad . Together , the incorporation of SNBPs , nucleolar breakdown and elevated pHisH3-ser10 levels suggest karyosome spermatocytes are largely transcriptionally down-regulated . Consistent with this interpretation , high levels of pCTD-ser2 in pachytene and diplotene nuclei abruptly decreased as karyosomes formed ( Figure 4D ) [38] . Karyosome formation therefore correlates with a decrease in global transcriptional activity . To understand how karyosome formation fits in with events of late meiotic prophase , we examined karyosome formation relative to synaptonemal complex ( SC ) disassembly . The SC , a proteinaeous scaffold , assembles prior to pachynema to facilitate and regulate recombination [18] , [43] . SCs are composed of two structures referred to as axial/lateral elements and central elements . Lateral/axial elements are composed of proteins , such as HIM-3 , that polymerize between sister chromatids along each homologous chromosome length [44] . Central region proteins , like SYP-1 , link the axes between homologous chromosomes [45] , and the loss of SYP-1 marks desynapsis . Interestingly , oocytes and spermatocytes differ in the dynamics of central element disassembly . In post-pachytene oocytes , SYP-1 becomes progressively restricted to axes distal to the chiasmata with regions of SYP-1 retained in all but the -1 oocyte . Complete SYP-1 removal occurs only after nucleolar breakdown and the appearance of pHisH3-ser10 [24] , [25] , [42] ( Figure 5A oocyte ) . In post-pachytene spermatocytes , SYP-1 undergoes a phased departure through diplotene but disappears prior to karyosome formation , well before chromosomes are fully condensed ( Figure 5A sperm ) . Notably , SYP-1 removal in late diplotene spermatocytes occurs just before nucleolar breakdown and the appearance of pHisH3-ser10 staining ( Figure 4B and 4C , Figure 5A , and data not shown ) . Thus SYP-1 is lost earlier in spermatocytes than it is in oocytes relative to meiotic stage , nucleolar breakdown , and degree of chromosome compaction . Such sex-specific differences in the timing of desynapsis relative to other late meiotic events may reflect fundamental differences in sperm chromatin composition and structure and hint at additional underlying sex-specific alterations in the corresponding meiotic machinery . The observation that karyosome formation initiates directly following SC disassembly suggested a possible link between the two events . To test whether SC formation was a prerequisite for karyosome formation , we analyzed karyosome formation in males with mutations in core SC elements . Despite their SC defects [45] , [46] , both syp-1 ( me17 ) and him-3 ( gk149 ) males produced spermatocytes that assembled karyosomes ( Figure 5D and data not shown ) . Thus , the presence of core SC components , and presumably the assembly of the SC , is not a prerequisite for spermatocyte karyosome formation . In contrast to SYP-1 , the lateral element component , HIM-3 , remains associated with oocyte chromosomes during the coiling-based process of diplotene chromosome shortening ( Figure 5B oocyte ) [24] , [44] , [45] . As such , we studied HIM-3 localization in spermatocytes as a marker for chromosome organization . HIM-3 assembled along the lengths of pachytene chromosomes and remained associated throughout the condensation zone . HIM-3 decorated the karyosome in distinct stripes then shifted to an X-shaped pattern along chromosome axes as individual bivalents resolved , similar to the localization pattern observed on chromosomes of oocytes at diakinesis ( Figure 5B sperm ) . This pattern suggests that homologs remain aligned during and following karyosome formation . Although HIM-3 was previously reported to persist at high levels on metaphase I spermatocyte chromosomes [44] , we detected a dramatic metaphase reduction of HIM-3 levels . We suspect that previous investigators , using only DNA or immunostaining of SC components , were unable to distinguish karyosome from metaphase I spermatocytes . In this adjusted analysis , HIM-3 localization patterns are similar during oogenesis and spermatogenesis . Another key player in meiotic progression is the aurora-like kinase AIR-2 [47] . As one of several “chromosomal passenger proteins” , AIR-2 mediates meiotic and mitotic chromosome condensation , chromosome-kinetochore attachments , sister chromatid release , and cytokinesis . During oogenesis , AIR-2 colocalizes with SYP-1 along the axes of pachytene chromosomes and then departs during early diplotene ( Figure 5C oocyte ) [24] . AIR-2 reassociates to the short arms of the bivalents in the -1 oocyte in the same pattern as the recently departed SYP-1 only when a signal from sperm residing in the spermatheca triggers the G2 to M transition ( Figure 5C oocyte ) [1] , [24] , [28] , [48] , [49] . Without sperm , hermaphrodites accumulate oocytes at diakinesis that are AIR-2 negative [49] . During spermatogenesis , AIR-2 localized along the axes of pachytene chromosomes and departed during early diplotene as in oogenesis ( Figure 5C sperm ) . However , AIR-2 exhibited a phased reassociation to chromosomes during karysome formation reaching high levels on late karyosome nuclei and concentrating in discrete regions visible on the external surface . As spermatocytes detached from the rachis and transitioned to prometaphase , AIR-2 localized on the short arm of the bivalents [2] , [48] . This shift in AIR-2 localization may reflect active cycles of AIR-2 unbinding and rebinding the chromatin or , alternatively , the passive movement of chromatin-bound AIR-2 on structurally dynamic chromosomes . Overall , we found AIR-2 reassociation in spermatogenesis is distinct from oogenesis – it is highly dynamic , signal independent , and occurs earlier . The apparent “exchange” of SYP-1 for AIR-2 within oocytes at late diakinesis suggested SYP-1 guides AIR-2 localization [18] , [24] . However , the temporal gap between SYP-1 loss and AIR-2 rebinding during spermatogenesis seemed inconsistent with this model . To test the dependency of AIR-2 chromosome association on SYP-1 , we examined AIR-2 localization in syp-1 ( me17 ) homozygous mutant males and hermaphrodites . In syp-1 spermatocytes , AIR-2 was undetectable on pachytene chromosomes , yet AIR-2 still reassociated with karyosomes ( Figure 5D ) . However instead of concentrating in discrete regions , AIR-2 bound diffusely thoughout the syp-1 karyosome mass , before binding unevenly to chromosomes at diakinesis . In syp-1 ( me17 ) hermaphrodites , AIR-2 was undetectable in oocytes at the pachytene and diakinesis stages [24] ( Figure 5D oocyte and data not shown ) . To distinguish whether the lack of AIR-2 staining in -1 oocytes in syp-1 mutants reflected the lack of “guiding” SYP-1 or the lack of a sperm signal from defective syp-1 sperm , we mated syp-1 ( me17 ) mutant hermaphrodites to wild-type males . In the presence of normal sperm , the syp-1 oocytes were triggered to mature and AIR-2 was detectable on -1 oocyte chromosomes , albeit in an abnormal pattern ( Figure 5D oocyte ) . These results suggest that SYP-1 is required in a sex-independent manner to recruit AIR-2 to pachytene chromosomes . Later , SYP-1 is dispensable for recruiting AIR-2 to chromosomes but required for properly localizing AIR-2 to the short arms of bivalents at diakinesis . Our analysis also refines the role of AIR-2 in phosphorylating serine 10 of histone H3 during the transition to meiotic divisions . Previous work has shown AIR-2 is required for HisH3-ser10 phosphorylation in both maturing C . elegans oocytes [42] and mouse spermatocytes [50] . However , in C . elegans spermatocytes , pHisH3-ser10 not only appeared earlier than AIR-2 but the two markers also exhibited distinct localization patterns on chromosomes during the diplotene , karyosome , and diakinesis stages . This suggests that another kinase is responsible for HisH3-ser10 phosphorylation in late diplotene and karyosome spermatocytes ( Figure 5C , Figure 3B and 3C , and Figure 4C ) . When the two proteins co-localize during prometaphase , AIR-2 may assume the role of phosphorylating HisH3-ser10 ( Figure 4C ) . The G2 to M transition marks the end of diakinesis and an irreversible commitment to meiotic divisions [25] , [27] , [28] . Typical transitional events include NEBD , changes in microtubule dynamics , centrosome separation , and several pre-division chromosome modifications . In C . elegans oocytes , the G2 to M transition initiates with nucleolar breakdown and HisH3-ser10 phosphorylation in the -3 and -2 oocytes followed by AIR-2 recruitment and NEBD in the -1 oocyte [28] . Oocytes of C . elegans and most other organisms lack centrioles [51]–[53] , thus their meiotic G2 to M transition does not involve centrosome nucleation and separation , and chromosome-mediated spindle assembly initiates only after NEBD . Because spermatocytes have centrioles , we anticipated that microtubule reorganization would mark the G2 to M transition . In all diplotene and most karyosome spermatocytes , immunostaining for SPD-2 , a core component of both active and inactive centrosomes [52] , [54] , revealed pairs of tiny , side-by-side SPD-2 foci ( quiescent centrosomes ) situated on one side of the nucleus ( Figure 4E ) at a stage when the spermatocyte cytoplasm was filled with unfocused microtubules ( Figure 3A and data not shown ) . In late karyosome spermatocytes , SPD-2 foci enlarged , indicating centrosome activation , and initiated separation . Cortical microtubule superstructures and small pairs of microtubule asters ranging from 0–90° of separation were also visible ( data not shown ) . In rachis-detached spermatocytes , microtubule asters continued to enlarge and separate until , by NEBD , they were fully opposed and the only remaining microtubule superstructure ( Figure 3B and 3C , Figure 6A ) . Thus , the G2 to M transition of spermatogenesis is associated with nucleation and separation of microtubule asters , exit from the karyosome stage , and rachis detachment . In contrast to oogenesis , spermatocyte meiotic spindle assembly is largely completed prior to NEBD . Having discovered that microtubule aster assembly and separation initiates in late karyosome spermatocytes , we also investigated the distribution of the cell cycle regulator polo-like kinase ( PLK-1 ) [55] , which has been implicated in cell division processes including mitotic spindle formation and mitotic entry [56] . Interestingly , repression of PLK by the PLK binding protein Matrimony maintains the G2 karyosome state in Drosophila oocytes [57] . In C . elegans diplotene and early karyosome spermatocytes , PLK-1 concentrates in a ring around the nuclear envelope and punctate structures throughout the cell ( Figure 6B and 6D ) . In late karyosome spermatocytes , PLK-1 localized to active centrosomes . After NEBD , PLK-1 continued to associate with centrosomes but also bound to metaphase chromosomes . This dynamic pattern is consistent with PLK-1 mediating microtubule nucleation [58] and centrosome separation [59] . It also suggests that PLK-1 relocalization marks the G2 to M transition in spermatocytes . Consistent with PLK-1 either activating or being activated by the universal regulator of the G2 to M cell cycle transition , Cdk1-cyclinB , cytoplasmic levels of cyclin B increase throughout the karyosome stage , and cyclin B switches to a predominantly nuclear distribution near the time of rachis detachment ( Figure 6C and 6D ) . Because spermatocytes and oocytes differ in chromatin composition and meiotic spindle structure and assembly , we anticipated that kinetochores , which link chromosomes to microtubules , might also vary in structure or assembly . During C . elegans mitosis , kinetochores are large , plaque-like structures , reflecting the holocentric nature of their chromosomes [60] , [61] . Studies of mitotic cells suggest a stepwise assembly of kinetochores [62] , [63] in which the evolutionarily conserved inner kinetochore components HCP-3CENP-A and HCP-4CENP-C establish a specialized chromatin base for the association of outer kinetochore proteins , which interface with spindle microtubules . However , meiotic-specific kinetochore structures may be required for orienting sister chromatids towards microtubules from the same spindle pole for the first meiotic division . Kinetochores of spermatocytes and oocytes may also differ since spermatocyte spindles are centriole-based while oocyte spindles are not . In fact , proteomic studies have identified gamete-specific differences in the levels of C . elegans kinetochore proteins [17] . Specifically , HCP-4CENP-C was enriched in spermatogenic chromatin while HCP-3CENP-A was enriched in oogenic chromatin . Similarly , the outer kinetochore protein HCP-1 was detected in chromatin preparations from oogenic germ cells but not from sperm . In this study , immunoanalysis of five different kinetochore proteins revealed striking sex-specific differences in the relative levels of specific kinetochore proteins . Although we found high levels of the inner kinetochore protein HCP-3CENP-A and the outer kinetochore protein HCP-1 in oocytes , these proteins were barely detectable in spermatocytes ( Figure 7A and 7E ) . Conversely , HCP-4CENP-C was highly abundant in spermatocytes ( Figure 7B ) . Because inner kinetochore components are intimately incorporated into chromatin , the near absence of detectable HCP-3CENP-A on spermatocyte chromosomes suggests that spermatocytes and oocytes differ in the organization of their kinetochore components . The CENP-F homologs , HCP-1 and HCP-2 are thought to function non-redundantly in mitotic spindle checkpoint assembly but redundantly in mitotic chromosome segregation [64]–[66] . The near absence of detectable HCP-1 suggests that HCP-2 may function non-redundantly in spermatogenesis ( Figure 7D and 7E ) . Notably , while all other kinetochore proteins were lost after anaphase II , HCP-4CENP-C perdured , encircling the post-meiotic sperm chromatin mass . Localization patterns revealed similarities and differences in the kinetochores of spermatocytes and oocytes . Interestingly , HCP-3CENP-A or HCP-4CENP-C localization differed strikingly in spermatocytes and oocytes ( Figure 7A and 7B ) . In oocytes , these proteins co-localize with the chromatin in an “inner” pattern , while in spermatocytes they “surround” the spermatocyte chromosomes in a pattern previously described for outer kinetochore proteins . Antibody inaccessibility is unlikely to account for this difference as co-immunostaining with anti-histone H1 antibody showed an even distribution of histone H1 on meiotically dividing chromosomes ( data not shown ) . HIM-10 , an evolutionarily conserved outer kinetochore protein surrounded the spermatocyte and oocyte chromosomes in a similar manner ( Figure 7C and [61] ) . The outer kinetochore protein HCP-2 also surrounded spermatocyte chromatin , yet in comparison to HIM-10 , its minimal co-localization with the chromatin suggests a less intimate association ( Figure 7C and 7D ) . HCP-2 localized completely between separating chromosomes in anaphase I then both asymmetrically and symmetrically on metaphase II chromosomes , indicating HCP-2 relocalizes from one side of the chromosomes to surround DNA prior to meiosis II ( data not shown ) . Thus in spermatocytes , inner kinetochore proteins exhibit a localization patterns more similar to outer kinetochores . This differential enrichment and localization of kinetochore components is the first evidence suggesting that the molecular machinery required for chromosome segregation in spermatocytes may differ from that in either oocytes or mitotically dividing cells . Sex-specific differences in the molecular composition of meiotic kinetochores may reflect differences either in the structure of the meiotic chromosomes or in the molecular requirements for interacting with structurally distinct meiotic spindle structures . The presence or absence of centrosomes not only affects the relative timing of meiotic spindle assembly but also influences the structure and mechanics of the spindles . In C . elegans , oocyte chromosomes apparently slide to metaphase congression between bundled microtubules as they segregate on a barrel-shaped , acentriolar spindle [67] , [68] . Anaphase movements are also distinctive with short-distance movements from the midline to the poles followed by further separation as the zone of midbody microtubules lengthens between the chromosome plates [26] , [68] . Although the meiotic cell divisions during C . elegans spermatogenesis have been described [30] , we used improved immunocytological methods to stage small and scarce dividing spermatocytes . To do this we characterized chromosome morphology in combination with DIC cell morphology [30] or microtubule dynamics [69] ( Figure 3 and Table 1 ) in flattened gonad preparations . Centriole based microtubule asters are first detectable in late karyosome spermatocytes and extend throughout the cell by the end of diakinesis as centrosomes separated to opposite sides of the nuclear envelope ( Figure 3B ) . Upon NEBD chromosomes attached to microtubules and congressed to form a metaphase I plate in which the X-chromosome is surrounded by autosomes ( Figure 3C ) . Metaphase I spermatocytes exhibited prominent astral microtubules . At anaphase I , spindle poles separated and disjoined homologs moved to opposite poles , with the X-chromosome frequently lagging between between ( Figure 2C and Figure 3D ) . Cytokinesis following anaphase I was often incomplete with a small cytoplasmic bridge connecting secondary spermatocytes [30] . Transition from anaphase I to metaphase II occurred without intervening chromosome decondensation and with metaphase II spindle poles setting up directly adjacent to the metaphase II plates ( Figure 3E ) . During anaphase II , microtubules concentrated at the spindle poles with chromosomes in between; midbody microtubules were not prominent ( Figure 3F ) . After a transient , shallow cleavage furrow formed and rapidly regressed , spermatids budded from a central residual body that accumulates materials not needed by mature sperm , including the bulk of the microtubules ( Figure 3G; [30] ) . During this final phase , the haploid chromosome complement quickly condensed further , forming a tiny , spherical , highly refractive chromatin mass . Importantly , these studies distinguish anaphase II from the asymmetric “budding division” that partitions spermatids from residual bodies as distinct and sequential events . Overall , the progression of meiotic division stages can now be reliably distinguished by a combination of chromatin , cell , and microtubule morphology ( Table 1 ) . In parallel studies , we studied the localization patterns of factors that facilitate meiotic divisions . These include the kinases AIR-2 and PLK-1 , as well as the AIR-2 targets pHisH3-ser10 and the meiotic cohesin protein REC-8 . AIR-2 phosphorylates REC-8 , which is present both between sister chromatids and between homologs , marking it for removal during the sequential metaphase-to-anaphase transitions [70] . Meiotic chromosome segregation thus requires AIR-2 to specifically localize between paired homologs during meiosis I and only relocalize between sister chromatids during meiosis II [2] , [48] , [71] . In fact , localization of AIR-2 , pHisH3-ser10 , and REC-8 to the mid-bivalent during spermatogenic meiotic divisions matches that described for meiotically dividing oocytes ( Figure 8A–8C ) [71] . PLK-1 , which regulates both microtubule and cell cycle events during meiosis , first associated with late karyosome centrosomes and persisted there through anaphase II ( Figure 6 and Figure 8D ) . PLK-1 also bound to chromatin during prometaphase and metaphase II but also localized between segregating chromosomes during anaphase I and II ( Figure 8D ) . The dynamic localization of PLK-1 may enable it to promote cohesion release during the metaphase-to-anaphase transition [72] and subsequently promote cytokinesis during anaphase [73] . Taken together , the localization of AIR-2 , pHisH3-ser10 , REC-8 , and PLK-1 in meiotically dividing spermatocytes can distinguish sub-stages of the meiotic divisions . Furthermore , while AIR-2 and PLK-1 exhibit sex-specific localization patterns before the meiotic divisions; their localization patterns during the meiotic divisions are remarkably non sex-specific .
In C . elegans , sperm and oocyte meiosis occur at remarkably different rates . Meiotic prophase lasts 54–60 hours during oogenesis and only 20–24 hours during spermatogenesis [74] . How is meiotic progression accelerated during sperm formation ? While previous studies suggest the lack of a DNA damage checkpoint during spermatogenesis shortens the pachytene period [74] , [75] , our studies reveal sperm-specific components and instances of overlapping developmental sub-programs that could potentially speed meiotic progression . For example , the sperm-specific presence of centrioles allows meiotic spindle assembly prior to NEBD [51] , [76] . These preformed spindles could accelerate the initiation of chromosome segregation in spermatocytes compared to oocytes , which must default to an alternate , chromosome directed mode of spindle assembly that can only begin after NEBD [77] , [78] . A second key event in spermatogenesis is the shaping and compaction of spermatid chromatin , which typically occurs during an extended period of post-meiotic differentiation [79] , [80] . C . elegans spermatids lack a prolonged post-meiotic differentiation phase [81] , yet achieve similar chromatin compaction . Shifting key events earlier may facilitate this process . For example , while mammals incorporate variant histones during meiosis and protamines after meiosis , C . elegans incorporates all currently known SNBPs at the end of pachytene while chromatin structure remains relatively more accessible [17] . Additionally , once sperm chromatin is condensed for meiotic divisions , it does not decondense for a post-meiotic round of transcription . Instead , final compaction of the haploid chromatin is reduced to a quick step following anaphase II , suggesting that required components may be pre-loaded and merely require an as yet unknown post-translational switch . Down-regulation of transcription is also shifted to an earlier stage in C . elegans sperm formation . During C . elegans oogenesis , nucleolar breakdown , HisH3-ser10 phosphorylation , and AIR-2 reassociation are coupled to the G2 to M transition [25] , [28] , [49] and delayed when cell cycle progression is halted by either the absence of sperm or depletion of cdk-1 [28] . In contrast , during spermatogenesis these events occur during the earlier diplotene to karyosome transition with AIR-2 reassociation following . Thus , events associated with transcriptional down-regulation appear to be uncoupled from the G2 to M transition in spermatocytes . Consistent with this model , spermatocytes still form compact , pHisH3-ser10 positive , chromatin masses even when the G2 to M events of spindle formation and NEBD are blocked by dominant “always on” mutations in the cell cycle regulator wee-1 . 3 [82] . Our studies predict that these mutant spermatocytes have completed the diplotene to karyosome transition but are subsequently arresting at the karyosome stage . We have found C . elegans spermatocytes form karyosomes , a feature of meiosis in more than 120 species including Drosophila , mouse and humans [10] , [41] , [83] , [84] . While karyosomes are proposed to prepare and gather chromosomes prior to meiotic divisions , our studies indicate that these functions can also be important for chromosomes that are holocentric and/or segregate on centriole-based spindles . Our studies also suggest that karyosome chromatin is both highly structured and dynamic; karyosome chromosomes exhibit organized stripes of the SC axial/lateral element protein HIM-3 and the aurora kinase , AIR-2 . Disruption of these patterns in the absence of proper SC formation suggests the chromosomal superstructure of karyosomes may “lock in” SC-related organizational information that would otherwise be lost after desynapsis . As in other organisms , C . elegans karyosomes form during or after diplonema [10] . We further found that karyosome formation coincides with nuclear envelope detachment , SC central element protein ( SYP-1 ) loss , HisH3-ser10 phosphorylation , and transcriptional down-regulation . Drosophila oocytes undergo a similar suite of events during karyosome formation and these events are collectively disrupted by mutations in the nucleosome histone kinase NHK-1 , also known as vaccinia related kinase VRK-1 [83] . Known substrates of NHK-1/VRK-1 include histone H3-ser10 [85] , as well as histone H2A-thr119 [83] and the chromatin-nuclear envelope linker BAF-1 [86] . Thus , NHK-1/VRK-1 is a prime candidate for linking karyosome formation to other cellular events in C . elegans spermatogenesis . Unfortunately , germline proliferation defects in nhk-1/vrk-1 mutants have thus far precluded us from testing this prediction ( data not shown ) . Other proteins , like the cell cycle regulator wee-1 . 3 , may control whether late prophase chromosomes aggregate into karyosomes or disperse , as in C . elegans oocytes at diakinesis [67] . In developing oocytes , RNAi depletion of wee-1 . 3 causes precocious maturation involving premature nucleolar breakdown and histone H3 phosphorylation [25] , [28] . Strikingly , these mutant oocytes exhibit chromosome aggregation reminiscent of karyosome formation , as well as ectopic microtubule aster formation prior to NEBD [28] . Thus wee-1 . 3 is an excellent candidate for an oocyte-specific regulator that delays the G2-to-M transition until oocytes are properly prepared for meiotic divisions and fertilization . Spermatocytes and oocytes also differ in rachis detachment timing . While rachis detachment accompanies the diplotene to diakinesis transition of oogenesis [35] , it accompanies the G2 to M transition of spermatogenesis ( this paper ) . Transcriptionally repressed , detached spermatocytes lack somatic support while rachis-detached oocytes at diakinesis endocytose yolk proteins from the pseudocoelom and maintain gap junction contact with surrounding somatic sheath cells until ovulation [87] . Thus , for spermatocytes , rachis detachment may represent a critical point of “cellular independence” . For oocytes , the analogous point is not rachis detachment but ovulation . Though late prophase spermatocytes and oocytes exhibit many differences , their metaphase I chromosomes have remarkably similar patterns of AIR-2 , pHisH3-ser10 , REC-8 , and PLK-1 localization . Thus an open question was whether kinetochores also differ in a gamete specific manner . Our finding that spermatocyte and oocyte kinetochores do differ in molecular composition and localization patterns suggests kinetochore structure may adapt to reflect sex-specific differences of meiotic spindles and underlying chromatin structure . Meiotic kinetochores also apparently differ from their mitotic counterparts . On mitotically dividing chromosomes , the inner kinetochore protein HCP-3CENP-A is required to recruit HCP-4CENP-C [62] , [63] . On meiotically dividing oocyte chromosomes , HCP-3CENP-A and HCP-4CENP-C are present at high levels , but their role is controversial . When RNAi was used to deplete HCP-3CENP-A and HCP-4CENP-C , live studies of chromosome segregation in GFP-histone tagged oocytes suggested that HCP-3CENP-A and HCP-4CENP-C were dispensible for oocyte meiosis [88] . However , analyses using fluorescence in situ hybridization ( FISH ) or restriction fragment length polymorphisms ( RFLPs ) to tag individual chromosomes revealed reproducible segregation defects ( A . Severson and B . Meyer , pers . comm . ) . Our own studies add to the puzzle , since HCP-3CENP-A was barely detectable on spermatocyte chromosomes while HCP-4CENP-C was highly enriched . This may reveal that very low levels of HCP-3CENP-A may be sufficient to recruit HCP-4CENP-C . Alternatively , histone replacement and SNBP incorporation during late pachynema [17] may alter chromatin structure and consequently influence subsequent chromatin-based events , like kinetochore assembly . For instance , incorporation of the histone variant H2A . X proved essential for heterochromatic chromatin formation of the XY body [89] . Likewise , protamine or protamine-like proteins package DNA in a non-nucleosomal configuration [16] , [90] . Therefore , SNBP-based chromatin packaging may itself provide a sufficient platform for HCP-4CENP-C recruitment or maintenance . Indeed , following meiotic divisions and the departure of all other kinetochore proteins , HCP-4CENP-C remains bound to sperm chromosomes . This study describing the dynamics of key markers throughout spermatogenesis establishes guidelines for staging C . elegans spermatocytes and characterizing spermatogenesis defects . Importantly , our findings are also relevant to the understanding of meiosis . The discovery of gamete-specific differences in SC disassembly timing raises new questions regarding how SC disassembly is linked to the G2 to M transition . Likewise , differences in kinetochore structure raise questions about how kinetochore assembly is modified for distinct chromosome segregation events . This study also establishes a framework for comparative studies . What can be gleaned about the enigmatic karyosome stage from comparative studies between C . elegans spermatocytes and the oocytes of Drosophila and Xenopus ? Which features of the rapid spermatogenesis program of C . elegans are shared with Drosophila and mammals ? Until now , studies of C . elegans spermatogenesis have focused on features of their non-flagellated spermatozoa; this study highlights the usefulness of C . elegans spermatogenesis as a model for understanding the fundamental biology of meiosis .
C . elegans strains were maintained as described by Brenner [91] . All nematode strains were cultured at 20°C except where noted . Strains used include Bristol N2 , CB1489 him-8 ( e1489 ) , DR466 him-5 ( e1490 ) , AV307 syp-1 ( me17 ) V/nT1[unc- ? ( n754 ) let- ? qIs50] ( IV;V ) . Males were obtained either by mating 3 N2 hermaphrodites with 7 N2 males at 19°C for 4 days or by culturing 4–6 him-8 ( e1489 ) or him-5 ( e1490 ) hermaphrodites on OP50 seeded NGM plates for 3–5 days . Animals were then collected and bleached to isolate embryos ( 15 parts double distilled water : 4 parts bleach : 1 part 10N sodium hydroxide ) . Embryos were hatched without food overnight at 19°C with shaking at 200 rpm . L1 larvae were then plated onto OP50 seeded NGM plates at 19°C for 2–3 days . Animals from these synchronous cultures were used for immunostaining . Alternatively , fourth larval stage ( L4 ) males were collected from mating plates and grown to adulthood for 24–48 hrs prior to analysis . Male gonads were dissected in 5–10 microliters of sperm salts on ColorFrost Plus slides ( Fisher Scientific ) using established protocols for antibody staining of C . elegans gonads provided in Wormbook [92] . Three different fixation methods were used in this study . For paraformaldehyde staining , animals were processed as described in [61] , [92] . For cold methanol or methanol/acetone fixation , animals were dissected in sperm salts , and then a coverslip with four corner dots of silicon grease was placed over the isolated gonad and gentle pressure was applied to generate partially flattened gonads and/or monolayers of spermatocytes and spermatids . The slide preparation was then placed either in liquid nitrogen or on dry ice . After freezing , the coverslip was removed . For methanol/acetone preparations , the slide was immersed in 95% methanol for 10 minutes followed by a 5 minute immersion in 100% acetone . Slides were allowed to air dry briefly . For −20°C methanol preparations , slides were kept in methanol overnight . Slides were washed with three consecutive 10 minute washes in PBS followed by a 30 min . room temperature incubation in blocking solution ( PBS+0 . 5% BSA and 0 . 1% Tween 20 ) . Primary and secondary antibody incubations were each diluted into blocking solution at conducted at room temperature in a humid chamber . For DIC/Hoechst preparations , Hoechst 33342 ( Sigma-Aldrich ) was used at 100 µg/ml . The following primary antibodies were used in overnight incubations ( unless otherwise noted ) with different fixation conditions . Commercial sources or labs kindly providing antibodies are also listed . Paraformaldehyde fixed preparations: 1∶200 mouse anti-REC-8 ( Abcam ) ; mouse anti-Nop-1 ( yeast fibrillarin mAbD77 , Aris lab ) used at a 1∶1000 dilution [93] , rabbit anti-HIM-3 ( Zetka Lab ) used at a 1∶400 dilution [44] , rabbit anti-CeLamin ( Gruenbaum lab ) used at a 1∶500 dilution , rabbit anti-AIR-2 ( Schumacher lab ) used at a 1∶500 dilution , rabbit anti-SPD-2 ( O'Connell Lab ) used at a 1∶500 dilution , guinea pig anti-SYP-1 ( Villeneuve Lab ) was preasborbed against homozygote syp-1 ( me17 ) mutant animals from the strain AV307 and used at a 1∶200 dilution [45] . Methanol-acetone fixed preparations: rabbit anti-HCP-1 used at a 1∶200 dilution , rabbit anti-HCP-3 used at a 1∶200 dilution , and rabbit anti-HCP-4 used at a 1∶200 dilution ( Moore lab ) , rabbit anti-HIM-10 ( Meyer lab ) used at a 1∶500 dilution [61] . The HCP-2 antibody , used at a 1∶200 dilution , is a rabbit polyclonal raised against the peptide NSVDDNSYCEPPRASSAHD that correspond to amino acids 93–110 of HCP-2 . Cold methanol preparations as described in [94]: 1∶400 rabbit anti-pHisH3-ser10 ( Upstate Biotechnology ) , 1∶100 FITC-conjugated anti-α-tubulin ( DM1A; Sigma-Aldrich ) , 1∶1000 rabbit anti-PLK-1 [55] ( Golden Lab ) . 1∶3 anti-cyclin B ( F2F4 monoclonal developed by P . O'Farrell , Developmental Studies Hybridoma Bank ) . All incubations were 2–3 hrs at room temperature except PLK-1 and cyclin B , which were incubated overnight at 4°C and room temperature , respectively . Secondary antibodies from Invitrogen include goat anti-rabbit AlexaFluor 488-labeled IgG ( used at 1∶100 ) , goat anti-rat AlexaFluor 488-labeled IgG ( used at 1∶100 ) and goat anti-mouse AlexaFluor 488-labeled IgG ( used at 1∶100 ) . Affinity purified secondary antibodies from Jackson Immunoresearch Laboratories include goat anti-rabbit TRITC-labeled IgG ( used at 1∶100 ) and goat anti-mouse FITC-labeled IgG ( used at 1∶100 ) . DNA was visualized using the DNA dye DAPI at 0 . 1 µg/ml . Slides were prepared with either VectaShield ( Vector Labs ) or GelMount ( Biomedia Corp . ) as a combined mounting and anti-fade media . Images were acquired via either confocal microscopy using a Leica TCSNT microscope , epifluorescence microscopy using a Zeiss Axiovert200M coupled with deconvolution via Slidebook 4 . 2 software ( Intelligent Imaging Innovations ) , or DIC and epifluorescence on an Olympus BX60 microscope equipped with a Cooke Sensicam . Images acquired by confocal microscopy include those to visualize fibrillarin , HIM-10 , SPE-11 , HCP-4 , and HCP-3 . For deconvolution , images were acquired at 2×2 binning and 0 . 2 µm step sizes through each gonad and processed using either constrained iterative or nearest neighbors deconvolution . Images obtained via deconvolution include lamin , SYP-1 , HIM-3 , AIR-2 , RNA pol II CTD ( ser 2 ) , SPD-2 , and HCP-2 . Epifluorescence images include pHisH3-ser10 , α-tubulin and PLK-1 .
|
Sperm and oocytes contribute equal but unique complements of DNA to each new life . Both types of cells arise from meiosis , a multi-step program during which chromosomes replicate , pair and recombine , then divide to generate haploid gametes . Simultaneously , each cell type also differentiates via distinct developmental programs . Spermatogenesis rapidly produces many small , motile sperm with highly protected chromatin , while oogenesis occurs at a slower rate to yield fewer large , immobile , nutrient-rich oocytes . We provide a detailed molecular analysis of key landmark events of spermatogenesis and identify spermatogenesis-specific features of meiosis in the model organism C . elegans . We find that , as in many meiotic programs , C . elegans spermatogenesis includes a chromosome aggregation or “karyosome” phase . This extended stage provides a period for chromosome and microtubule remodeling prior to the meiotic divisions . Our analysis identifies several gamete-specific features of the meiotic program that may contribute to the differential timing , pace , and mechanics of meiotic progression . Our findings provide a foundation for understanding how differentiation influences meiosis , which is an essential step in identifying universal features required for reproductive success in all organisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology/germ",
"cells",
"cell",
"biology/nuclear",
"structure",
"and",
"function",
"cell",
"biology/cell",
"signaling",
"cell",
"biology/cell",
"growth",
"and",
"division",
"genetics",
"and",
"genomics/nuclear",
"structure",
"and",
"function",
"cell",
"biology/developmental",
"molecular",
"mechanisms",
"genetics",
"and",
"genomics/chromosome",
"biology",
"developmental",
"biology/molecular",
"development",
"developmental",
"biology/developmental",
"molecular",
"mechanisms"
] |
2009
|
Spermatogenesis-Specific Features of the Meiotic Program in Caenorhabditis elegans
|
In this study we compared the utility of two molecular biology techniques , cloning of the mitochondrial 12S ribosomal RNA gene and hydrolysis probe-based qPCR , to identify blood meal sources of sylvatic Chagas disease insect vectors collected with live-bait mouse traps ( also known as Noireau traps ) . Fourteen T . guasayana were collected from six georeferenced trap locations in the Andean highlands of the department of Chuquisaca , Bolivia . We detected four blood meals sources with the cloning assay: seven samples were positive for human ( Homo sapiens ) , five for chicken ( Gallus gallus ) and unicolored blackbird ( Agelasticus cyanopus ) , and one for opossum ( Monodelphis domestica ) . Using the qPCR assay we detected chicken ( 13 vectors ) , and human ( 14 vectors ) blood meals as well as an additional blood meal source , Canis sp . ( 4 vectors ) . We show that cloning of 12S PCR products , which avoids bias associated with developing primers based on a priori knowledge , detected blood meal sources not previously considered and that species-specific qPCR is more sensitive . All samples identified as positive for a specific blood meal source by the cloning assay were also positive by qPCR . However , not all samples positive by qPCR were positive by cloning . We show the power of combining the cloning assay with the highly sensitive hydrolysis probe-based qPCR assay provides a more complete picture of blood meal sources for insect disease vectors .
Blood-feeding insects in the subfamily Triatominae are vectors of Trypanosoma cruzi , the parasite that causes Chagas disease . Approximately 140 species of Triatominae range across the Americas [1] and vary in their role in transmitting human Chagas disease in part because of their preference for sylvatic ( wild ) , peridomestic ( immediate surroundings of a house ) or domestic ( within house ) ecotopes . Although it is generally assumed that domestic and peridomestic vectors are important in disease transmission , the role of sylvatic vectors in disease transmission is less understood . Because sylvatic vectors have the potential to colonize houses or simply enter houses to feed and then leave , collecting sylvatic vectors and analyzing their blood meal profiles can provide insight into their movement among domestic , peridomestic and sylvatic ecotopes and their potential role in disease transmission [2] . Not only vector species , but also vertebrate hosts vary in relevance for human transmission . Vectors that feed on human blood are important in disease transmission but because some mammals are more likely to transmit the parasite to the vectors , insight into the spectrum of blood meal sources is also epidemiologically important . For example , non-infected Triatoma infestans , often cited as the most important vector of Chagas disease [3] , were 50 times more likely to become infected when feeding on dogs compared to humans [4] . Dogs are also a more dependable food source relative to other hosts , evident by vectors feeding more consistently on dogs across study sites and seasons [5] . A strong correlation between vector parasite infection and vectors feeding on dogs has been reported [6] . Several techniques have documented blood meal profiles of these triatomine insect vectors such as protein-based assays ( e . g . , antisera and precipitin tests [6]–[8] ) , DNA tests based on the polymerase chain reaction ( PCR , e . g . , gel electrophoresis [9] , [10] , melt curve analysis [11] , [12] ) , direct sequencing [13] , [14] and cloning followed by sequencing of PCR products [15] . DNA-based approaches have the advantage of being more amenable to the degraded DNA often found in the vector digestive tract . Except for techniques involving DNA sequencing , these approaches require antibodies , PCR primers , or restriction enzymes designed to detect specific taxa , and the scope is limited by a priori knowledge of potential blood meal sources [16] . For domestic and peridomestic vectors , common blood meal sources ( e . g . , humans , domestic animals ) are known [10]–[13] , [15]–[18] . However , it is challenging to identify potential sources of blood meals for sylvatic vectors . Amplifying DNA from the blood meal using vertebrate or mammalian specific primers followed by cloning of the PCR products has the advantage of being able to cast a wide net in detecting vector blood meal sources [15]; however , cloning is more costly and time consuming than other DNA-based approaches . Cloning , reliant on conventional PCR may be biased toward more recent blood meals because of vector digestion of older blood meals [16] , [19] , and thus less sensitive to all blood meal sources , relative to modern highly sensitive DNA-based approaches such as qPCR [16] , [20] . The relative sensitivity and specificity of cloning vs . qPCR in detecting blood meal sources collected from the harsh environment of the vector digestive tract has not been examined . We explore the application of these two molecular techniques to examine the role of sylvatic vectors in the transmission of human Chagas disease , by determining the blood meal sources of sylvatic vectors collected from a region in Bolivia with high disease incidence [21] . We compared two approaches designed to detect blood meal sources from DNA extracted from the vector abdomen . The first approach , cloning of PCR products amplified with general vertebrate primers , broadly identifies all potential vertebrate blood meal sources [15] . The second approach uses highly sensitive qPCR taxa-specific primers and hydrolysis probes to survey for chicken , Canis sp . ( e . g . , dog , wolf and coyote ) and human blood meals . To our knowledge this is the first study to use hydrolysis probe-based qPCR to analyze blood meal sources of insect disease vectors and compare the two DNA-based methods for sensitivity and specificity .
The protocol for handling animal specimens in this study was approved by Universidad de San Francisco Xavier de Chuquisaca Animal Research Committee and follows the European Directive ( Directive 2010/63/EU revising Directive 86/609/EEC on the protection of animals used for scientific purposes , project license number D . F . C . Q . F Y B . N° 237 ) . Located in the Andean highlands of the department of Chuquisaca , Bolivia , the rural landscape of the study area includes xeric valleys [22] , with thorny shrubs and cacti adapted to low , seasonal rainfall that defines the dry ( 8 . 4 cm average rainfall of May to October ) and wet seasons ( 35 . 8 cm average rainfall of December to March ) [23] . The sampling locations ( Mean Center calculated at 65o 08′ 2 . 31″ W , 18o 46′ 44 . 26″ with ArcGIS , Ver . 10 . 1 , ESRI Inc . , Redlands , California , USA ) were within 200 m of Rio Chico; six sampling locations were evenly divided between the west and east banks . Elevation of these locations range between 1740 m and 1780 m above mean sea level ( AMSL , Figure 1 ) . The nearest community to the sampling locations is Zurima , ∼600 m to the east ( Figure 1 ) . Located along Rio Chico , Zurima has an elevation of 1730 m ( Mean Center at 18o 46′ 34 . 96″ S , 65o 07′ 42 . 36″ W ) . As of the latest national census in 2001 , Zurima had a population of 495 ( 257 females , 238 males ) living in 135 houses [24] . The residents are mostly indigenous or of Spanish descent and practice subsistence agriculture [24] , [25] . Six isolated houses , ranging 14 m to 870 m apart and ∼60 m from the closest trap location , are located on the west riverbank ( Figure 1 ) . These isolated houses were identified with satellite imagery but never surveyed . The town boundary of Zurima was defined as a 50 m buffer from houses along the edge of the community . Six isolated houses were geolocated based on freely available satellite imagery ( ArcGIS , Ver . 10 . 1 , ESRI Inc . , Redlands , California , USA ) [26] ( Figure 1 ) . A proximity function ( Near ) analysis was used to calculate the distance between the traps and town boundaries and thus , the distance of sylvan vectors from the town as well as the distance between the traps and isolated houses . ArcGIS was used to perform all GIS analyses ( ArcGIS , Ver . 10 . 1 , ESRI Inc . , Redlands , California , USA ) . Trapping many insect vectors is challenging due to the low success rates of traps [27] . We baited traps with live mice ( also known as Noireau traps ) because a review of the literature suggested these traps attract T . infestans , the principal insect vector in the region , more successfully than traps without mice [28] in laboratory studies , and that they have been successful in previous field studies [27] , [29] . Traps consisted of opaque bottles ( 15×7 cm ) covered with double-sided tape . A mouse was placed inside with a small piece of apple; and the opening was sealed with a metal screening mesh to prevent adult vectors and large nymphs from entering the trap ( Figure 2 ) . One to four traps were placed at six georeferenced sampling locations for a total of 17 traps ( Figure 1 ) . Traps west of the riverbed were placed in sylvan areas , while the eastern traps were between the village of Zurima and agricultural areas . Traps were placed in the field at sunset on September 4th , 2012 at ∼18:00 and recovered ∼12 hours later . Vectors were gently removed from the double-sided tape with forceps , placed individually into plastic flasks , labeled according to the collection site , and transported live to the laboratory . Each trap location was recorded with a handheld GPS receiver ( Garmin model 76 , WGS 1984 ) . Trypanosoma cruzi infection was determined using microscopy by USFX ( Universidad San Francisco Xavier , Sucre , Bolivia ) researchers trained in the safe handling of infectious agents using published methods [21] . DNA was subsequently extracted from the posterior abdomen of each insect using the DNeasy Kit ( Qiagen , Valencia , CA ) following the manufacturer's protocol , as previously described [10] . DNA concentrations of abdomen extractions were measured with the NanoDrop ND-100 spectrophotometer . These abdomen extractions , a complex mixture of degraded DNA containing parasite DNA ( if infected ) , digested vector blood meal and vector tissue derived DNA , were also tested for T . cruzi by USFX researchers using PCR with the previously reported TCZ primers [21] , [30] . PCR results were confirmed by SEDES ( Servicio Departamental de Salud ) in La Paz , Bolivia . Blood meal sources were assayed using two methods: cloning of PCR products amplified with general vertebrate primers and qPCR using taxa specific primers and hydrolysis probes for chicken , Canis sp . and human blood meals ( Table 1 ) . We analyzed the blood meal sources of all 14 vectors using the cloning assay following previously published methods [15] , [31] . Briefly , because the DNA extractions from the insect abdomens potentially contain blood meal DNA from multiple vertebrates , the initial PCR used primers specific for mitochondrial DNA coding for the 12S ribosomal RNA gene of vertebrates ( hereafter referred to as 12S primers ) . Two sets of vertebrate 12S primers were used [22] , [32] , and are referred to hereafter as the Kitano and Melton assays . An ethidium bromide stained , 2% agarose gel was used to verify the ∼150 bp PCR products , which were then cloned with the pGEM-T kit ( Promega , Madison , WI , USA ) . Cloned DNA from 12 colonies per insect ( or 24 from two vectors because the first 12 did not have a single interpretable sequence ) was PCR amplified using the same 12S primers , sequenced using BigDye v3 . 1 ( Applied Biosystems , Foster City , CA , USA ) and subsequently analyzed with an ABI PRISM 3730xl DNA analyzer ( Beckman Coulter , Fullerton , CA , USA ) . Sequence alignments and editing were done with Sequencher v4 . 10 ( Gene Codes Corporation , Ann Arbor , MI , USA ) . Taxonomic identification of the sequences was determined as ≥99% match of 101 bp ( Kitano ) or 107 bp ( Melton ) using a BLAST search . To rule out contamination , two controls with nuclease free , DNA grade water instead of template DNA went through each step of the cloning and sequencing procedure for each set of primers ( i . e . , Melton and Kitano ) . Based on the results of the cloning assay , combined with previous studies analyzing blood meal sources of vectors from Bolivia , Brazil and Argentina [5] , [6] , [10] , [33] , [34] , species-specific qPCR assays were developed to test for chicken , Canis sp . and human blood meal sources . The blood meal qPCR assays were modified from previously described assays targeting the mitochondrial cytochrome b gene ( hereafter referred to as Cytb , [35] . The guidelines for the “Minimum Information for Publication of Quantitative qPCR Experiments” ( MIQE ) were followed to test the assays and interpret results [20] . We used the previously described [35] primer and probe concentrations for Canis sp . However , because the reported conditions for the chicken and human assays did not produce consistent results , we tested primer concentrations of 50–900 nM per reaction for chicken and human using DNA from known sources . After selecting the concentration yielding consistent amplification curves , we varied the probe concentration 50–900 nM . In addition , we varied annealing temperatures from 60°C to 62°C for the chicken and human assays . Following optimization , we analyzed the sensitivity of chicken , Canis sp . and human qPCR assays by varying the template DNA concentration using 10-fold serial dilutions spanning 10 orders of magnitude ( 100 vs . 10−10 ) in triplicate . There were eight no template controls ( NTC ) consisting of DNA grade H2O instead of template . The 100 template concentrations were 12 . 46 ng/uL of chicken DNA , 7 . 06 ng/uL of Canis sp . DNA and 7 . 90 ng/uL of human DNA . After optimization and sensitivity analyses , chicken and human assays were run in duplicate including all samples , positive controls spanning five orders of magnitude ( 100 to 10−4 ) and three or four NTC . We tested for Canis sp . in singlicate using half the amount of template used for the other qPCR assays because of limited DNA template . Because this is the first study to use hydrolysis probe-based qPCR to detect blood meal sources , the interpretation of the qPCR results of one replicate of each sample and at least five positive and three or four NTC controls , were verified by sequencing the qPCR products in one direction using the forward qPCR primer and the same sequencing protocol as the cloning assay . In three cases , because the duplicate trials differed , we sequenced both replicates . Although there was no evidence of NTC amplification before cycle 40 , we sequenced NTC with the samples to rule out false positive results . The qPCR reactions included 10 uL PerfeCTa qPCR ToughMix ( Quanta Biosciences , Gaithersburg , MD , USA , Catalog # 95112-250 ) , 8 uL of template DNA for human and chicken and 4 uL template DNA for Canis sp . , forward primer ( 200 nM chicken and Canis sp . , 300 nM human ) , reverse primer ( 60 nM chicken , 200 nM Canis sp . , 300 nM human ) , probe ( 230 nM chicken , 200 nM Canis sp . , 300 nM human ) and nuclease free water to make 20 uL . A two-step qPCR cycling protocol was used for all three assays . The chicken and human assays had an initial denaturation at 95°C for 10 minutes , followed by 45 cycles of 95°C for 20 seconds and 60°C for 1 minute , and a final extension of 72°C for 10 minutes . For the Canis sp . assay , the initial denaturation was for 5 minutes , followed by 45 cycles of 95°C for 5 seconds and 57°C for 40 seconds with final extension at 72°C for 10 minutes . Both probes and primers were purchased from Biosearch Technologies ( Novato , CA ) ; and qPCR reactions were run on a LightCycler™ 480 thermocycler ( Roche , Indianapolis , USA ) . Reagents and master mixes were prepared in a vented hood sterilized by UV light and wiped with 0 . 4% sodium hypochlorite to minimize contamination . Autoclaved tubes and DNA grade water , along with the DNA-free 96-well plates , filtered tips , pipettes and tube racks wiped with 0 . 4% sodium hypochlorite were placed 5–10 cm from a UV light source for 1 h before preparing all reagents and master mixes [36] . To compare sylvatic vector blood meal feeding samples to the positive and negative controls , the crossing point ( Cp ) , defined as the cycle where the sample begins to amplify above the background noise [20] , [37] , [38] , was calculated using the Absolute Quantification Fit Points method from the LightCycler 480 GeneScanner Software V1 . 5 ( Roche , Indianapolis , USA ) . Reactions were run for 45 cycles . The 40 cycle threshold represents the cycle where significant amplification should be visible if there is at least one molecule of target DNA in the original sample [37] .
Chagas disease vectors were collected at three of the six sampling locations , located furthest from Zurima , and from five of the 17 ( 30% ) traps . All 14 vectors were collected within 100 m of an isolated house ( Table S1 ) ; and all were T . guasayana , a species that has been previously reported in domestic [2] , peridomestic and sylvatic [39] ecotopes . The concentration of extracted DNA averaged 33 . 3 ng/uL ( range 2 . 4–175 . 22 ng/uL ) . T . cruzi was detected in only one vector ( 7% ) with complete agreement between microscopy and PCR . The initial PCR products from the 14 T . guasayana for both the Melton and Kitano assays were the expected size and thus used in the cloning reactions; even though there was no visible band from the negative control , it was treated the same as the samples to control for potential contamination . The PCR products from the cloning colonies of all 14 samples were also the expected size; the colonies from the negative control did not produce the expected band size indicating there was no contamination . For the Melton assay , 12 colonies from each sample were sequenced . Because no interpretable sequences were obtained from the first 12 , an additional 12 colonies were sequenced for samples Tg04 and Tg11 . For the Kitano assay , 12 colonies from each sample were sequenced . Of the 14 T . guasayana analyzed , four different blood meal sources were detected with the Melton cloning assay; three of these sources were found in more than one T . guasayana ( Table 2 ) . The most common blood meal source was human , which was found in six T . guasayana . Five T . guasayana had evidence of feeding on chicken ( Gallus gallus ) and five had fed on unicolored blackbird ( Agelasticus cyanopus ) . Opossum ( Monodelphis domestica ) was detected in one T . guasayana . The Melton cloning assay detected 57% ( 8 of 14 ) of T . guasayana feeding on birds from both domestic and sylvatic ecotopes , and 14 . 3% ( 2 of the 14 ) of the vectors feeding on birds were positive for both chicken and unicolored blackbird . Based on the Melton cloning assay , 50% ( 6 of 14 ) fed exclusively on domesticated taxa; 14 . 3% ( 2 of 14 ) fed exclusively on sylvatic taxa ( i . e . , unicolored blackbird ) ; 28 . 6% ( 4 of 14 ) fed on vertebrates from both domestic and sylvatic ecotopes . We were unable to detect a blood meal source for 7% ( 1 of 14 ) vectors despite analyzing 24 colonies instead of 12 . For the Kitano cloning assay , only one blood meal source , human , was detected , and from only one vector ( Table 2 ) . Overall , the Kitano primers were less successful than the Melton primers , the Kitano assay blood meals were detected from 7% vs . 86% of samples for the Melton primers . Although both authors claim the primers are for species identification among vertebrates , a comparison of sequences in this region of the 12S gene shows that the match varies among taxa and for the Kitano primers , appears to be higher in mammals than birds ( Table 3 ) . Interestingly , the Melton primer set did not detect the same blood meal source as the one blood meal the Kitano primer set detected . Although dogs are a common blood meal source in this region , they were not detected as a blood meal source by either of the cloning assays . The cloning NTC in duplicate , for both the Melton and Kitano assays , did not produce interpretable DNA sequence , ruling out contamination in the cloning assays . The limit of detection ( LOD ) , or template concentration that can be detected within 95% certainty [20] , is 10−4 ng/uL for all three qPCR assays ( Figure 3 ) . Although positive controls were detected at more dilute concentrations , we could only yield consistent and reproducible results until 10−4 ng/uL . Averages of three replicates per concentration are 3 . 9 ( range 3 . 7–4 . 6 ) cycles apart for chicken ( Figure 4 ) , 3 . 8 ( range 3 . 1–4 . 3 ) cycles apart for Canis sp . ( Figure 5 ) and 3 . 3 cycles ( range 2 . 2–3 . 8 ) apart for human ( Figure 6 ) , respectively , a little higher than the expected 3 . 3 cycles [37] . The more dilute samples ( <10−5 ng/uL ) sometimes amplified after the 40 cycles at which one DNA molecule should be detectable ( Figure 7 ) [37] . Four to five positive controls for chicken , Canis sp . and human qPCR products were sequenced , including the 10−4 ng/uL sample , and correctly identified in a BLAST query . For the qPCR assay , we found the prevalence of feeding on Canis sp . , chicken and human was 29% , 93% , and 50 or 100% ( 4 , 13 and 7 or 14 of 14 ) respectively ( results summarized in Table 2 , Cp values for each sample in each assay are shown in Table S2 ) . DNA sequencing confirmed the amplicon was the target DNA in all instances . Attempts to sequence the qPCR product from the NTC were unsuccessful for the chicken and Canis sp . assays . In contrast , two of the four human NTC had amplicons identified as human by BLAST search; however , the NTC amplified five cycles later than the lowest positive control ( 10−4 ng/uL , LOD ) and six cycles later than the latest amplifying sample ( Table S2 ) . Because of the human DNA amplification from the NTC , we conservatively conclude that at least 50% ( the number positive by the cloning assay ) of the samples are positive for human . Note that the samples positive for human by cloning tend to have the lowest Cp values , indicating higher concentrations of DNA ( samples positive for human by cloning average: Cp = 27 . 6 , average for samples negative for human by cloning: Cp = 34 . 0 ) . The qPCR assay detected its targeted taxa ( i . e . , chicken , Canis sp . and human ) in more blood meals than the cloning assay . For chicken , all five samples identified as positive by the cloning assay were also positive by qPCR; however , qPCR showed that an additional sample had fed on chicken . Similarly for human , all seven samples that were positive by the cloning assays were positive by qPCR and qPCR indicated that the remaining seven were positive for human; however , due to reported NTC amplification ( see QPCR blood meal detection ) , only the human cloning results can be unambiguously established . Although the cloning did not detect any feeding on Canis sp . , the qPCR assayed was positive for four samples .
We collected 14 sylvatic T . guasayana during the dry season ( September ) in the Andean highlands of the department of Chuquisaca , Bolivia and found a single individual positive for T . cruzi by both microscopy and PCR . Using the cloning assay and our newly developed , highly sensitive qPCR assay , we also found at least 50% were positive for human blood meal indicating T . guasayana should be further examined for potential epidemiological importance because our study did not explicitly examine parasite transmission . Our study reports on a limited number of T . guasayana because cloning is very resource intensive ( see Cloning blood meal detection ) ; however , additional specimens should be examined for a more complete picture of sylvatic T . guasayana feeding habits . T . guasayana from northern Argentina have been reported previously to have a high rate of feeding on humans ( 3 of 5 blood meals analyzed against blood antisera ) and carry a T . cruzi-like parasite ( 2 of 5 analyzed with microscopy ) , although none of the vectors were positive for both human and the parasite [6] , [40] . We detected , on average , 1 . 4 blood meal sources per vector using the cloning assays . Evidence of domestic ( chicken , Canis sp . and human ) blood meal sources was found in all vectors with qPCR . Literature reports adult T . guasayana as having high night flight dispersal during the dry season [27] , [40] . Previous studies of sylvatic Chagas disease vectors in Argentina found the dispersion index for T . guayasana was 4 . 5 times higher for females and 2 times higher for males compared to T . infestans the most prevalent domestic vector in this region [41] . They found T . guasayana had the highest number of flying individuals and although it does not colonize houses , adults often invade houses . Based on our trap locations , these vectors appear highly mobile , travelling between ∼60 m ( nearest isolated house ) and ∼600 m ( nearest house in Zurima ) to obtain blood meals from domestic sources . The collection sites for T . guasayana from this study are 60 m . ( Table S1 ) to the nearest isolated house , which is within the 80 m nymph dispersal distance reported in a study from Argentina . Thus previous studies indicate it is plausible for T . guasayana to move to houses , however the food sources also move . Evidence ( e . g . trash , fire pits , hunting remains ) of humans sleeping/resting in sylvan areas has been reported [42] , [43] and chickens often roam freely [44] , so this result could be from either vectors or hosts moving , or more likely , a combination of both . All 14 T . guasayana from this study had evidence of feeding on domestic animals , but about half also fed on sylvatic animals ( unicolored blackbird and opossum ) . Sylvatic T . guasayana from Argentina in previous studies have been reported to exclusively feed on domestic hosts based on an antisera test relying on controls gathered from local animals [45] . The majority of controls used in this Argentinian study were animals not typically found in sylvan areas ( 72 . 7% , 8 of 11 ) [6] , [45] . Our cloning method reported here does not rely on the availability of , nor is it limited by a priori knowledge of domestic , peridomestic and sylvatic fauna for testing blood meals . Although qPCR indicates T . guasayana in our 14 samples are not feeding on Canis sp . as frequently as humans and chickens , dogs have been suggested as sentinels of T . cruzi infection in humans [46] . A rural town in northern Argentina reported as many as 65% of the dogs seropositive for T . cruzi infection [46]; and T . guasayana was correlated with the T . cruzi infection of the dog population [46] . The incidence of T . cruzi in sylvatic hosts can give some insight into the importance of this broad feeding on the potential transmission of T . cruzi to humans . Sylvatic rodents have been reported as blood meal sources for sylvatic vectors in Central [18] and South America [17] and are an important T . cruzi reservoir with reportedly high parasitemia [47]; likewise , opossum are an important reservoir for T . cruzi throughout its range [48] . Although birds cannot sustain T . cruzi infection , having chickens near domiciliary areas may decrease vector parasitemia , while increasing vector abundance [7] . Fluorescence based qPCR assays are capable of detecting minute amounts of nucleic acids while being quick , simple and specific [20] . Of the two basic qPCR approaches ( i . e . , high resolution melting ( HRM ) and hydrolysis probe assays ) , HRM assays have recently been reported for detecting multiple blood meals; however , the resolution of taxa with similar Tm [11] ( for example chicken and human 86 . 27 and 85 . 79 , respectively ) could yield indistinguishable genotype signatures . Similarly , a recent study used HRM to detect blood meals using 12S vertebrate primers; however , multiple feedings by a vector was not examined [12] . This is the first report of a hydrolysis probe-based qPCR assay for detecting vector blood meal sources . Probe-based assays can be multiplexed and have the potential to be more specific , more sensitive , and more reproducible . Although the assays reported here examined each taxa in a separate assay , multiplex assays could be developed using probes with different fluorescence for each potential vertebrate host . There are few controlled experiments for blood meal detection . Although insect vectors require a feeding to molt into each lifecycle stage and many Triatominae can live up to two years [49] , [50] , researchers have detected blood meals in lower abdomen extractions only 70 days post feeding using less sensitive PCR methods [18] . Numerous studies have used days post feeding as an indicator of assay sensitivity [9] , [11] , [18] , [19] . An alternative approach is to assess PCR sensitivity with serial dilutions , while examining any temporal effects of blood meal digestion using days post feeding . One study successfully detected template until 21 days post feeding with sensitivity reported until 10−1 , 10−3 and 10−2 for chicken , dog and human , respectively [51] . Our study is the first to examine qPCR sensitivity using serially diluted controls . We found a lower detection limit and thus present more sensitive assays than reported for conventional end-point PCR ( chicken: 3 orders of magnitude , Canis sp . : 1 order of magnitude and human: 2 orders of magnitude ) [51] . Temporal effects of blood meal digestion has not yet been examined for hydrolysis probe-based qPCR . Despite extreme efforts to control for contamination , the qPCR assay showed amplification of human DNA in the NTC , suggesting caution in interpreting results . We ran samples in duplicate or triplicate with high volumes of template DNA to reduce the likelihood of false positives . Sequencing of qPCR products , used to verify results , showed we were unable to eliminate human contamination; however , the combination of the cloning assay results and the qPCR amplification of samples at least 5 cycles before the NTC supports our conclusion that a significant number of these vectors are feeding on humans . Previous studies have shown some reagents may be contaminated with human DNA and have encouraged researchers to position tubes on their sides with UV treatment within 2 . 54 cm [52] , [53] , differing from our upright placement of tubes with DNA grade water within 10 cm of a UV light source . Therefore we speculate our DNA grade water , used to prepare other reagents , as a potential source of contamination . This study demonstrates the power of combining the cloning assay with the highly sensitive hydrolysis probe-based qPCR to provide a more complete picture of the blood meal sources of insect disease vectors . Our data show that humans and chickens are major food sources for sylvatic T . guasayana based on the qPCR assay , while the cloning assay was able to discover that T . guasayana are also feeding on wild animals . Future studies should include sensitivity analysis of qPCR after blood meals have been subject to various lengths of digestion by the vector and , in addition , would benefit from this dual approach , the cloning assay to gather a priori information on actual food sources perhaps on a larger sample size of pooled triatomine blood meals , followed by the highly sensitive qPCR assay to determine feeding habits of individual triatomines . Combined with Noireau traps , these methods can help us unravel the role of sylvatic insect vectors in the transmission of human Chagas disease .
|
The World Health Organization ( WHO ) estimates that 7 to 8 million people are currently infected with Trypanosoma cruzi , the parasite that causes Chagas disease . The WHO recommends insect vector control as the primary prevention method; and insecticide spraying is the most commonly used intervention technique . Sylvatic insect vectors are a special concern because they are a source of reinfestation after insecticides have been applied to living quarters ( domestic ) and immediate surroundings ( peridomestic ) . To better understand sylvatic insect vector movement , we used two molecular biology techniques to detect the blood meal sources of sylvatic insect vectors . The first technique , cloning of 12S PCR products , allows us to cast a wide net and detect blood meal sources with no previous knowledge of vertebrates or mammals in the study site . After acquiring knowledge of vertebrates in the study site ( either through the aforementioned cloning technique , literature review or survey of the area ) , the second technique , the species-specific hydrolysis probe-based qPCR provides a highly sensitive assay for particular taxa .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"chagas",
"disease",
"neglected",
"tropical",
"diseases",
"tropical",
"diseases",
"protozoan",
"infections",
"parasitic",
"diseases"
] |
2014
|
Sources of Blood Meals of Sylvatic Triatoma guasayana near Zurima, Bolivia, Assayed with qPCR and 12S Cloning
|
Diverse sex-chromosome systems are found in vertebrates , particularly in teleost fishes , where different systems can be found in closely related species . Several mechanisms have been proposed for the rapid turnover of sex chromosomes , including the transposition of an existing sex-determination gene , the appearance of a new sex-determination gene on an autosome , and fusions between sex chromosomes and autosomes . To better understand these evolutionary transitions , a detailed comparison of sex chromosomes between closely related species is essential . Here , we used genetic mapping and molecular cytogenetics to characterize the sex-chromosome systems of multiple stickleback species ( Gasterosteidae ) . Previously , we demonstrated that male threespine stickleback fish ( Gasterosteus aculeatus ) have a heteromorphic XY pair corresponding to linkage group ( LG ) 19 . In this study , we found that the ninespine stickleback ( Pungitius pungitius ) has a heteromorphic XY pair corresponding to LG12 . In black-spotted stickleback ( G . wheatlandi ) males , one copy of LG12 has fused to the LG19-derived Y chromosome , giving rise to an X1X2Y sex-determination system . In contrast , neither LG12 nor LG19 is linked to sex in two other species: the brook stickleback ( Culaea inconstans ) and the fourspine stickleback ( Apeltes quadracus ) . However , we confirmed the existence of a previously reported heteromorphic ZW sex-chromosome pair in the fourspine stickleback . The sex-chromosome diversity that we have uncovered in sticklebacks provides a rich comparative resource for understanding the mechanisms that underlie the rapid turnover of sex-chromosome systems .
Genetic sex determination ( GSD ) is prevalent in vertebrates and is often accompanied by the presence of a heteromorphic chromosome pair in one sex . Birds and snakes have a ZW heteromorphic pair , where the W sex chromosome is female-limited; however , neither the bird nor the snake sex determination locus has been identified ( reviewed in [1] ) . Most mammals have an XY heteromorphic pair [2] , and the male-limited Y sex chromosome bears SRY , a male-determining gene [3]–[5] that is found in all but a handful of mammals [6]–[10] . However , this broad conservation of sex-chromosome systems across large taxonomic groups is not universal in vertebrates . Both simple ( XY and ZW ) and complex ( polygenic ) forms of GSD , as well as environmental sex determination ( ESD ) , are seen in teleost fish , lizards , turtles and amphibians [1] . Even closely related species within a genus might have different sex-determination systems . For example , the only other known vertebrate sex-determining gene , DMY in the medaka fish Oryzias latipes [11] , [12] , is not found in many closely related Oryzias species [13] , [14] . Comparative studies of sex-chromosome systems have supported the assertion that sex-chromosome pairs were autosomes prior to the acquisition of a sex-determination locus [15] . Sex-chromosome heteromorphy initially arises as a consequence of selection for a loss of recombination between linked sex-determining loci [16] . Once recombination is suppressed , intrachromosomal inversions and deletions and mobile sequence elements tend to accumulate in the nonrecombining region of the Y or W chromosome [2] , [17]–[20] . These physical changes to the sex chromosome result in heteromorphy seen in metaphase chromosome spreads , although it is not possible to state a priori whether the hemizygous sex chromosome will be the larger or smaller chromosome of a heteromorphic pair [21] . Theoretical studies have suggested that a further reduction of recombination on a sex chromosome might be favored when genes with alleles of sexually-antagonistic effect are linked to a sex-determination locus ( SEX ) [22] . For example , an allele that increases male fitness and reduces female fitness in an XY system benefits from absolute linkage with the male-determining SEX locus . Selection for linkage of sexually-antagonistic genes to SEX might also explain the rapid turnover of sex-determination loci and sex chromosomes between closely related species [23] . Several mechanisms could bring about this rapid turnover , including the appearance of a novel SEX locus on an autosome [15] , the transposition of a SEX locus between chromosomes in different lineages [24] , or fusions of an existing sex chromosome with an autosome [25] . Investigation of the process of sex-chromosome turnover requires detailed molecular , genetic , cytogenetic and phylogenetic analyses of sex determination systems that differ between closely related species . Teleost fishes are a particularly useful group to explore turnover of sex chromosome systems because different sex-determination mechanisms exist in closely related species [26]–[29] . For example , there is evidence for the evolution of a novel SEX locus in Oryzias [28] and transposition of the existing SEX locus in salmonids [24] . Furthermore , both XY and ZW GSD systems have been identified in species of Oryzias [28] , [30] , [31] , Xiphophorus [32] , and tilapiine cichlids [29] , [33] , [34] . Thus , teleost fishes also provide the opportunity to ask whether distinct forms of GSD found in closely related species have interconverted or evolved independently . Among fishes , the sticklebacks ( Gasterosteidae ) provide a particularly interesting system in which to investigate the evolution of sex determination and sex chromosomes . The first cytogenetic survey in this family reported the presence of a heteromorphic XY pair in the black-spotted stickleback ( Gasterosteus wheatlandi ) and a heteromorphic ZW pair in the fourspine stickleback ( Apeltes quadracus ) [35] . In the same study , evidence of a heteromorphic pair was not seen in the threespine stickleback ( G . aculeatus ) . The findings of this 1970 study , along with phylogenetic relationships between the stickleback species , are summarized in Figure 1 . Although later studies [36] , [37] also did not find evidence of a heteromorphic sex-chromosome pair in G . aculeatus , genetic mapping subsequently identified the presence of XY genetic sex determination on linkage group ( LG ) 19 in this species [38] . Using fluorescence in situ hybridization ( FISH ) , we have recently demonstrated that there is a heteromorphic XY pair corresponding to LG19 in G . aculeatus [39] . Genetic mapping has now demonstrated that the sex-determination locus in the ninespine stickleback ( Pungitius pungitius ) maps to LG12 , which is distinct from the G . aculeatus sex chromosome LG19 [40] . Taken together , these data suggest that different sex-determination systems and sex chromosomes have evolved within the stickleback clade . Combined with the recent development of genetic and genomic resources for both G . aculeatus and P . pungitius [40]–[43] , these small teleost fish are an excellent system in which to study the evolution of sex chromosomes and GSD . To systematically characterize the relationships between the sex-determination mechanisms and sex-chromosome systems in the stickleback family , we genetically mapped sex-determination loci and searched for heteromorphic sex-chromosome pairs using FISH in the North American stickleback species: G . wheatlandi , P . pungitius , the brook stickleback ( Culaea inconstans ) , and A . quadracus . Using FISH in P . pungitius , we identified a heteromorphic XY pair corresponding to LG12 , where the sex-determination locus has been mapped in this species [40] . We confirmed the presence of a heteromorphic pair in G . wheatlandi , although we find that males of this species have a diploid chromosome number of 41 , not 42 as previously reported [35] . Genetic mapping and molecular cytogenetics demonstrate that the G . wheatlandi Y chromosome consists of a fusion between LG12 and LG19 , resulting in an X1X2Y sex-chromosome system . However , neither LG12 nor LG19 is associated with a sex-determination locus or a heteromorphic sex-chromosome pair in C . inconstans or A . quadracus . Consistent with previous work [35] , we find that A . quadracus has a heteromorphic ZW pair , while C . inconstans has no heteromorphic sex-chromosome pair . The results we report here , summarized in Figure 1 , demonstrate the remarkable diversity of genetic mechanisms and chromosomal systems of sex determination that can be present within a clade of teleost fish that diverged approximately twenty million years ago ( MYA ) [44] .
Master sex-determination loci map to different Y chromosomes in G . aculeatus ( LG19 ) and P . pungitius ( LG12 ) [38] , [40] . The relationships between linkage groups in these two species were established by including markers of known genomic locations derived from G . aculeatus in the P . pungitius linkage map [40] . Furthermore , the sequences of the P . pungitius microsatellite markers used for map construction were BLASTed against the G . aculeatus genome ( http://www . ensembl . org/Gasterosteus_aculeatus/Info/Index ) , and the positions of 92% ( 156/169 ) were unambiguously identified [40] . The nomenclature for the P . pungitius linkage groups corresponds to the G . aculeatus nomenclature , and LG12 and LG19 represent distinct chromosomes in these two species [40] . To determine whether markers from either sex-chromosome pair ( LG12 or LG19 ) are linked to SEX in G . wheatlandi , we genotyped the 80 progeny ( 41 females and 39 males ) of three G . wheatlandi crosses with LG12 and LG19 markers derived from both G . aculeatus and P . pungitius ( Table S1 ) . Five markers from LG19 and six markers from LG12 were heterozygous in at least one of the parents of the three crosses ( Table S1 ) . For the nine markers that were heterozygous in a male parent , there was perfect concordance between the marker genotype inherited from the father and the sex phenotype of the progeny ( Table S2 ) , demonstrating that G . wheatlandi males are the heterogametic ( XY ) sex and that markers from both LG12 and LG19 are sex-linked in G . wheatlandi . For all five LG19 markers , the Y-linked allele is a null allele ( i . e . no PCR product is amplified ) , while none of the Y-linked alleles of LG12 markers are null . To further explore the relationship between LG12 and LG19 markers in these G . wheatlandi crosses , the complete genotypes of the ten markers informative in all three crosses were used to create a linkage map . Using a stringent LOD score of 10 . 0 , all ten markers were found in a single linkage group ( Figure 2A ) . When only the male meiotic data were used to create a linkage map , all markers were completely linked to each other and to SEX ( Figure 2B ) . However , when only the female meiotic data were used to create a linkage map , two independent linkage groups representing LG12 and LG19 were found ( Figure 2C ) . Our genetic mapping data suggest that one chromosome 12 and one chromosome 19 might be fused in male , but not female , G . wheatlandi . Consistent with these results , a karyogram made from a male G . wheatlandi metaphase spread ( Figure 3A ) contained 41 chromosomes ( 19 pairs and three unpaired ) . The presence of 41 chromosomes in male somatic tissue was seen in multiple metaphase spreads from multiple individuals from two different populations ( Table 1; Materials and Methods ) . The three unpaired chromosomes consist of a large submetacentric , a medium submetacentric , and a medium acrocentric chromosome . By contrast , the female karyogram comprises 21 chromosome pairs ( Figure 3B ) . Absence of the large submetacentric chromosome from the female karyogram defines it as the Y ( Figure 3A ) . To examine the relationship between LG12 and LG19 in G . wheatlandi , G . aculeatus and P . pungitius , we hybridized LG12 and LG19 FISH probes to metaphase spreads from females and males of all three species . In females of all three species , the LG12 and LG19 pairs appear homomorphic ( Figure 4 ) . We had previously demonstrated that the G . aculeatus LG19 is heteromorphic in males [39]; here we demonstrate that LG12 is homomorphic in males ( Figure 4 ) . There is a heteromorphic pair in the male P . pungitius karyogram ( Figure 5A ) that is absent from the female karyogram ( Figure 5B ) . However , we found that LG12 , and not LG19 , comprises the heteromorphic pair in male P . pungitius ( Figure 4 ) . Because both copies of chromosome 12 ( the X chromosome ) in female P . pungitius are metacentric , the metacentric chromosome 12 in P . pungitius males is the X , and the submetacentric chromosome 12 , which appears larger than the X , is the Y chromosome ( Figure 4; Figure 5A ) . In male G . wheatlandi metaphase spreads , both the LG12 probe and the LG19 probe for the Stn303 locus [39] hybridized to a large unpaired submetacentric chromosome , which is the Y chromosome ( Figure 3A; Figure 4 ) . The LG12 probe also hybridized to the medium unpaired acrocentric chromosome , while the LG19 probe hybridized to the medium unpaired submetacentric chromosome ( Figure 4 ) . In female G . wheatlandi metaphase spreads , the LG12 probe hybridized to a pair of acrocentric chromosomes , and the LG19 probe hybridized to a pair of submetacentric chromosomes ( Figure 4 ) . Based on a comparison of chromosome morphologies across species , we define the submetacentric chromosome 19 to be X1 and the acrocentric chromosome 12 to be X2 for G . wheatlandi males ( Figure 3A; Figure 4 ) . To further assess the physical relationship between the G . wheatlandi and G . aculeatus Y chromosomes , we used two additional LG19 FISH probes , one from the Idh locus and one from the Wt1a locus [39] . Each probe hybridized to a chromosome pair in female G . wheatlandi . However , each probe only hybridized to a single chromosome in males and did not hybridize to the large submetacentric Y ( Figure 6 ) , indicating that loci present on the G . aculeatus Y might be deleted from the G . wheatlandi Y . These data are consistent with the finding of null Y-linked alleles for all five LG19 markers used to genotype the G . wheatlandi mapping crosses . We next asked whether SEX-linked markers from LG12 or LG19 were associated with a single locus controlling male or female sexual development in two additional stickleback species . We genotyped a single C . inconstans cross and a single A . quadracus cross and found no such associations ( Table S1 ) . Seven LG12 markers ( 4 from G . aculeatus and 3 from P . pungitius ) were informative in our A . quadracus cross , yet none were linked to SEX ( Table S1 ) . Similarly , two LG19 markers , Stn194 and Pun117 , were linked to SEX in G . aculeatus but not in A . quadracus ( Table S1 ) . Although very few LG12 and LG19 markers were informative in our C . inconstans cross , neither the LG12 marker Pun234 nor the four LG19 markers , Stn186 , Cyp19b , Pun168 , and Pun268 , were linked to SEX in C . inconstans ( Table S1 ) . To determine whether any G . aculeatus or P . pungitius genetic markers are linked to a sex-determination locus in these species , we genotyped the C . inconstans cross and the A . quadracus cross with all available G . aculeatus and P . pungitius markers [40] , [41] . There was no evidence for an association between any marker and sex phenotype in either species . However , many G . aculeatus and P . pungitius markers either failed to work in C . inconstans or A . quadracus , or were not heterozygous in the parents of the crosses ( Table S3 ) . Therefore , we were unable to test markers from all G . aculeatus and P . pungitius linkage groups . We hybridized LG12 and LG19 FISH probes to metaphase spreads from A . quadracus and C . inconstans males and females . Consistent with the genetic mapping data in these species , neither of the chromosome pairs identified by hybridization with LG12 and LG19 probes were heteromorphic in either sex of either species ( Figure S1 ) . There was no evidence for obvious heteromorphy of any chromosome pair in C . inconstans males or females ( Figure S2 ) . By contrast , there is a heteromorphic chromosome pair in the female A . quadracus karyogram , in which the female-limited W chromosome appears larger than the Z at metaphase ( Figure 7 ) .
Our cytogenetic survey of the family Gasterosteidae has uncovered a diversity of sex- chromosome systems not previously identified [35] . Although initial cytogenetic surveys of stickleback fish did not find a heteromorphic XY pair in P . pungitius [35] , [36] , genetic mapping has shown that P . pungitius have an XY pair corresponding to LG12 [40] . Here , we used FISH to demonstrate that chromosome 12 is a heteromorphic pair in P . pungitius males ( Figure 4 ) . Although we only examined karyotypes from a single Canadian population of P . pungitius , our results are very similar to those obtained in a recent cytogenetic study of two Polish populations , in which a heteromorphic XY pair was also identified [45] . While our G . wheatlandi male karyogram supports the prior report of a heteromorphic XY pair , the same study also reported a male diploid chromosome number of 42 [35] . Our karyogram shows that male G . wheatlandi have a 2n = 41 karyotype , while females are 2n = 42 ( Figure 3 ) . It is unlikely that presence of an odd diploid chromosome number in males is due to experimental artifact , as this result was obtained in multiple metaphase spreads from multiple individuals obtained from natural populations in Canada and Massachusetts ( Table 1; Materials and Methods ) . Furthermore , the genetic mapping data supporting the physical linkage between LG12 and LG19 was obtained using a third G . wheatlandi population from Maine . Although we have not found evidence for intraspecific polymorphism in the stickleback sex chromosomes studied here , we cannot rule this out as a possible reason for the discrepancy between the current study and previous work [35] . However , we believe that the higher resolution of the molecular cytogenetic techniques used here explain our findings of additional heteromorphic sex-chromosome systems in the Gasterosteidae . Genetic and cytogenetic evidence support the conclusion that a fusion between the G . aculeatus Y chromosome ( YLG19 ) and an autosome ( LG12 ) created a neo-Y chromosome in G . wheatlandi males . This fusion created an X1X2Y sex chromosome system , which explains the odd diploid chromosome number in G . wheatlandi males . Our data suggest that the X1X2Y system in G . wheatlandi is derived from the G . aculeatus Y chromosome , rather than the P . pungitius Y chromosome . All available phylogenies support a closer relationship between G . aculeatus and G . wheatlandi than between G . wheatlandi and P . pungitius ( Figure 1 ) . Our cytogenetic and FISH data also support a closer relationship between the G . aculeatus and G . wheatlandi karyotypes: LG12 is acrocentric in G . aculeatus and G . wheatlandi females , but metacentric in P . pungitius females , while LG19 is the submetacentric X chromosome in G . aculeatus and G . wheatlandi , but metacentric in P . pungitius ( Figures 3–5 ) . Our data are consistent with other studies in teleost fish , where species with X1X2Y systems often have one less chromosome than sister taxa [46] , [47] . If the G . wheatlandi X1X2Y male karyotype had been created by fission of an ancestral X rather than a Y-autosome fusion , we should have observed the diploid chromosome number 2n = 43 in males , rather than the observed 2n = 41 ( Figure 3A ) . These data suggest that derived Y chromosome-autosome fusions might be the predominant source of X1X2Y sex chromosome systems in teleost fishes , although alternative mechanisms likely account for X1X2Y systems discovered in insects and mammals [48] . Our FISH data further suggest that the fusion of the ancestral YLG19 chromosome and the acrocentric LG12 autosome resulted in loss of one arm of the YLG19 in G . wheatlandi males ( Figure 6 ) . The two G . aculeatus LG19 FISH probes that did not hybridize to the G . wheatlandi Y ( Figure 6 ) are from the q arm of the G . aculeatus Y ( Yq ) [39] , while the LG19 probe that did hybridize to the G . wheatlandi Y ( Figure 4 ) is from the p arm of the G . aculeatus Y [39] . Furthermore , the Y-linked alleles of all four G . aculeatus Yq markers segregated as null alleles in G . wheatlandi , consistent with loss of the q arm . In further support of the hypothesis that LG12 was autosomal prior to fusion with the Y , none of the LG12 markers have null alleles , suggesting that extensive degeneration has not yet occurred on the LG12-derived region of the G . wheatlandi Y . This is consistent with the interpretation that one copy of LG12 fused to the existing YLG19 chromosome in G . wheatlandi males in the past 10 million years since G . wheatlandi and G . aculeatus diverged [44] . The loss of one arm of YLG19 is consistent with the creation of the G . wheatlandi Y by an unbalanced Robertsonian translocation between a metacentric Y and an acrocentric autosome . To more precisely map the rearrangements that have occurred between the two X chromosomes and the Y chromosome in G . wheatlandi , we will need a more extensive cytogenetic analysis as we have accomplished for the G . aculeatus X and Y [39] . However , the current data suggest that additional deletions of Y chromosome material have occurred on the G . wheatlandi Y relative to the G . aculeatus Y . No hybridization of FISH probes containing the Wt1a or Idh genes is observed on the G . wheatlandi Y chromosome ( Figure 6 ) , although these two probes do hybridize to the G . aculeatus Y chromosome [39] . In particular , the deletion of the region around the Idh FISH probe on the G . wheatlandi Y ( Figure 6B ) explains why a male-specific allele of this locus was not identified in our previous study [38] , leading to the erroneous conclusion that the Y chromosomes of these two species were unrelated . Finally , the putative loss of an entire arm of the Y chromosome in G . wheatlandi males , as well as the existence of a 6 Mb deletion on the G . aculeatus Y chromosome [39] , suggests that dosage compensation mechanisms may have evolved in sticklebacks . We plan to investigate this possibility in the future . Our data suggest that XY sex chromosomes have been independently derived on LG19 and LG12 in the Gasterosteus and Pungitius lineages , respectively . It is possible that sex determination also arose independently in these lineages . Precedent for the independent evolution of XY sex determination in closely related species exists in the teleost clade Oryzias [28] . However , it is also possible that G . aculeatus , G . wheatlandi , and P . pungitius all share a common sex-determining locus , but that SEX has transposed between LG12 and LG19 in the two lineages; a similar transposition of the SEX locus to four different chromosomes has been seen in salmonids [24] . Finally , although we have argued that the G . wheatlandi Y was derived from the G . aculeatus Y , it is possible that sex determination arose independently in these species . In order to distinguish these possibilities , the identity of SEX in these three species must be determined . If all three species share a common sex-determining factor , then transposition of SEX between linkage groups has likely occurred; the identification of different sex-determining factors in the three species would provide evidence of the independent evolution of XY sex determination in sticklebacks . In any of these scenarios , LG12 appears to have been selected for SEX linkage at least two independent times: by fusion to the YLG19 chromosome in G . wheatlandi and by acquisition of either a transposed or newly evolved SEX locus in P . pungitius . It has been suggested that selection for linkage between autosomal genes with sexually antagonistic effects and SEX might drive Y-autosome fusions [25] , the appearance of a new SEX locus on an autosome [23] , or the transposition of an existing SEX locus to an autosome [23] . Thus , linkage between LG12 and SEX in both G . wheatlandi and P . pungitius suggests that LG12 might have an abundance of genes with differential fitness effects in males and females and thus be predisposed to becoming a sex chromosome . Comparative genomic analysis of the autosomal and sex-linked forms of LG12 in the different stickleback species might yield insight into the presence and types of sexually antagonistic alleles that play an important role in the evolution of sex chromosomes in sticklebacks . In addition to multiple XY systems , we also found evidence for both XY and ZW systems in sticklebacks . Both cytogenetic and genetic data suggest that the ZW system of A . quadracus is not related to the XY systems of G . aculeatus , G . wheatlandi or P . pungitius , raising the possibility that the ZW system arose independently . However , we do not know what the ancestral sex-chromosome state is for the sticklebacks , so an accurate parsimony-based reconstruction of the evolution of XY and ZW GSD in sticklebacks is not currently possible . It will be useful to karyotype the European fifteenspine stickleback ( Spinachia spinachia ) , a close relative of A . quadracus , to determine whether it too has 46 chromosomes and a heteromorphic ZW pair . We are currently working to identify SEX-linked sequences in A . quadracus using unbiased methods and to identify the linkage group comprising the Z and W chromosomes by FISH with G . aculeatus BAC probes . These studies will allow us to determine which autosome ( s ) gave rise to this ZW pair . Additional efforts will focus on identifying the sex determination mechanism of C . inconstans . Although we have not yet identified sex-linked markers in either C . inconstans or A . quadracus , these studies have been limited by the availability of polymorphic markers . Therefore , it is still possible that there is a simple genetic sex determination mechanism in C . inconstans . However , it is also possible that C . inconstans uses ESD or complex GSD . Knowing the sex determination mechanism in C . inconstans might shed light on the transitions between XY and ZW systems in this family . Transitions between XY and ZW GSD occur via indirect or direct mechanisms . For example , an interim period of ESD might facilitate an indirect transition between two forms of GSD [1] . A more direct transition between XY and ZW forms might occur , as exemplified by recent work . In tilapiine fishes , two species ( Oreochromis aureus and O . mossambicus ) have complex genetic sex determination in which LG1 and LG3 are both associated with sex determination loci [29] , [33] . The phylogenetic positions of these species provide a direct link between two related species in which LG1 is associated with a simple XX/XY system and two other species in which LG3 is associated with a simple ZZ/ZW system [29] . A similar link exists in the platyfish ( Xiphophorus maculatus ) , where some populations have W , X , Y and Z chromosomes and closely-related species have either XY or ZW GSD [32] . In the Japanese frog ( R . rugosa ) , there is evidence that an existing XY sex chromosome became a ZW sex chromosome in a derived population [49] . Finally , comparative mapping of the platypus sex chromosome chain suggests that the monotreme XY GSD system might have directly evolved from a bird-like ZW system , while the therian XY system might have evolved independently [1] , [10] , [50] , [51] . Additional genetic and genomic analyses of all stickleback species might elucidate whether the ZW and XY systems directly interconverted or were independently derived in sticklebacks . Teleost fishes are useful organisms in which to study the evolution of sex determination and sex chromosomes . In particular , several teleost fish species have X1X2Y sex chromosome systems [46] , [47] , [52]–[57] . While a benefit of fusion of sex chromosomes to autosomes has been suggested [25] , to our knowledge this is the first report of the evolution of an X1X2Y system in which the fused Y comprises two chromosomes that are used as distinct Y chromosomes in two closely related species . X1X2Y sex chromosome systems are not reported as frequently as XY or ZW systems , which could mean that they lead to evolutionary dead ends or exist as transitional states [58] . For example , following the fusion of a Y to an autosome , the X1 and X2 chromosomes might fuse to restore diploidy . The discovery of an X1X2Y system in G . wheatlandi , sister species to G . aculeatus , for which many molecular , genetic and genomic tools have been developed [41]–[43] , will facilitate further characterization of the mechanisms and evolutionary forces underlying the transition between simple XY sex chromosomes and X1X2Y neo-sex chromosomes . Furthermore , the availability of these tools in sticklebacks will also be important to improve our understanding of the transition between male and female heterogamety that is evident in this group of fishes with diverse sex chromosome systems .
All animal work was approved by the Fred Hutchinson Cancer Research Center Institutional Animal Care and Use Committee ( protocol #1575 ) . Three G . wheatlandi crosses were generated using males and females collected from Wells , ME in May 2003 . Sperm from a single G . wheatlandi male was used to fertilize the eggs of a single G . wheatlandi female ( cross 1 ) ; the sperm of a second G . wheatlandi male was used to fertilize the eggs of two different G . wheatlandi females ( crosses 2 and 3 ) . The progeny of each of the three crosses were grown in separate tanks . A single C . inconstans cross was generated using a female collected from Fox Holes Lake ( Northwest Territories , Canada ) and a male collected from Pine Lake ( Wood Buffalo National Park , Alberta , Canada ) in June 2005 . A single A . quadracus cross was generated using a single female and a single male collected from Pilgrim Lake ( Cape Cod National Seashore , MA ) in May 2004 . For all crosses , the sex of the progeny was determined by visual inspection of the gonads . DNA was prepared from the caudal fin of each individual by phenol-chloroform extraction , followed by ethanol precipitation . PCR genotyping with G . aculeatus microsatellite markers and P . pungitius microsatellite markers was performed as previously described [40] , [41] , except that the reactions were run on an ABI 3100 and the genotypes were analyzed using ABI GeneMapper 3 . 7 ( Applied Biosystems ) . Genetic linkage maps were created in JoinMap3 . 0 [59] using default parameters . Both Kruskal-Wallis tests for significant associations between genotype and sex phenotype , as well as interval mapping , were performed in MapQTL4 . 0 [60] . Metaphase spreads were prepared as described [39] using G . wheatlandi males collected from Baie de L'Isle-Verte National Wildlife Area ( Québec , Canada ) in May 2003 , G . wheatlandi males and females collected from Demarest Lloyd State Park ( Dartmouth , MA ) in May 2005 and May 2007 , P . pungitius and C . inconstans males and females collected from Pine Lake ( Wood Buffalo National Park , Alberta , Canada ) in June 2007 , A . quadracus males collected from Pilgrim Lake ( Cape Cod National Seashore , MA ) in May 2005 , and A . quadracus males and females collected from Demarest Lloyd State Park ( Dartmouth , MA ) in May 2005 and 2007 . Fluorescence in situ hybridization ( FISH ) experiments were performed on metaphase spreads as described [39] . Bacterial artificial chromosomes ( BACs ) from the G . aculeatus CHORI-213 library [42] were used as FISH probes . Linkage group ( LG ) 19 probes ( Stn303: CH213-035N15; Idh: CH213-101E08; and Wt1a: CH213-180J08 ) were identified previously [39] . The LG12 probe ( CH213-140B10 ) , spanning 0 . 35 to 0 . 56 Mbp on the G . aculeatus public genome assembly of scaffold 6 ( part of the LG12 sequence ) , was found to contain genetic marker Stn144 [41] at 0 . 51 Mbp using a computational approach [39] .
|
Sex chromosomes have independently evolved many times in animals , plants , and fungi . Although some sex chromosomes have been maintained across different species for over a hundred million years , there are other systems in which sex chromosomes appear to turn over within a few million years . Because sex determination is such a fundamental biological process , this rapid turnover of sex chromosomes is puzzling . Theoretical work has attempted to identify the mechanisms that might underlie the transitions between sex-chromosome systems . However , there are few empirical data to test these hypotheses . In this study , we have uncovered an unanticipated diversity of sex-chromosome systems in the stickleback fishes ( Gasterosteidae ) . We have found that each stickleback species examined has a unique sex-chromosome system , with evidence for two independent XY systems , a derived Y-autosome fusion that has created an X1X2Y neo-sex chromosome , and a ZW system . Thus , stickleback fishes are excellent model species for investigating the mechanisms that lead to rapid turnover of sex chromosomes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/chromosome",
"biology",
"evolutionary",
"biology/evolutionary",
"and",
"comparative",
"genetics"
] |
2009
|
Turnover of Sex Chromosomes in the Stickleback Fishes (Gasterosteidae)
|
Efficient cognitive decisions should be adjustable to incoming novel information . However , most current models of decision making have so far neglected any potential interaction between intentional and stimulus-driven decisions . We report here behavioral results and a new model on the interaction between a perceptual decision and non-predictable novel information . We asked participants to anticipate their response to an external stimulus and presented this stimulus with variable delay . Participants were clearly able to adjust their initial decision to the new stimulus if this latter appeared sufficiently early . To account for these results , we present a two-stage model in which two systems , an intentional and a stimulus-driven , interact only in the second stage . In the first stage of the model , the intentional and stimulus-driven processes race independently to reach a transition threshold between the two stages . The model can also account for results of a second experiment where a response bias is introduced . Our model is consistent with some physiological results that indicate that both parallel and interactive processing take place between intentional and stimulus-driven information . It emphasizes that in natural conditions , both types of processing are important and it helps pinpoint the transition between parallel and interactive processing .
Research on human action control typically distinguishes between two types of action: re-actions issued in response to some external stimulus event and voluntary actions based on an internal decision to act [1] , [2] . There has been a long-standing debate on whether these two types of action are controlled by two different brain systems . Some studies using neurophysiological , behavioural or neuropsychological methods suggest that this is indeed the case [3]–[6]; [ for review] , [ see Refs . 2 , 7] . Patients suffering from “utilization behaviour” ( UB ) , for example , show a strong tendency to use objects they spot in the environment without any clear need or purpose [8] . This behaviour has been explained in terms of a lack of inhibition and modulation of the external action system due to damage in the voluntary system [8] . Other studies , however , point to the existence of common control mechanisms . In a recent study [9] , for example , participants were asked to prepare and execute left- or right-hand voluntary actions . Occasionally , the voluntary action preparation was interrupted by a stimulus requiring either a left- or right-hand response . The results showed that increased voluntary motor preparation , as assessed by the readiness potential , produced faster stimulus-driven responses on congruent trials ( i . e . , when participants voluntarily prepared the same hand that was also used in response to the target stimulus ) than on incongruent trials . This suggests that voluntary and stimulus-driven actions share some central preparatory mechanisms . It is evident that voluntary and stimulus-driven action control are at least to some extent based on separate mechanisms , for the simple reason that stimulus-driven but not voluntary action control needs to be linked to the perceptual system . Conversely , it is also obvious that voluntary and stimulus-driven action control are to some extent based on common mechanisms , as it seems undisputed that the most final steps of action execution use the same cerebral structures . Therefore , we propose that an appropriate model to account for voluntary and stimulus-driven actions should be composed of two stages , a first stage where the two types of actions are dissociated and a second stage where they are combined . Thus , the issue is not whether there are one or two systems but rather where the transition between the two stages is . As a consequence , to better understand the interaction between voluntary and stimulus-driven action control , we need to develop tools that embrace the notion of the existence of both differences and commonalities between voluntary and stimulus-driven action and that are capable of pinpointing them . The goal of the current study is to provide evidence for a two-stage model of action control . Our study is very tightly rooted in a new methodological approach developed by Stanford and colleagues [10]–[12] . Participants are required to initiate a choice response to a left or right target stimulus . Importantly , the target is presented with a variable delay , called gap , after participants began to prepare their action . For very short gaps the action is truly stimulus-driven , whereas for long gaps , it is truly intentional , as participants no longer have the time to take the stimulus into account . For intermediate gap durations , the target is possibly able to influence the ongoing voluntary action preparation . This paradigm mimics situations in which we have to anticipate and start preparing an action even before the stimulus the action is supposed to respond to is available , for instance , when a goal keeper has to anticipate where the opponent will shoot the ball and prepare his action even before having any visual cue . The paradigm enables us to investigate whether , and if so , how and from when on voluntary and stimulus-driven action preparation interact . In one of the studies the present expriment is based on investigated monkeys were required to perform internally/externally chosen left/right saccades [10] . The authors proposed a race-to-threshold model where motor preparation for left/right responses accumulates over time to account for the data [13]–[15] . Models where perceptual evidence accumulates over time have indeed been supported by both behavioral and neurophysiological studies [13]–[17] . However , this particular type of task [10] is not purely perceptual in nature . The participant first starts to prepare an action without the perceptual information being available . It is clear that the evidence accumulated during this stage is not “perceptual” , but rather internal . To this extent , the paradigm is close in spirit to another one used in action control [18] . To capture this aspect of the task , the current research focuses on how , precisely , internal ( voluntary ) and external ( stimulus-driven ) accumulation of evidence interact . To model our results we conceptualize internal and external accumulation of evidence as a two-system process having both a separate and a common stage . This hybrid model includes a transition threshold below which the signals accumulate separately , and above which they interact ( resulting in facilitation in case of congruent and interference in case of incongruent actions ) . Our experiments also allows us to quantify the relative importance of each stage , thus demonstrating that they are both necessary .
In our task [10] , participants were presented with a Go Signal followed by the appearance of a Target Signal to the left or the right of fixation and a distractor object on the other side ( see Fig . 1 ) . The target was defined by its color ( red or green ) chosen randomly for each trial and indicated to the participant as the color of the fixation point . Participants were instructed to press a key ( Go Key ) in response to the Go Signal and immediately afterwards to initiate a choice response to the side of the target ( left or right ) . The Target Signal was presented after the Go Key with a variable delay called gap . Therefore , in trials where the gap is too long , participants had to initiate their choice response even before appearance of the Target Signal . In order to prevent participants from waiting for the presentation of the target and from answering systematically too quickly without taking into account the target , a feedback procedure was introduced that encompassed both speed and accuracy performance ( see Methods for details ) . All participants had an accuracy above chance ( mean = 71 . 5% correct , SD = 5 . 8% ) . We present data pooled across all ten participants . As anticipated , accuracy significantly decreased as the gap increased ( F ( 10 , 99 ) = 9 . 4 , p<0 . 0001 ) , simply because participants did not have any opportunity to revise their initial decision if the gap was too long . In addition , Response Times ( RT ) , that is , the time between the Go Key press and the choice response , increased gradually with increasing gap ( F ( 10 , 99 ) = 6 . 6 , p<0 . 0001 ) . However , the slope of RT increase with gap was less than one for all participants ( mean estimated slope = 0 . 34 , maximum of the upper bound of the 95% individual confidence interval = 0 . 71 ) indicating that participants possibly waited for the target to appear on some trials but not all ( Fig . 2 A–B ) . A key variable of the analysis is the raw Processing Time ( rPT ) introduced by Stanford and colleagues ( 10 ) . The rPT is the time during which the target information was available before the choice response was carried out . More precisely , the raw processing time is defined as: rPT = . Thus , positive values correspond to a choice response after target onset , and negative values correspond to a response before target onset . The percentage of correct choice responses increased sharply with rPTs ( Fig . 2C ) : under a critical value of rPT ( 202 ms± . 4 , s . e . computed with a Bootstrap procedure ) responses were given at random whereas above this value accuracy reached quickly 100% . Fig . 2D shows the normalized distribution of the rPTs separately for correct and error trials . The distribution of correct choice responses looks like the superposition of two component distributions , one identical to the rPT distribution in the case of erroneous responses , and one specific to correct responses . This component reflects actions carried out without perceptual information being taken into account ( because the rPT was too short ) . The second component of the correct distribution , corresponding to longer rPTs , reflects actions in response to the target . We propose a hybrid model in which the decision of a particular choice response is the result of a two-stage race between an internal variable that codes randomly for one or the other response and an external variable that codes for the target side . During the first stage of the race , the two variables accumulate independently , each at a constant rate drawn from a lognormal distribution of a certain mean and variance . The internal variable starts accumulating as soon as the Go key is pressed while the external variable starts accumulating only at the appearance of the target ( see Fig . 3 ) . Indeed , until target presentation the participant has no sensory information to rely on . It is only after both variables crossed a first threshold , the Transition Threshold , that the common stage starts where the external variable influences the internal one . We distinguish congruent from incongruent trials depending on whether internal and external variables code for the same or different responses . In congruent trials , the external facilitates the internal variable . In contrast , in incongruent trials , the external inhibits the internal variable . The first variable to cross the second threshold , the Response Threshold , triggers execution of the action it codes . In our task the action is most often triggered by the internal variable because the internal variable always starts accumulating first . However , it can happen ( especially for short gaps ) that the external crosses the second threshold before the internal crosses the first threshold . In that case that resembles a traditional reaction time experiment , the action is purely triggered by the external stimulus . Please notice that the RT distributions are necessarily a mixture of congruent and incongruent trials ( where internal and external variables coded for the same or different responses ) . However , the congruency of the two variables is in principle unknown to the experimenter , as the internal decision of the participant is not known . To deal with this problem , we present a method to estimate the distributions for the two types of trials ( congruent and incongruent ) in the second Experiment . The model includes seven parameters: the means and standard deviations of the lognormal distribution of the internal and external accumulation rates , the Transition Threshold ( the Response Threshold is set arbitrarily to 1000 , all other parameters being scaled relative to this value ) , the acceleration factor A that represents the time needed for the external variable to fully influence the internal one , and an execution delay . Table 1 presents the values of the parameters corresponding to the best fit for each participant obtained by maximizing the likelihood of the rPT distribution ( the averaged model fit is represented as continuous lines in Fig . 2 ) . Because our task is similar to that used by Stanford et al . [10] , we can compare the performance of our hybrid model with their model , which is a single race-to-threshold between two decision variables , representing right and left choices . The two variables start accumulating with randomly drawn rates , then , after the gap , color discrimination affects the decision process by accelerating the variable representing the target side and decelerating the variable of the distractor side . Without affecting the spirit of their model , we used a slightly simplified version that contained only eight parameters rather than eleven ( see Text S1 ) . We also compared our model to a simpler version of it that was used in another task [19] , a version of the drift diffusion model [13] adapted to our task and an Ornstein-Uhlenbeck process [20] . In this model there is only an independent race stage without then transition threshold . We compared the models at the group level [21] using BIC ( Bayesian Information Criterion ) and AIC ( Akaike Information Criterion ) for each model and each subject as a measure of evidence . The performance of our model was superior to the other models: the exceedance probability [21] , i . e . the probability that our model was the more frequent in our population of subject , was greater than . 95 both for the AIC criteria and the BIC criteria ( see Table S1 , S2 in Text S1 for the individual AIC and BIC values ) . Our model is based in part on the accumulation of evidence of an internal variable that is aimed arbitrarily to one side or the other . When this internal variable crosses the transition threshold , it starts to be affected by the stimulus-driven process . At that point , the stimulus-driven process can either facilitate or inhibit the intentional process , depending on whether the initial decision is congruent or incongruent with the stimulus . Unfortunately , in our first experiment , there is no way to analyze separately the congruent and incongruent conditions , simply because we do not know what the initial decision was . Rather than asking explicitly to the participant what his initial decision was at the end of each trial , we slightly modified the design of our first experiment . In order to estimate the rPT distributions separately for congruent and incongruent trials , we introduced a frequency bias in the second experiment . Unbeknown to the participants , the probability of the target being on one or the other side depended on the pitch of the Go signal . We refer to the side that had the higher probability ( 65% ) to be the target side as the “more frequent side” . Eleven out of fourteen participants expressed the expected response bias for the more frequent side . The bias ranged from 55% to 75% ( mean = 65 . 3% , SD = 6% ) . The data of these eleven participants were analyzed together . The remaining three subjects were excluded from the analysis either because the bias was too extreme or absent . The basic features of the results of Experiment 2 are identical to Experiment 1 ( see Fig . S2 in Text S1 ) . We adapted our hybrid model to this second experiment . In the model described earlier , both sides were chosen with the same a priori probability . The response bias in the current experiment is modeled by introducing a bias α on the choice of the internal variable for the more frequent side . If participants did adapt perfectly to our experimental setup , the value of α should be 0 . 65 . The new hybrid model including this additional bias parameter was adjusted to the individual data; the average of the best fits for each subject is superimposed on the curves in Fig . S2 in Text S1 and their parameters are shown in Table 2 . In the model , the external and the internal variables either code for the same side or they code for different sides , thereby defining congruent and incongruent trials . As a consequence , rPT distributions can be seen as the mixtures of two distributions corresponding to congruent and incongruent trials . Because of the bias in the model , the proportion of correct trials on the more frequent side is α for congruent trials and ( 1- α ) for incongruent trials . On the less frequent side , the proportion of correct trials is ( 1- α ) for congruent trials and α for incongruent trials . Using a linear transformation accounting for the bias we can thus estimate the rPT distributions for correct congruent and incongruent trials ( Fig . 4 ) . The figure reveals that target information needs more time to influence incongruent trials than congruent ones , notably the incongruent model distribution peaks 48 ms later than the congruent one . Keeping in mind that the external variable always codes for the correct side , error trials are thus necessarily incongruent trials . Therefore , the rPT distribution of error incongruent trials can be computed by taking all the error trials together .
The aim of the present study was to assess and model the interaction of stimulus-driven and voluntary action control . Participants were presented with the stimulus of a speeded perceptual discrimination task during the preparation of a voluntary action . Depending on the time gap between the beginning of the trial and the presentation of the target stimulus , voluntary action preparation was differently advanced when the target appeared , thus , enabling us to trace the effect of voluntary action preparation on stimulus-driven behavior . In the first experiment , a good predictor of accuracy was the raw processing time rPT , which is the time during which the target information was available before the stimulus-driven choice response was executed . For short rPTs , response accuracy was at chance . However , as soon as rPTs exceeded a critical value , it rapidly increased to reach ceiling performance . In other words , for short rPTs participants performed in an internal mode , choosing the response at random . It is only for longer rPTs that they take target information into account . The transition between purely internal and based-on-evidence modes was very fast although we analyzed the pooled data set . This shows that both the critical rPT value and the fast transition are quite robust among participants , since inter-subject variability would smooth the curve . Overall , our behavioural results are thus very much in line with the two monkeys' behaviour in in Stanford et al . 's study [10] . A critical difference in the behavioural pattern concerns the RT that are longer in our study , which is consistent with the fact that saccades initiations ( as in the monkey study ) are much faster than manual response [22] . To account for the shape of the distribution of rPTs , we suggested a physiologically founded model , the hybrid model , which is in essence a two-stage race model between a variable of the external and a variable of the internal action system . Our model is “hybrid” in two ways . First because it incorporates two systems , external and internal , secondly because the race has two stages: a stage at which evidence in the two systems accumulates independently and a later stage at which evidence accumulation in the two systems interacts . The distinction between internal and external systems follows the literature of stimulus-driven and voluntary action [5] , [6] , [8] , [18] , [23] . A recent study [24] that combines functional magnetic resonance imaging ( fMRI ) and pattern recognition shows that this distinction between two systems is also relevant in the framework of perceptual decision making . Indeed , this study shows that decisions regarding highly visible stimuli are predicted by visual brain areas , whereas it is not the case for low visibility when participant decisions are at chance level . In the latter case , the precuneus , which has been shown to encode “free” decisions [25] , is a good predictor of the final choice . However , our hybrid model does not reduce to a simple “switch” between a guessing system and a perceptual decision system since it has two stages . This fits with the growing body of evidence that information processing is not purely sequential [26] . The distinction of these two stages is based on the fact that the external and the internal action routes are necessarily separated at least up to some point , since the external system needs to make contact to visual processing , and probably merged from that point on . As outlined above , the hybrid model fits our data very well , better than models that only incorporate a single system or a single stage [10] , [13] , [19] , [20] . Since , in the hybrid model , accumulation of evidence occurs in two systems , an external and an internal system , a particular trial can be congruent or incongruent , external and internal variables either coding for the same response or for different responses , respectively . The congruency of a trial determines how the two variables interact in the second stage of processing . In congruent trials the internal variable is facilitated , whereas in incongruent trials , it is inhibited . As a consequence rPTs of the based-on-evidence component should be longer in incongruent than in congruent trials . To test our model it was therefore essential to have rPT distributions separately for congruent and incongruent trials . However , note that it is difficult to know whether a particular trial is congruent or incongruent , since this information requires access to the variable coding for the internally chosen action . One solution to this problem is to ask participants at the end of the trial which action they initially prepared and , using some brain imaging technique ( e . g . , EEG ) , to cross-check the introspection of the subject ( e . g . , by means of ERPs; see Ref . [9] ) . Here we chose a different approach: instead of relying on participants' introspection , we decided to deduce the congruent/incongruent distributions analytically in the second experiment . To do so , we made the probability of the target being presented on one or the other side depends on the pitch of the Go signal . Due to the induced response bias α , the rPT distributions of correct trials result from a proportion of α congruent and ( 1-α ) incongruent trials for the more frequent response and from a proportion of ( 1-α ) congruent and α incongruent trials for the less frequent response . Thus , in Experiment 2 , we were able to estimate the rPT distributions for correct congruent and incongruent trials by means of a linear transformation; this was not possible in Experiment 1 , where α = 1- α , making a linear transformation impossible . The results show that “based-on-evidence” incongruent trials correspond to longer rPTs than congruent trials , suggesting interaction between the internal and the external system . This is in agreement with the results reported in a recent study [9] in which the authors found reaction times to be longer for incongruent than for congruent trials . The hybrid model also allows us to estimate the relative amount of processing required in the two stages ( independent and common ) , as the relative importance of the two stages is captured by the value of the transition threshold . Experiments 1 and 2 revealed that a considerable amount of processing is done in both stages so that both stages are necessary . In both experiments , the mean threshold was approximately around 50% of the overall evidence accumulation , showing that about half of processing is done separately and the other half in a combined way . Moreover , the individual participant variability is well capture by our model . In most previous modeling attempts , only one stage is present: either the two variables accumulate independently of each other or they interact during the whole race . For instance , a recent study [27] introduced several race-to-threshold units to model a Go-No Go task . If the variable in the Stop unit reaches its threshold before the variable in the Go unit , it cancels the race of the later ( see also Ref . [28] ) . It is important to keep in mind that even though they are simultaneous , the races in the different units stay independent of each other: the rate of accumulation of one variable is not influenced by any other . In other models , like the leaky accumulator [15] , variables inhibit each other during the whole process . For instance , on the late distractor effect in saccadic inhibition [29] has been successfully modeled by a dual input combined with mutual . To our knowledge , only one recent study has also proposed a two-stage diffusion model . In an effort to account for behaviors that resemble a change of mind in the course of a manual action , Resulaj and colleagues [30] introduced a change of mind bound and a change of mind deadline to the original diffusion process . This model can be seen as the concatenation of two diffusion processes where a second diffusion can be initiated after a first decision . In contrast , our model clearly distinguishes a first stage where two variables are processed independently from a second stage where these variables interact . We believe that models based on two stages of processing will be inspiring for future attempts of modeling race-like phenomena .
Participants were voluntary and gave their informed consent . Research was approved by the Ethics committee for biomedical research ( CERB ) Ile de France II . Ten healthy participants with normal or corrected-to-normal vision participated in the first experiment ( 6 males , 4 females; mean age: 23 years and 3 months , SD: 1 year and 3 months ) . Fourteen healthy participants with normal or corrected-to-normal vision participated in the second experiment ( 8 male , 6 female; mean age: 23 years and 5 months , SD: 2 years and 1 month ) . All subjects were naive with respect to the goal of the experiment . Visual stimuli were presented on a computer screen LIYAMA HM 903 DTA ( 19 in ) . The experiment was controlled using Matlab and the Psychtoolbox [31] , [32] . Visual stimuli were two colored dots ( target and distractor ) , one red and one green presented each on one side of a third dot . The color of the central dot varied randomly from trial to trial . It was either red or green . Each dot had a radius of 44 pixels and the distance between the central dot and the peripheral dots was of 134 pixels , viewing distance was 50 cm . An auditory signal of 50 ms duration indicated the beginning of the task ( Go signal ) . Its pitch varied randomly on each trial . It was either high ( 1500 Hz ) or low ( 600 Hz ) . In the two experiments we analyzed the pooled data of all subjects excluding trials in which reaction time ( RT ) deviated more than 4 standard deviations from the mean . We fitted the Hybrid model , the single independent stage model [19] , the diffusion model [13] , the Ornstein-Uhlenbeck process [20] and Stanford et al . 's model [10] by maximizing the likelihood of the rPT distribution for each subject independently . We used the same procedure in the second experiment to fit the Hybrid model with the internal bias as an additional parameter . In the second experiment , we inverted the matrix to find the weights of the renormalized distribution of rPT for congruent and incongruent trials as a mixture of the renormalized distribution of rPT for “more frequent” and “less frequent” trials . We set α = . 65 , the true bias . In the Hybrid model , the decision process is split into two stages . During the first stage , internal and external variables accumulate without interacting until they reach their transition threshold and enter the common part . In the common part , internal and external variables interact in a way that depends on the congruency of internal and external variables . The label ( i . e . , left or right ) of the internal variable is drawn randomly from a Rademacher distribution . The initial accumulation rate of the internal variable during the first stage is drawn from a lognormal distribution of mean and standard deviation , i . e . . Similarly , the accumulation rate of the external variable during the first stage is drawn from a lognormal distribution of mean and standard deviation , i . e . . The internal variable starts accumulating as soon as the go Key has been pressed at time whereas the external variable starts accumulating as soon as the cues have appeared on the screen , i . e . at time ( we consider no afferent delay for sake of simplicity ) . We can thus write the value of the variables at each time during the first phase: and ( for ) . Once both variables have reached the transition threshold at time , where is the first to verify and , the accumulation rate of the internal variable is influenced by the external variable . More precisely , we have a distinction between congruent and incongruent trials: The transition between and is continuous and follows the differential equation: The first variable to reach the common threshold determines the side chosen , and the time at which it crosses the response threshold plus the execution delay ( ) gives the reaction time . We compared our Hybrid model with three other models based on the rPT distribution obtained for each participant in the first experiment . We used simulations to compute the likelihood of each participant distribution of rPT under each model and sets of parameters ( see Text S1 ) . The best set of parameters was obtained by maximizing the likelihood ( L ) for each model , then we computed the AIC and BIC for each subjects using the following formulae for a model using p parameters , and n observed rPTs:
|
The topic of our study is the interaction between intentional and externally-driven actions . The contemporary literature on motor control in primates clearly distinguishes two neural mechanisms for these two types of actions . We believe that this distinction is artefactual and comes in part from the fact that intentional and externally-driven actions have been studied by different groups of researchers , using different methodologies . In real life however , voluntary planned actions such as making a cup of tea are often interrupted by other actions in response to the outside world . In the present study , we specifically investigate the interaction between intentional and externally-driven actions . We asked participants to anticipate their response to the delayed appearance of a target . We find strong interactions between the prepared action and the target-triggered response: perceptual decisions are quicker when the two actions are congruent . To account for our behavioral results , we propose a computational model that is based on two stages , the first in which external and internal evidence accumulate separately , and the second in which internal processes are modulated by externally-driven ones . This model allows us to establish that intentional and externally-driven actions really interact only in half of the decision making process .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"psychology",
"cognitive",
"psychology",
"social",
"and",
"behavioral",
"sciences",
"mental",
"health",
"computational",
"neuroscience",
"biology",
"neuroscience"
] |
2013
|
Dual Process for Intentional and Reactive Decisions
|
Cell differentiation is remarkably stable but can be reversed by somatic cell nuclear transfer , cell fusion , and iPS . Nuclear transfer to amphibian oocytes provides a special opportunity to test transcriptional reprogramming without cell division . We show here that , after nuclear transfer to amphibian oocytes , mitotic chromatin is reprogrammed up to 100 times faster than interphase nuclei . We find that , as cells traverse mitosis , their genes pass through a temporary phase of unusually high responsiveness to oocyte reprogramming factors ( mitotic advantage ) . Mitotic advantage is not explained by nuclear penetration , DNA modifications , histone acetylation , phosphorylation , methylation , nor by salt soluble chromosomal proteins . Our results suggest that histone H2A deubiquitination may account , at least in part , for the acquisition of mitotic advantage . They support the general principle that a temporary access of cytoplasmic factors to genes during mitosis may facilitate somatic cell nuclear reprogramming and the acquisition of new cell fates in normal development .
Normal development , as well as nearly all cases of experimentally induced changes in gene transcription , is accompanied by cell division . It is therefore hard to distinguish those molecular events which prepare cells for , or engage them in , mitosis from those that are required specifically for transcriptional reprogramming . The relationship between the cell cycle and cell fate decisions has for a long time attracted interest [1] . Transition through mitosis is a time when many transcription factors are displaced from chromatin , potentially permitting new transcription factors to occupy chromatin sites on mitotic exit and so direct a postmitotic cell fate change [2]–[5] . Mitotic remodelling has been shown to be of great importance for the efficient replication of erythrocyte nuclei by Xenopus egg extracts [6] , [7] . For new transcription , cell division seems to be needed in some cases [8] , [9] but not in others [10] , [11] . Here we have used nuclear transfer to amphibian oocytes to compare directly the ability of mitotic chromatin or interphase nuclei to be reprogrammed in the absence of cell division . Germinal vesicle ( GV ) stage oocytes do not replicate or divide . They therefore provide an opportunity to test whether the cell cycle phase of donor nuclei affects the efficiency of nuclear reprogramming as judged by active transcription of previously silenced genes [12] . To our surprise , we found that a mitotic state of donor nuclei dramatically increases the efficiency of activating certain quiescent pluripotency genes in these nuclei . Our results support an idea that a brief period during mitosis facilitates an exchange of gene regulatory factors on chromatin and that this could be an important mechanism to help cells embarking on new cell lineages during normal development .
Permeabilized mouse C2C12 cells , a cultured myoblast cell line which we have used extensively in our oocyte nuclear transfer experiments , were arrested at specific stages of the cell cycle ( Figure S1a ) and were injected into the GV of oocytes ( Figure 1a ) . The DNA content of these donor cell populations ( Figure 1b ) confirmed cell cycle arrest in each of the cell cycle stages . The transcriptional reactivation of three silent genes quiescent in C2C12 cells ( Nanog , Oct4 , and Sox2 ) was assessed by RT-qPCR 38 h after nuclear transplantation ( Figure 1c ) . Nuclei at a late stage of the cell cycle ( M ) show greatly enhanced transcription of each of the genes when compared to unsynchronized nuclei ( predominantly G1 and S ) , whereas an already active gene ( c-jun ) shows little increase in transcript level . Particularly impressive is the 100-fold enhancement in Sox 2 expression from mitotic donor nuclei when compared to interphase donor nuclei ( Figure 1c ) . In over 50 experiments , donor cells arrested in mitosis or in late G2 always generated more Sox2 transcripts from reactivated genes at 25–48 h after injection to oocytes than unsynchronized donor cells . This difference ranged from a few fold to over 100-fold and is much affected by the exact duration of nocodazole treatment . Sox2 is a gene that is more widely expressed than most others , notably in early embryos , in most stem cells , and in the nervous system [13] . To test whether this result is a peculiarity of this donor cell type ( C2C12 myoblasts ) or is a nonspecific effect of nocodazole , we repeated these experiments with 10T1/2 donor nuclei ( Figure S1b ) or prepared mitotic C2C12 donor cells without any inhibitors by a shake-off procedure ( Figure S1c ) . In both cases , enhanced transcription from mitotic donors was observed , although the magnitude of mitotic advantage was lower ( particularly in the case of the shake-off samples , many cells of which appeared to be apoptotic by visual inspection ) . Mitotic donor nuclei were also prepared using another cell synchronization agent ( Taxol ) , and the mitotic advantage was again seen ( Figure S1d ) . When G1/G0 cells were exposed to nocodazole for the same period of time as used to prepare mitotic cells , no enhancement of transcription of the genes was observed ( Figure S1e ) . These results indicate that the observed mitotic advantage is not due to a nonspecific activity of nocodazole nor to a peculiarity of one line of cells ( C2C12 ) . To ask if this mitotic advantage applies more widely in the genome than to the pluripotency genes so far tested , we compared by RNAseq the genes transcribed in injected oocytes by interphase nuclei or mitotic chromatin . We focussed our analysis on genes that were found to be consistently expressed by interphase nuclei after nuclear transfer . One experiment indicated that 617 genes were transcribed in oocytes at least 2-fold more in mitotic nuclei compared to interphase nuclei . Of these mitotically up-regulated genes , Sox2 was 4-fold more transcribed than in interphase nuclear transfers , and over half of the 617 genes were more strongly transcribed than Sox2 . The list of these genes is in Table S1 . The enhanced reprogramming from mitotic donor material could be due to an increased rate or to a greater eventual level of reprogramming . To distinguish these ideas and to measure the rate of reprogramming , we measured the incorporation of GFP-tagged histone B4 ( an early marker of oocyte reprogramming ) [14] and the association of Cherry-labelled histone H2B by live imaging of mixed populations of mitotic and interphase donor cells after injection into oocytes ( see Figure S2a for design ) . Mitotic donor material becomes very rapidly marked with both histone B4 and histone H2B , whereas interphase donor nuclei show a lag in the association of both and particularly of H2B ( Figure 2a ) . In support of a difference in the rate of reprogramming , we find that oocyte-derived TBP2 marks the transplanted mitotic cells more strongly than interphase donor cells ( Figure S2b; compare white mitotic with yellow interphase arrows ) . We then asked if there is a more rapid association and activation of RNA polymerase II with mitotic chromatin . We used immunostaining for the elongating form of RNA polymerase II on a mixed population of mitotic and interphase nuclei injected into oocyte GVs . Mitotic donor material is clearly marked with elongating Pol II before interphase donor material ( Figure 2b , compare panels ii and iv for pol II ) . In view of this difference between the two nuclear types in the onset of global pol II transcription after nuclear transfer , we asked whether reprogrammed genes are activated at a different rate in mitotic donor cells compared to interphase cells or if the magnitude of activation is greater . A time course of reprogramming from oocytes injected with either interphase or mitotic donor cells was assessed by RT-qPCR and revealed that genes from mitotic donor cells are activated more rapidly than the same genes from interphase cells ( Figure 2c ) ; the accumulation of transcripts reached by 63 h is similar . We conclude that the difference in reprogramming between interphase and mitotic donor material giving this mitotic advantage reflects the rate of reprogramming rather than the eventual magnitude of transcript generation from these two types of nucleus . The most obvious explanation for this mitotic advantage is the absence of a nuclear envelope in the mitotic karyoplasts . We have quantitated this difference in membrane permeability by time course imaging a mixture of injected interphase nuclei and mitotic karyoplasts . We carried out a “double permeabilization , ” in which both the cell and nuclear membranes , of interphase or mitotic donor cells , were permeabilized as illustrated in the scheme in Figure 3a . We then compared the rate of oocyte factor uptake with the rate of reprogramming by RT-qPCR . A difference in the amount of B4 and H2B uptake is indeed seen after plasma permeabilization with digitonin ( Figure 3b ) but is no longer seen after double permeabilization of the nuclear envelope with Triton ( Figure 3c ) . Nevertheless , the mitotic difference between interphase and mitotic chromatin does persist in respect of the transcriptional reprogramming of silenced genes ( Figure 3d ) . We have confirmed this conclusion using permeabilization by different reagents . Streptolysin 0 ( SLO ) permeabilizes the plasma membrane but not the nuclear membrane; SLO and Lysolecithin ( LL ) together permeabilize the plasma membrane and nuclear membrane [15] . Permeabilization was tested using different sizes of dextran ( Figure S3a ) . We then compared transcription from transplanted nuclei , comparing those treated with SLO alone and those treated with SLO and LL . The transcription ratio following these two procedures shows no advantage when the nuclear envelope is permeabilized ( Figure S3b ) . We conclude that the presence of an intact interphase nuclear envelope does not explain the mitotic advantage . Because the difference in reprogramming rate between interphase and mitotic donor cells is maintained after extensive permeabilization of the interphase nuclear membrane , we asked if the source of the difference lies in the chromatin of the two donor cell preparations . To answer this , we mildly sonicated both interphase and mitotic donor cell preparations to give fragments of chromatin of similar sizes ( Figure 4a and b ) , injected these preparations in parallel with a permeabilized cell preparation into oocyte GVs , and assessed gene reactivation by RT-qPCR ( Figure 4c ) . It is clear that the difference in the rate of gene reactivation from interphase and mitotic nuclei is maintained when the injected material is sonicated chromatin as opposed to whole nuclei . This suggests that the “mitotic advantage” is present in the chromatin of mitotic cells . This result also confirms that the difference between interphase and mitotic donor cells is not due to the interphase nuclear membrane , nor to any other aspect of nuclear organization that is eliminated by sonication . The difference between interphase and mitotic reprogramming is , however , abolished when genomic DNA prepared from donor nuclei is injected into oocyte GVs ( Figure 4d ) ; this excludes differences at the DNA level ( sequence and DNA methylation for example ) as possible sources of the difference in reprogramming between interphase and mitotic samples . The possibility of DNA methylation accounting for the mitotic effect was further excluded by bisulphite analysis of specific loci on mitotic and interphase DNAs , as this revealed no mitosis-specific differences ( Figure 4e ) . These two results indicate that whatever accounts for the difference between mitotic and interphase donor cells is not present at the level of genomic DNA itself but is in non-DNA components of chromatin . As the mitotic advantage is likely to be due either to the loss or gain of chromatin binding factors , we removed most of these from our donor suspension of interphase nuclei by incubating such nuclei in a high-salt Triton buffer . We thereby tested whether a loss of chromatin binding factors at mitotic entry could remove the mitotic advantage . We also largely depleted chromatin binding factors from permeabilized mitotic cells and thus removed many chromatin factors that may be gained by cells entering mitosis . The depletion of chromatin binding factors was achieved with 300 mM salt and Triton , which removed most nonhistone DNA binding factors . A scheme of the cell preparation and examples of the proteins removed are shown in Figure 5a and 5c . It can be seen that the great majority of the nonhistone chromosomal proteins that have been tested and that normally exist in interphase nuclei have been removed from mitotic chromatin by 300 mM salt and Triton . Nevertheless interphase nuclei depleted of salt soluble nuclear protein ( 300 mM sample ) do not acquire the same reprogramming responsiveness as mitotic donor material ( Figure 5b ) . Likewise , extensive protein removal from mitotic donor material ( Figure 5c ) before nuclear transplantation does not abolish the mitotic advantage ( mitotic 300 mM; Figure 5b ) , indicating that the acquisition of chromosomal proteins by mitotic nuclei does not account for this advantage . Independently of salt release experiments , we tested topoisomerase II whose activity increases from S phase to the end of G2 . The inhibition of topoisomerase II and of its adaptor molecules 14-3-3z and H3S10ph by inhibitors , inhibitory peptides , and antibody injection in transplanted mitotic nuclei did not reduce the mitotic advantage . We also found that salt release removes topoisomerase from mitotic chromatin , as in experiments shown in Figure 5c , but does not change the mitotic advantage . We conclude that a loss of salt-soluble chromatin binding factors does not account for the mitotic advantage . It is likely therefore that the source ( s ) of the difference is either a non-salt-soluble factor ( gained or lost at mitotic entry ) , a covalent modification of chromatin , or the spatial arrangement of nucleosomes . We next considered covalent histone modifications that may be lost or gained on mitotic chromatin compared to interphase chromatin . A large number of histone modifications are associated with mitotic entry [16] , [17] , as well as changes in nucleosome positioning and in chromatin compaction . We first tested the most striking changes involving global histone deacetylation , phosphorylation , and some small increases in histone H3 lysine 4 and 9 methylation that have been seen on mitotic chromatin [16] , [18] , [19] . Histone deacetylation in mitotic cells is successfully inhibited during mitotic synchronization by the histone deacetylase inhibitor TSA ( Figure 6a ) . Histone phosphorylation in mitotic cells is inhibited by the Aurora B/JAK inhibitor AT9823 ( Figure 6b ) . Nevertheless , the mitotic advantage persists after both of these treatments ( Figure 6c and 6d ) . Similarly , the removal of mitotic histone phosphorylation from the Sox2 gene by protein phosphatase treatment of mitotic and interphase donor cells before nuclear transplantation also failed to abrogate the mitotic effect ( Figure S4a and b ) . A small ( 2-fold ) local increase in Sox2 locus histone methylation at mitosis ( Figure S4c and d ) , seen by ChIP in mitotic chromatin , is eliminated by the methylation of MTA ( not shown ) , but the mitotic advantage is retained ( Figure 6e ) . In normal cells , histone ubiquitination ( primarily H2AK119Ub and H2BK120Ub ) is dramatically reduced at mitotic entry [16] . H2AK119Ub is associated with transcriptional repression [20] . Thus , a reduction in H2A ubiquitination at mitosis is an attractive candidate to explain the enhanced reprogramming of mitotic chromatin . We first tested the effect of increasing ubiquitination of mitotic chromatin , by preparing nuclei for injection in the presence of iodoacetimide ( IAA ) , which inhibits deubiquitinases ( Figure 7a ) [21] . Under normal conditions , interphase chromatin is at least five times more globally ubiquitinated than mitotic chromatin ( Figure 7a , b ) . The inhibition of mitotic deubiquitination by IAA increases the ubiquitin level in mitotic chromatin , so that it is nearly equal to that of interphase nuclei ( Figure 7c ) . When tested for transcription in oocytes , hyperubiquitinated mitotic chromatin by IAA does not show an advantage over interphase chromatin ( Figure 7d ) , in accord with the idea that the deubiquitinated state of normal mitotic chromatin could account for its special transcriptional advantage . As further support for this idea , we tried to remove ubiquitin from interphase nuclei with a recombinant deubiquitinase ( Ubp-M ) , and then tested the effect of this by oocyte injection followed by RT-qPCR . Treatment of interphase nuclei with Ubp-M removes histone ubiquitination ( Figure 7g ) . However , unexpectedly , we see that the removal of histone ubiquitination by Ubp-M does not significantly enhance the reprogramming of interphase nuclei , so that they behave the same , in this respect , as mitotic chromatin ( Figure 7h ) . This suggested that deubiquitination itself is not sufficient to confer mitotic advantage . We hypothesized that H2A deubiquitination is a required step in a series of chromatin remodelling events that eventually lead to mitotic advantage . We therefore chose to reduce the ubiquitinated state of interphase nuclei in living cells in order to allow events downstream of ubiquitin-depleted chromatin to take place . The inhibitor MG-132 is thought to lower histone ubiquitination by reducing the pool of free ubiquitin through inhibition of the proteasome . MG-132 treatment of interphase nuclei before injection to oocytes ( Figure 7a ) gave a partial but significant reduction in H2A ubiquitination on Sox2 ( Figure 7e compared to 7b ) ; it also resulted in a substantial enhancement of oocyte-induced transcription from interphase , but not mitotic chromatin ( Figure 7f ) . The increase in ubiquitination of transplanted mitotic nuclei , coupled with the removal of the mitotic advantage , shows that chromatin ubiquitination contributes to the mitotic advantage . It may , however , not be sufficient to explain the whole phenomenon , because we do not achieve a complete mitotic advantage in interphase nuclei by deubiquitination .
Our results show a substantial effect of the cell cycle stage of donor nuclei in nuclear transfer experiments . The reason why this has not been seen before in some of the older nuclear transfer experiments with eggs is probably for several reasons . One is that the somatic nuclei used as donors were not able to be well synchronized [22]–[24] . Another is that tests have involved the normality of development , rather than gene activity . Third , and most importantly , tests have been carried out on cell dividing eggs , whereas our work has tested gene transcription in the complete absence of DNA replication or cell division . The more recent results of [6] are in agreement with the work described here . The success of the first mammal cloning work was attributed in part to the use of donor cells in G0 [25] , [26]; this result was not , however , found by others [27]–[30] . Lemaitre et al . [6] have described a dramatic effect of mitosis on the efficiency of DNA replication by egg cytoplasm , but transcription could not be tested in their extract experiments . Egli et al . [3] have proposed an important role for mitosis in permitting chromosomal protein exchange in mouse nuclear transfer experiments but not necessarily on gene transcription . Our results are therefore in accord with previous work , but reveal a specific effect of mitosis on gene transcription . An important component responsible for the acquisition of mitotic advantage appears to be the removal of ubiquitin from histone H2A or H2B in mitotic chromatin . Ubiquitination of histones is primarily monoubiquitination , and we assume that this is the modification involved in mitotic advantage . H2A ubiquitination on lysine 119 is associated with transcriptional repression , particularly of lineage-specifying genes in ES cells [31] , possibly through its association with members of the polycomb repressive proteins [32] . It could be envisaged that the deubiquitination of chromatin seen at mitotic entry permits this mitotic advantage by removing this inhibitory mark or the associated binding proteins [33] . In keeping with this idea , we have been able to partly simulate the mitotic effect by biochemically removing histone ubiquitination in interphase donor cells . What could be the significance of the mitotic advantage identified here ? In our experiments , the mitotic advantage takes place during the early stages of transcriptional activation , and is no longer seen after 2 d . In normal dividing cells , mitosis lasts for only a few hours . We therefore think that mitosis is a time when cells can most easily change their chromatin state , exchange transcription factors , and embark on a new lineage . When a cell has adopted a new fate , its daughter cells will usually follow the same lineage , unless an exchange of nuclear components takes place . The acceleration of postmitotic transcriptional activation [34] may be an associated phenomenon .
Cells were cultured in DMEM ( D5796 , Sigma; E15-810 , PAA; 41965-062 , Invitrogen ) with 10% FCS ( 10270106 , Invitrogen ) , 100 units/ml Penicillin-Streptomycin ( 15140-122 , Invitrogen ) , and 0 . 25 µg/ml Fungizone ( 15240-096 , Invitrogen ) . Inhibitors used for various experiments include the following: 5′-Deoxy-5′- ( methylthio ) adenosine ( MTA ) ( D5011 , Sigma ) used at 1 µg/ml , Aphidicolin ( A0781 , Sigma ) at 1 µg/ml , AT9823 ( gift from M . Dawson ) used at 100 nM; ICRF-90 ( I4659 , Sigma ) at 1 µg/ml , iodoacetamide ( IAA ) made freshly and used at 10 µM , MG-132 ( Sigma ) used at 4 µM , Nocodazole ( M1404 , Sigma ) at 75–100 nM; Taxol ( T7402 , Sigma ) at 1 µM , Thymidine ( T1895 , Sigma ) , and Trichostatin A ( T8552 , Sigma ) at 1 µg/ml . Cell synchronization was achieved according to the scheme in Figure S1A . In general , media containing the desired inhibitor were applied to unsynchronized cells a day after seeding for 16–20 h . For mitotic cells , seeded cells were initially arrested in 2 mM thymidine for 16–24 h , washed 3× in PBS , released into fresh media for 6–12 h , and then media replaced with Nocodazole or Taxol containing media for 10–16 h , after which rounded cells were detached by “shake-off” and the culture media harvested for the mitotic cell fraction . G1 arrest was achieved by Serum starvation for 72 h . Cells in suspension ( either from mitotic shake-off or trypsinization of adherent cells ) were washed twice in PBS , transferred to SuNaSP , and permeabilized with Digitonin ( 40–100 µg/ml ) for 3 min on ice . The reaction was stopped by addition of and excess of SuNaSP-BSA and the nuclei concentrated to an appropriate volume for GV transfer [12] . Nuclear transplantation to oocyte GVs was performed as described in [12] . The following media were used for permeabilization: SuNaSP , 0 . 25 M Sucrose , 75 mM NaCl , 0 . 5 mM Spermidine , 0 . 15 mM Spermine; SuNaSP-BSA , SuNaSP with 3% ( w/v ) BSA; and HPRicLS , Final [1×] Hepes 20 mM , KCl 75 mM , MgCl2 1 . 5 mM . Cells were permeabilized with Digitonin and incubated for 15 min in prebuffer ( 20 mM Hepes , 75 mM KCl , 1 . 5 mM MgCl2 , 25 mg/ml Gelatin , 60 mg/ml BSA ) and washed twice into permeabilization buffer ( 20 mM Hepes , 75 or 300 mM KCl , 1 . 5 mM MgCl2 , 0 . 2% TritonX100 , 12 mg/ml Gelatin , 30 mg/ml BSA ) . Cells were then extensively washed ( 20 mM Hepes , 75 mM KCl , 1 . 5 mM MgCl2 ) and resuspended in a suitable volume of SuNaSP-BSA for nuclear transplantation . Cells were permeabilized in Digitonin , incubated in “prebuffer , ” washed into permeabilization buffer ( 75 mM salt ) , washed in SuNaSP , and then transferred into a suitable reaction buffer with or without recombinant enzyme . Dephosphorylation was performed with Protein Phosphatase I ( NEB , P0754S ) in HPRicLS ( 20 mM Hepes , 75 mM KCl , 1 . 5 mM MgCl2 ) and Deubiquitination performed using recombinant enzyme prepared from insect cells ( as described in [33] ) in buffer ( 50 mM NaCl , 50 mM Tri-Cl PH 8 . 0 , 1 mM DTT , 1× Complete EDTA-free Protease inhibitor [Roche] ) . After enzyme treatment , cells were washed in HPRicLS and SuNaSP and resuspended in a suitable volume of SuNaSP-BSA . PBS washed cell suspensions were ( fixed in ethanol and stained with 50 µg/ml propidium iodide ) . DNA content analyses were then performed on a FACSCalibur cytometer ( BD Bioscience ) . The following antibodies were used: αAurura B ( AIM1 ) ( Cell Signalling , 3094 ) , αBmi1 ( Cell Signalling , 6964 ) , αBRD4 ( Cell Signalling , 12183 ) , αCTCF ( Cell Signalling , 3418 ) , αCyclinB ( Cell Signalling , 4138 ) , αDNMT1 ( Abcam , ab92453 ) , αH2AK119Ub ( Cell Signalling , 8240 ) , αH2BK120Ub ( Cell Signalling , 5546 ) , αH3 ( Abcam , ab1791 and Cell Signalling , 4620 ) , αH3K4me3 ( Abcam , ab8580 ) , αH3K9ac ( Cell Signalling , 9649 ) , αH3K9me2/3 ( Cell Signalling , 5327 ) , αH3S10ph ( Sigma , H 0412 ) , αH4 ( Abcam , ab31830 ) , αHP1α ( Cell Signaling , 2616 ) , αphosphoSer2-PolII ( Covance , MMS-129R ) , αRunx2 ( Cell Signalling , 8486 ) , and αTBP ( Abcam , ab62125 ) . RT-qPCR was performed as described in [11] . Unless otherwise stated , results are normalized to VegT ( correcting for intrasample RNA extraction variation ) and G3PDH ( correcting for nuclear number differences between injected oocyte samples ) . Error bars indicate SME or standard deviation , and significance is determined by unpaired Student t test , with p<0 . 05 being considered significant . All experiments presented were single experiments representative of at least three experimental repeats unless otherwise noted . Genomic DNA was prepared using DNeasy blood and tissue kits ( 69504 , Qiagen ) , bisulphite conversion was performed using EpiTect Bisulfite Kit ( 59104 , Qiagen ) , and primer sequences for DNA preparation were designed using Qiagen . Pyrosequencing was performed on a Qiagen Pyromark Q96 ID using PyroMark Gold Q96 Reagents ( 972804 , Qiagen ) and PyroMark PCR Kit ( 978705 , Qiagen ) , as per the manufacturer's recommendations .
|
Cells are dividing very actively at a time in development when new gene expression and new cell lineages arise . At mitosis , most transcription factors are temporarily displaced from chromosomes . We show that , after transplantation to oocytes , somatic cell nuclei that have been synchronized in mitosis can be reprogrammed to pluripotency gene expression up to 100 times faster than interphase nuclei . We find that , as cells traverse mitosis , their genes pass through a temporary phase of unusually high responsiveness to oocyte reprogramming factors ( mitotic advantage ) . Many other genes in the genome have also shown a mitotic advantage , which affects the rate rather than the final level of transcriptional enhancement . This is attributable to a chromatin state rather than to more rapid passage of reprogramming factors through the nuclear membrane . Histone H2A deubiquitination at mitosis is required for the acquisition of mitotic advantage . Our results support the general principle that a temporary access of cytoplasmic factors to genes during mitosis facilitates somatic cell nuclear reprogramming and the acquisition of new cell fates in normal development .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology",
"and",
"life",
"sciences"
] |
2014
|
Mitosis Gives a Brief Window of Opportunity for a Change in Gene Transcription
|
Adaptive collective systems are common in biology and beyond . Typically , such systems require a task allocation algorithm: a mechanism or rule-set by which individuals select particular roles . Here we study the performance of such task allocation mechanisms measured in terms of the time for individuals to allocate to tasks . We ask: ( 1 ) Is task allocation fundamentally difficult , and thus costly ? ( 2 ) Does the performance of task allocation mechanisms depend on the number of individuals ? And ( 3 ) what other parameters may affect their efficiency ? We use techniques from distributed computing theory to develop a model of a social insect colony , where workers have to be allocated to a set of tasks; however , our model is generalizable to other systems . We show , first , that the ability of workers to quickly assess demand for work in tasks they are not currently engaged in crucially affects whether task allocation is quickly achieved or not . This indicates that in social insect tasks such as thermoregulation , where temperature may provide a global and near instantaneous stimulus to measure the need for cooling , for example , it should be easy to match the number of workers to the need for work . In other tasks , such as nest repair , it may be impossible for workers not directly at the work site to know that this task needs more workers . We argue that this affects whether task allocation mechanisms are under strong selection . Second , we show that colony size does not affect task allocation performance under our assumptions . This implies that when effects of colony size are found , they are not inherent in the process of task allocation itself , but due to processes not modeled here , such as higher variation in task demand for smaller colonies , benefits of specialized workers , or constant overhead costs . Third , we show that the ratio of the number of available workers to the workload crucially affects performance . Thus , workers in excess of those needed to complete all tasks improve task allocation performance . This provides a potential explanation for the phenomenon that social insect colonies commonly contain inactive workers: these may be a ‘surplus’ set of workers that improves colony function by speeding up optimal allocation of workers to tasks . Overall our study shows how limitations at the individual level can affect group level outcomes , and suggests new hypotheses that can be explored empirically .
Many systems in biology and engineering , from cells to mobile networks and human societies , consist of several or many interacting units that contribute ‘work’ towards a central goal [1–6] . Each of these systems employs a ‘task allocation mechanism’ , i . e . , individual workers choose , or are allocated to , a specific part of the total workload , a task , which they then attempt to complete . The simplest such task allocation mechanism might be one where each individual picks a task randomly; another simple ( from an algorithm standpoint ) mechanism might be one where each individual is preprogrammed to always pick a defined task . For example , in a simple multicellular organism such as the alga Gonium [7] , each cell processes nutrients that it happens to encounter , and each cell is equally likely to reproduce . Conversely , a car may be made up of lots of elements that need to work together to make the car run , but these elements have no flexibility with regard to how they contribute to this goal: each part fulfills its preprogrammed and unchangeable function . However , most biological systems , and many engineered ones , do not behave according to either of these extremes . Instead , individuals have to choose how to contribute , and may use various types of information about the need for different types of work to make this choice ( note that we are using the term ‘choice’ in the sense of possessing an algorithm that leads to task selection , and do not imply free will ) . The goals of any such task allocation mechanism are to achieve efficiency and robustness of system function . For example , in a developing embryo , multiple cells have to select which organs or tissues to develop into [8] . The task allocation mechanism used has to ensure that the right cells are allocated to all necessary organs; at the same time , it has to tolerate the occasional loss of cells . Similarly , in cloud computing , the demand for different types of computation may change dynamically over time , and so might the availability of individual processors [9 , 10] . The ideal task-allocation mechanism used here again has to achieve a match of allocated processors with current needs , which likely requires repeated re-allocation . Is task allocation a difficult problem , and does it matter which algorithm is chosen ? If task allocation is an easy problem , then the match of work to workers should be close enough to the theoretical optimum that the efficiency and robustness of the evolved biological systems and designed/engineered systems are not substantially reduced . However , there is evidence from theoretical computer science that indicates that task allocation ( referred to as ‘resource allocation’ ) is difficult [11–13] in that it requires a non-negligible amount of resources ( such as time , memory , and/or communication messages ) . In particular , [12] shows that if individuals also differ in how well they can perform different types of work , then in the model they consider , task allocation is an NP-hard problem . Another line of evidence for the idea that task allocation is difficult is the number of workers in distributed systems that are in fact not allocated to any tasks [14] . In social insect colonies in particular , a large fraction of workers do not appear to work [15]; in addition , at any point in time , there is another substantial group of individuals who are thought to be actively looking for work [16] . This may indicate either that these workers are in excess of the number needed to perform tasks , or that they are result of a task allocation mechanism that either costs time ( in the form of workers looking for work ) or produces inadequate allocation ( unemployed workers that could be employed ) . Either way , this would indicate that task allocation is not an easy problem ( several other hypotheses , unrelated to task allocation , have also been proposed [15] ) . In distributed computing , extra computing devices ( in addition to the number necessary to complete the tasks ) are often used to achieve fault tolerance and increase efficiency by replicating information and computation over multiple devices [9 , 10] . Both of these phenomena might indicate that task allocation is neither effective nor fast: if task allocation were easy to achieve quickly , then there would not be a need for costly buffering . If task allocation is a difficult problem , we would expect to see complex systems employ imperfect mechanisms that lead to approximate solutions , or which sometimes fail to allocate workers to tasks correctly , or we might see additional strategies that compensate for mistakes of imperfect task allocation , or trade-offs between the resources invested and the quality of task allocation achieved . Thus , in these cases we expect the chosen task allocation mechanism to contribute significantly to system performance or biological fitness . It will not then be possible to understand the evolution of system organization , or to design an efficient and robust system , without also understanding the constraints imposed by the process of task allocation . Here we aim to contribute to an understanding of what limits flexible and robust task allocation . To do this , we develop a model of task allocation in social insect colonies . We are specifically interested first in how group size , i . e . the number of individuals that may be allocated to work , affects the difficulty of correct task allocation , and second , in the effects of having more workers available than work ( which would lead to inactive workers ) . We also discuss the effect of the number of distinct task types to which workers have to be allocated . We quantify performance of three generalized task allocation mechanisms that differ in the amount of information available to workers about the demand for work in different tasks . We are thinking of our model as representing individual insect workers making choices among such tasks as foraging or brood care . However , our model is kept general in many respects , and is thus likely to apply to many similar systems where individuals are making choices about tasks using local information . Group size is typically thought to be a central factor in determining complex system function [17]: multicellular organisms [18] , human societies and organizations [19 , 20] , and social and computer networks [21] all have been argued to develop more complexity , acquire new functionalities , and be competitively superior at larger group sizes , and all of this has also been argued for social insect colonies [22] . In many cases , although not unequivocally [19 , 22] , larger group size has been associated with more specialized , and possibly less flexible , individuals within the group; this may result from the smaller variance typically experienced by larger groups because of the ‘law of large numbers’ [23] . Larger groups may also benefit from ‘economies of scale’ when there are fixed costs that do not scale linearly with the number of individuals [24]; for example , broadcast signals reach more individuals in larger groups at the same cost [25] . Biological accounts of the evolution of larger groups , at any level of organization , typically focus on these benefits of group size [17] . In computer science , on the other hand , research has often focused on the costs of group size [13 , 26] . Generally speaking , algorithms that require interactions between individuals take much longer to execute in larger groups , because the number of possible interactions increases faster than linearly with group size ( with N2 for pairwise interactions , exponentially when any number of interactants is possible ) . Indeed this effect of group size on ‘naive’ distributed problem-solving algorithms is so great that the group size is typically equated with ‘problem size’ , and the performance of algorithms is measured mainly in terms of how strongly they depend on group size or other measures closely related to group size [13 , 27] . This makes sense if one assumes that the effect of group size will outweigh the effects of any constant factors on the performance of the algorithm , even for moderately large groups . As stated above , we are using social insect colonies as a model system to study the effect of group size on the difficulty of task allocation . Social insects such as bees , ants , wasps , and termites typically live in colonies that contain one or a few queens , who are the source of colony reproduction , and many , anywhere from a handful to millions of workers , who do not reproduce but complete all other tasks [28–30] . These tasks include foraging ( finding and collecting food ) , nest building and repair , brood care ( caring for immature individuals; Hymenopteran insects such as bees and ants spend ≈ 10 − 30% of their lifespan in an immature stage in which they cannot move and have to be cleaned , fed , defended , and kept at a tolerable temperature much like the most dependent mammals in their infant stage ) , colony defense , and various other tasks that may include thermoregulation ( such as by ventilation or heating ) , nest cleaning , undertaking ( removing dead individuals ) , etc . [15] . The need for work in these different tasks typically fluctuates in daily and seasonal patterns as well as stochastically [31] . Social insect colonies are self-organized , meaning that neither the queens nor any other workers ‘direct’ the task choices of other workers , although interactions between individuals such as communication signals and aggression may affect task selection [29 , 32] . There are more than 10000 species of ants alone , and different species of social insects may use different task allocation mechanisms . Any task allocation mechanism consists of two parts: the traits of individuals that predispose them to particular tasks , and the behavioral rules that lead them to select a particular task at a given moment ( the individual-level algorithm; [33] ) . In social insects , body size , age , physiological and nutritional status , sensory abilities , and other internal factors are thought to create variation among individuals in task preferences and skills; in addition , individual experience , interactions with other workers , spatial and hierarchical position in the colony , and random encounters with tasks will do so as well , in the short and long term [14 , 32 , 34] . In different species , some or all of these factors may play a role in task allocation , and to differing degrees . The behavioral rule set , i . e . the algorithm , by which individuals choose a task to work on in the moment , is typically thought to involve a comparison between an individual’s task preferences and the need for a particular task; this is sometimes referred to as the ‘task stimulus response threshold mechanism’ ( because workers are thought to have different thresholds at which they decide to work on a task , depending on the level of ‘task stimulus’ which communicates demand for work in the task , [35] ) . However , it is worth noting that the actual precise algorithm is seldom defined in the insect literature; e . g . ‘thresholds’ may actually be continuous probabilistic functions , and it is unclear how multiple task stimuli are evaluated ( in random order , or at the same time , and do they interact or not ) . It is also typically unclear how the factors listed above interact to produce variation in preferences across tasks or across individuals; e . g . are the preferences for different tasks independent of one another or not [36] . All of this may also vary across species . Despite this uncertainty about the precise mechanism , the fact that social insects achieve task allocation is well studied . Workers in a colony specialize to a large or small degree on different tasks , and may switch tasks as needed [37] , although this may come at additional cost [38] . Colonies are typically able to effectively compensate for worker loss ( [36] , although see [39] ) or changes in demand for different tasks [14] . However , it is also the case that inactive workers are common: at any point , often > 50% , sometimes > 70% , of the colony appear not to be performing any tasks [15] . This may be in part due to need for rest , selfish reproduction by workers [40 , 41] , or immaturity of workers [42]; but it has also been suggested that completely inactive and ‘walking’ ( without apparently getting anything done ) workers may either be looking for work and failing to find it [16] , or in fact be a surplus of workers not necessary to complete the work of the colony [14] . Inactive workers , i . e . units within a complex system that are not contributing , may also be common elsewhere , both in biology and engineering [43 , 44] . Here we examine the effect of such a buffer of apparently redundant workers on task allocation efficiency . This study aims to contribute to understanding why social insects evolved the task allocation mechanisms that they did , and , more generally , what limits effective task allocation in distributed sytems . We contribute to these aims by measuring the performance of task allocation mechanisms under different assumptions . To achieve this , we derive how quickly task allocation can be achieved using distributed computing theory methods to analyze algorithm performances . We use a generalized task allocation mechanism with three different assumptions about how individual workers can acquire information about the need for more work in specific tasks ( what we call the ‘deficit’ ) . This approach then leads us to insights about whether and how task allocation is limited by group size , the relationship of group size to the total need for work ( what we call the ‘demand’ ) , the information available to workers , the number of tasks , and how precisely the colony must match the allocation of workers to demands for work across tasks . The rest of this paper is organized as follows: in the Methods section , we describe the tools and techniques we use from distributed computing theory , together with a formal model of the task allocation system we consider; in the Results section , we mathematically derive bounds ( that is upper limits ) on the time for ants to allocate themselves to tasks in the various versions of our formal model , and also provide some intuitive explanations and numerical examples of the results; in the Discussion section , we emphasize the implications of our results for actual ant and bee species and we address some caveats and open questions .
The specific abstraction of the task allocation problem that we study involves a distributed process of allocating all workers to tasks with the goal of satisfying the demand for each task . The demand for each task can be thought of as a work-rate required to keep the task satisfied . We consider all workers to be equal in skill level and preferences . While this is an abstraction , we focus here on simply the challenge of allocating generalist workers among tasks . We do not attempt to model how the demand for a task is computed or measured empirically . Instead , we assume that as a result of workers trying to maximize the fitness of the colony , there is some optimal number of workers performing each task , and this is what the workers should attempt to match . At each time step , each worker decides what task to work on based on simple feedback from the environment informing the worker of the state of the tasks . In particular , we consider two specific types of environment feedback: ( 1 ) whether the worker is successful at its current task , and ( 2 ) which task does the work choose next . We analyze whether this general algorithm is able to successfully allocate the workers so that all tasks are satisfied , and the time for this process to terminate . In particular , we focus on upper bounds for the time to satisfy all tasks ( i . e . how long it is expected to take given the worst possible starting conditions ) as a function of colony size , the number of tasks , and the total amount of work in the presence or absence of extra workers ( beyond the minimum to satisfy all tasks ) in the colony . See S1 Text for a more detailed version of this section . Let A denote the set of workers and T denote the set of tasks . Each task i ∈ T has an integer demand di that represents the minimum number of workers required to work on task i in order to satisfy the task . Let wi denote the total number of worker units of work currently supplied to task i . Let w → and d → denote the vectors of wi and di values , respectively , for each 1 ≤ i ≤ |T| . The d → vector is static , while w → changes over time depending on the different tasks workers choose to work on . Clearly , in order for all demands to be met , there should be sufficiently many workers in the colony . We assume that there exists a real c ≥ 1 such that |A| = c ⋅ ∑i∈T di .
In this section , we give some basic definitions and results that will be used in the subsequent analyses of the convergence times for the various choice options . A task is satisfied at time r if di ≤ wi ( r ) . Let S ( r ) denote the set of satisfied tasks at time r . Let U ( r ) = T \ S ( r ) denote the set of unsatisfied tasks at time r . For each task i ∈ T and each time r , let Φi ( r ) = max{0 , ( di − wi ( r ) ) } be the deficit of task i at time r . If i ∈ U ( r ) , then Φi ( r ) = di − wi ( r ) . We define the total deficit at time r: Φ ( r ) = ∑ i ∈ T Φ i ( r ) . Define a worker to be inactive in round r , for r > 0 , if it is in state q⊥ at time r − 1 or if it receives 0 from success in round r . In other words , a worker is inactive if it is not working on any task , or if it unsuccessful at the current task it is working on . For a full list of the parameters used in the model and analysis , see Table 2 . Based on the basic properties of the success and choice components , we can establish the following facts: Next , we analyze the three variations of the choice component . In this section , we consider the first option for the choice component , where in each round choice returns a task i with probability 1/|T| . This section includes only proof overviews and approximate running times . For detailed proofs of the results in this section , refer to S2 Text . One of the main results for this option of the choice component states that for any success probability 1 − δ that we choose , the time until workers re-allocate correctly is at most O ( | T | c − 1 ) ( ln Φ ( 0 ) + ln ( 1 / δ ) ) . We can see the time is linearly proportional to the number of tasks |T| , logarithmically proportional to the total amount of work needed ( Φ ( 0 ) ) and the inverse of the failure probability , and inversely proportional to c , the ratio of the colony size to the total sum of demands of tasks . Theorem 1 . For any δ , 0 < δ < 1 , with probability at least 1 − δ , all tasks are satisfied by time O ( | T | c − 1 ) ( ln Φ ( 0 ) + ln ( 1 / δ ) ) . Proof Idea: We know that the number of inactive workers in round r + 1 is at least c ⋅ Φ ( r ) ( by fact 4 ) . By the definition of choice in this section , each inactive worker starts working on each task i with probability 1/|T| . Therefore , we can show that , in each round , the expected number of new workers to join each unsatisfied task is at least c ⋅ Φ ( r ) /|T| . First , consider the case when c ≤ 2|T| and consider some time r . After some workers join task i in round r + 1 , it is not guaranteed that the entire new set of workers remains working on task i because some workers may be unsuccessful if task i does not require that many workers . Assuming c ≤ 2|T| , since the total deficit is Φ ( r ) and there are |T| tasks , we can show that in expectation the total deficit in the next round is at least c ⋅ Φ ( r ) /|T| ( which can be 0 if all tasks are satisfied ) . Therefore , in expectation , at least c ⋅ Φ ( r ) /|T| of the new workers that join tasks will remain working on them . This implies that the expected total deficit Φ ( r ) decreases by approximately c ⋅ Φ ( r ) /|T| in round r + 1 . Next , we consider the case of c > 2|T| . We can express c as a multiple of |T|: c = c′ ⋅ |T| for some c′ > 2 . We can show that in each round , the probability to satisfy each task is at least some constant , and consequently ( using fact 5 above ) , we conclude that the expected number of unsatisfied tasks and the total deficit decrease by a constant fraction in each round . Finally , we start at time 0 , when the total deficit is Φ ( 0 ) , and inductively apply the conclusions above in the cases of c ≤ 2|T| and c > 2|T| . By facts 2 and 3 , we know that both |U| and Φ are non-increasing , so we just need to analyze how fast they decrease . For the case of c ≤ 2|T| , the expected total deficit Φ ( r ) decreases by approximately c ⋅ Φ ( r ) /|T| in each round r + 1 . So it will take approximately ( |T|/c ) ln Φ ( 0 ) rounds until the total deficit decreases to 0 . To turn this into a more formal probabilistic claim , we can add approximately ln ( 1/δ ) rounds , for some 0 < δ < 1 , in order to ensure that the tasks are satisfied not only in expectation , but with probability at least 1 − δ . This trick works by applying a simple Markov bound ( see S2 Text ) . The second main result for this option of the choice component studies the time until workers re-allocate in such a way that , for any success probability 1 − δ and any fraction ϵ that we choose , a ( 1 − ϵ ) -fraction of the total work Φ ( 0 ) is satisfied with probability at least 1 − δ . The time to re-allocate in this case is at most O ( | T | c − 1 ) ( ln ( 1 / ϵ ) + ln ( 1 / δ ) ) . Similarly to the first result in this section , the time is linearly proportional to the number of tasks |T| , logarithmcally proportional to the inverse of the failure probability , and inversely proportional to c , the ratio of the colony size to the total sum of demands of tasks . However , here , we do not have a dependence on Φ ( 0 ) , but only a logarithmic dependence on 1/ϵ . Theorem 2 . For any δ and ϵ , 0 < δ , ϵ < 1 , with probability at least 1 − δ , the deficit at time O ( | T | c − 1 ) ( ln ( 1 / ϵ ) + ln ( 1 / δ ) ) is at most ϵ ⋅ Φ ( 0 ) . Proof Idea: Following the same structure as the proof above , we can also compute the number of rounds until the tasks are satisfied approximately . Suppose we only want a ( 1 − ϵ ) fraction of Φ ( 0 ) to be satisfied for 0 < ϵ < 1 . Recall that for c ≤ 2|T| , the expected total deficit Φ ( r ) decreases by approximately c ⋅ Φ ( r ) /|T| in each round r + 1 . So it will take only ( |T|/c ) ( ln ( 1/ϵ ) + ln ( 1/δ ) ) rounds to ensure this is true with probability at least 1 − δ ( again , the ln ( 1/δ ) factor is to ensure the probability guarantee ) . For the case of c > 2|T| , we proceed similarly . Recall that in this case c′ = c/|T| and the expected number of unsatisfied tasks and the total deficit decrease by a constant fraction in each round ( this constant depends on c′ ) . So , with probability at least 1 − δ , all tasks are satisfied by time approximately ( 1/c′ ) ( min{ln |T| , ln Φ ( 0 ) } + ln ( 1/δ ) ) . The reason for having a minimum is to take advantage of the smaller value between |T| and Φ ( 0 ) . And similarly , if we only want to satisfy the tasks approximately the ln Φ ( 0 ) term turns into ln ( 1/ϵ ) . In this section , we consider the second option for the choice component where in each round choice returns a task i ∈ U ( r ) with probability 1/|U ( r ) | . This section includes only proof overviews and approximate running times . For detailed proofs of the results in this section , refer to S2 Text . One of the main results for this option of the choice component states that for c ≥ 1 and any success probability 1 − δ that we choose , the time until workers re-allocate correctly is at most O ( ln Φ ( 0 ) + ln ( 1 / δ ) ) . We can see the time is logarithmically proportional to the total amount of work needed ( Φ ( 0 ) ) and the inverse of the failure probability . Since c may be extremely close to 1 , we do not get any effect of c in this result . Theorem 3 . For c ≥ 1 and for any δ , 0 < δ < 1 , with probability at least 1 − δ , all tasks are satisfied by time min { | T | , O ( ln Φ ( 0 ) + ln ( 1 / δ ) ) } . Proof Idea: Suppose c ≥ 1 and consider some time r . We can show that in round r + 1 at least one of the following happens: ( 1 ) the total deficit decreases by a constant fraction , or ( 2 ) the number of unsatisfied tasks decreases by a constant fraction . To show the first property holds , we consider tasks with a fairly high deficit , which are not likely to get satisfied in one round . We show that the number of new workers joining such tasks is enough to decrease the total deficit by a constant fraction . To show the second property ( the number of unsatisfied tasks decreases by a constant fraction ) , we focus on tasks with fairly low deficit which are likely to get satisfied within one round . We can show that these tasks are enough to decrease the total number of unsatisfied tasks by a constant fraction in one round . For showing both ( 1 ) and ( 2 ) , we first prove a bound on the probability to satisfy any given task in a single round and then use fact 5 to get a bound on the expected number of unsatisfied tasks and the expected total deficit . Finally , we start at time 0 , when the total deficit is Φ ( 0 ) and the number of unsatisfied tasks is at most |T| , and inductively apply the two results above . By facts 2 and 3 , we know that both |U| and Φ are non-increasing , so we just need to analyze how fast they decrease . If it is the case that the expected total deficit Φ ( r ) decreases by a constant factor in each round , then it will take approximately ln Φ ( 0 ) rounds until the total deficit decreases to 0 . If it is the case that the number of unsatisfied tasks decrease by a constant factor in each round , then it will take approximately ln |U ( 0 ) | rounds until the total deficit decreases to 0 . Since Φ ( 0 ) ≥ |U ( 0 ) | , we know either Φ ( 0 ) or |U ( 0 ) | will decrease to 0 in approximately 2 ln Φ ( 0 ) rounds . To turn this into a more formal probabilistic claim , we can add approximately ln ( 1/δ ) rounds , for some 0 < δ < 1 , in order to ensure that the tasks are satisfied not only in expectation , but with probability at least 1 − δ . This trick works by applying a simple Markov bound ( see S2 Text ) . The minimum in the final bound follows by fact 6 in the General Facts section . The second main result for this option of the choice component states that for c > 1 and any success probability 1 − δ that we choose , the time until workers re-allocate correctly is at most O ( 1 / ln c ) ( ln | T | + ln ( 1 / δ ) ) . Similarly to the result above , the time is logarithmically proportional to the total amount of work ( Φ ( 0 ) ) needed initially , and the inverse of the failure probability . Now , c is strictly greater than 1 , so we see that the time is also inversely proportional to the natural logarithm of c . Theorem 4 . For c > 1 and for any δ , 0 < δ < 1 , with probability at least 1 − δ , all tasks are satisfied by time min { | T | , O ( ( 1 / ln c ) ( ln | T | + ln ( 1 / δ ) ) ) } . Proof Idea: Suppose c > 1 and consider some time r . Unlike the case of c ≥ 1 , where in round r + 1 either the total deficit or the number of unsatisfied tasks decreases by a constant fraction , here we can show that the number of unsatisfied tasks decreases by at least a constant fraction in round r + 1 . We consider all tasks with a fairly low deficit , which are likely to get satisfied in a single round . The total deficit at time r is Φ ( r ) , and the total number of inactive workers in round r + 1 is at least c ⋅ Φ ( r ) . The fact that the number of inactive workers is at least a constant fraction greater than the total deficit lets us show that the expected number of low-deficit tasks is at least a constant fraction of all unsatisfied tasks . Therefore , by satisfying these low-deficit tasks the number of unsatisfied tasks decreases by a constant fraction in expectation . Again , we can show this by proving a bound on the probability to satisfy any given task and then using fact 5 . The value of that constant fraction by which the number of unsatisfied tasks decreases is what determines the dependence of the running time on 1/ln c in this case . Finally , we start at time 0 , when the total deficit is Φ ( 0 ) and the number of unsatisfied tasks is |U ( 0 ) | , and inductively apply the result above to show that the workers will re-allocate correctly within O ( ln | U ( 0 ) | + ln ( 1 / δ ) ) rounds . Note that ln |U ( 0 ) | ≤ ln |T| and ln |U ( 0 ) | ≤ Φ ( 0 ) . The minimum in the final bound follows by fact 6 in the General Facts section . We can combine the results of the two theorems in this section . Clearly , if c is extremely close to 1 , the 1/ln c term becomes very large , and in the limit the running time becomes ∞ . Therefore , we can take the minimum of the running times in the cases of c ≥ 1 and c > 1 to get the overall running time of the algorithm . Essentially , the running time is determined mostly by the case of c > 1 , except for the small range of values for c when c is very close to 1 . In this section , we consider the third option for the choice component where in each round choice returns a task i ∈ U ( r ) with probability ( di − wi ( r ) ) /Φ ( r ) . This section includes only proof overviews and approximate running times . For detailed proofs of the results in this section , refer to S2 Text . One of the main results for this option of the choice component states that for c ≥ 1 and any success probability 1 − δ that we choose , the time until workers re-allocate correctly is at most O ( ln Φ ( 0 ) + ln ( 1 / δ ) ) . We can see the time is logarithmically proportional to the total amount of work needed ( Φ ( 0 ) ) and the inverse of the failure probability . Since c may be extremely close to 1 , we do not get any effect of c in this result . Theorem 5 . For c ≥ 1 and for any δ , 0 < δ < 1 , with probability at least 1 − δ , all tasks are satisfied by time min { | T | , O ( log Φ ( 0 ) + log ( 1 / δ ) ) } . Proof Idea: Since an inactive worker starts working on a task i with probability ( di − wi ( r ) ) /Φ ( r ) , and since there are at least Φ ( r ) inactive workers in round r + 1 , the expected number of new workers to join task i in round r + 1 is at least a constant fraction of di − wi ( r ) , which is exactly the deficit of the task at time r . We can show that each task is satisfied in round r + 1 with probability 1/2 , and so , by fact 5 the total number of unsatisfied tasks and the total deficit decreases by half in expectation . Finally , we start at time 0 , when the total deficit is Φ ( 0 ) and inductively apply the observation above to show that the workers will re-allocate correctly in approximately logΦ ( 0 ) rounds . The minimum in the final bound follows by fact 6 in the General Facts section . The second main result for this option of the choice component states that for c > 1 and any success probability 1 − δ that we choose , the time until workers re-allocate correctly is at most O ( 1 / c ) ( ln Φ ( 0 ) + ln ( 1 / δ ) ) . Similarly to the result above , the time is logarithmically proportional to the total amount of work needed ( Φ ( 0 ) ) and the inverse of the failure probability . Now , c is strictly greater than 1 , so we see that the time is also inversely proportional to the natural logarithm of c . Theorem 6 . For c > 1 and for any δ , 0 < δ < 1 , with probability at least 1 − δ , all tasks are satisfied by time min { | T | , O ( 1 / c ) O ( ln Φ ( 0 ) + ln ( 1 / δ ) ) } . Proof Idea: For the case of c > 1 , similarly to the case of c ≥ 1 , we show that each task is satisfied with a constant probability , so the number of unsatisfied tasks and the total deficit decrease by a constant fraction in each round . The value of that constant fraction is what let us show that the running time depends on 1/c . The minimum in the final bound follows by fact 6 in the General Facts section . We can combine the results the two theorems in this section . Clearly , if c is extremely close to 1 , the 1/c term becomes very large , and in the limit the running time becomes ∞ . Therefore , we can take the minimum of the running times in the cases of c ≥ 1 and c > 1 . For the various options for the choice feedback component ( keeping the success component the same ) , we study the time to correctly re-allocate all workers: the number of steps workers need to take until the demands of all tasks are satisfied or over-satisfied . In particular , we show three types of results , which differ in precisely what conditions are set on this performance measure ( rows in Table 3 ) . First , we consider the case where the demand D has to be fully satisfied with a high probability ( 1 − δ ) . For this case , in options ( 2 ) and ( 3 ) , we see that if the number of task types ( |T| ) is small , the time to allocation only depends on this parameter ( see also Table 4 ) . If the number of task types is high , we see a positive ( logarithmic ) dependence of the time to correctly re-allocate all workers on the deficit across all tasks ( i . e . the value of Φ ) . That is , correct allocation takes longer if more workers have to be re-allocated; this relationship is not linear but saturates over time . In the case of option ( 1 ) ( where workers can only check for demand in different tasks sequentially rather than instantaneously ) , we also see a linear positive dependence on the number of tasks |T| . Finally , as the workers-to-work-ratio ( c ) increases , the time to re-allocate all workers decreases: this means that if there are ‘extra workers’ ( workers in excess of the total demand for work ) , task allocation becomes faster . In options ( 1 ) and ( 3 ) , that dependence is approximately 1/c , and in option ( 2 ) , the dependence is slightly weaker: 1/ln c ( Fig 1 ) . However , note that extra ants do not contribute towards a faster task allocation until c is large enough ( approximately until c ≥ e ) . Second , we studied the time until the demand D in different tasks is satisfied approximately ( to within a ( 1 − ϵ ) fraction ) rather than exactly as above ( but still with high probability of 1 − δ ) . In general , the effect of different parameters on performance is similar to the case where task demands are satisfied exactly . However , we show that in this case , for all options of choice , surprisingly , the time to re-allocate all workers does not depend on the total deficit ( Φ ) at all . Instead , it depends on the value of ϵ . In particular , the smaller ϵ gets , the more accurately we need to re-allocate all workers , leading to a longer time to do so , until the same time as for the exact case is reached ( as in the first row in Table 4 ) . The results in both cases ( exact and approximate matching of task demands ) are the same for ϵ = 1/Φ . This implies that for very large Φ , ϵ needs to be very small in order to have equal values in the two rows . Approximate task allocation is achieved faster than precisely accurate task allocation when Φ > 1/ϵ . Finally , for the third option of the choice component , we also study the time to re-allocate all workers under some noise in the success and choice components . In particular , we assume the success component can make a limited number of ‘mistakes’ ( at most z flipped bits from 0 to 1 and vice versa ) and the choice component may return a task with a probability slightly larger or smaller than we require in option ( 3 ) ( change the probability of a task being suggested to a worker by at most a factor of 1 − y ) . We show that the best the workers can do in re-allocating is to satisfy all but z units of work , and the time to reach such a re-allocation increases as the range of the probabilities of choice increases . Here , we choose some sample values for the parameters in the model and calculate numerical results ( Table 4 and Fig 2 ) . The expressions used to generate these values roughly correspond to the first two rows of the table in Table 3 , with the difference that here the values are exact upper bounds and not asymptotic ( big-oh ) notation ( see S2 Text for how they are calculated ) . The most obvious pattern here is that task allocation takes a lot more rounds under option ( 1 ) ( workers are not able to assess quickly which tasks need more work ) than under options ( 2 ) and ( 3 ) for choice . Is task allocation then a ‘difficult’ problem that requires a significant amount of time ? This depends on how long , in real time , a ‘round’ is . If workers require time on the order of minutes to choose a task , attempt to perform work in it , and assess whether they have successfully contributed to the colony with this work , then the results for option ( 1 ) imply that a colony will need one or several hours to correctly match workers to tasks when the demand for work in the different tasks changes . For the examples given here , that would imply a definite cost , in terms of not being able to maintain a correct match of workers to the tasks that need work ( since the level of demand for work is likely to change more frequently than every few hours , or because a lag in matching demand in the realm of hours implies a significant cost ) . If workers only require seconds to assess demand for work across all tasks ( e . g . because task stimuli are volatile pheromones , or global variables like temperature ) , and can choose a task based on this information , then the time cost of correct allocation in options ( 2 ) and ( 3 ) is likely insignificant . This would imply that a correct allocation can be achieved quickly , and thus workers should be dynamically and optimally reallocated to changing demands on a timescale of less than a minute . Another pattern emerging from these calculations is that under options ( 2 ) and ( 3 ) for choice , it is primarily the number of task types ( |T| ) that affects how fast task allocation proceeds . Neither the number of extra workers ( c ) nor the size of the initial work deficit ( Φ ) play a major role; also neither does ϵ , i . e . allowing a small amount of error in allocation does not decrease the time to successful reallocation in a meaningful way . How accurate are these conclusions , given that we are only examining somewhat arbitrarily chosen parameter combinations ? Our results in Table 3 give a more complete picture , as do the plots in Fig 2; this table is only intended as an illustration of the results . However , the parameter values illustrated here are not entirely arbitrary , but represent best-guesses given empirical data ( see Table 4 ) . For example , many authors have tried to examine the number of task types in social insects , and our results cover the range generally found ( 2–30; Table 4 ) .
The process of task allocation and its typical outcome , division of labor , have received a lot of attention in the social insect literature . Empirical studies typically focus on determining the individual traits or experiences that shape , or at least correlate with , individual task specialization: e . g . when larger or older individuals are more likely to forage ( e . g . [53] ) or when interaction rates or positive experience in performing a task affect task choices [32 , 64] . Generally the re-allocation of workers to tasks after changes in the demand for work often needs to happen on a time scale that is shorter than the production of new workers ( which , in bees or ants , takes weeks or months , [65] ) , and indeed empirical studies have found that the traits of new workers do not seem to be modulated by colonies to match the need for work in particular tasks [66] . Therefore , more recent empirical and most modeling studies focus on finding simple , local behavior rules that generate individual task specialization ( i . e . result in division of labor at the colony level ) , while simultaneously also enabling group-level responsiveness to the changing needs for work in different tasks [35 , 67 , 68] . For example , in classic papers , Bonabeau et al . [69] showed theoretically that differing task stimulus response thresholds among workers enable both task specialization and a flexible group-level response to changing task needs; and Tofts and others [70 , 71] showed that if workers inhabit mutually-avoiding spatial fidelity zones , and tasks are spread over a work surface , this also enables both task specialization and flexible response to changing needs for work . In this paper we examined how well we should expect task allocation to be able to match actual demands for work , and how this will depend on group size and the number of ‘extra’ , thus inactive , workers . Neither of the modeling studies cited above explicitly considered whether task allocation is improved or hindered by colony size and inactive workers . In addition , while several studies find increasing levels of individual specialization in larger groups , the empirical literature overall does not show a consensus on how task allocation or the proportion of inactive workers is or should be affected by group size ( reviewed in [14 , 22] ) . In general , few studies have cosidered the efficiency of the task allocation process itself , and how it relates to the algorithm employed [72] , often in the context of comparing bio- ( ant- ) inspired algorithms to algorithms of an entirely different nature [73 , 74] . For example , Pereira and Gordon , assuming task allocation by social interactions , demonstrate that speed and accuracy of task allocation may trade off against each other , mediated by group size , and thus ‘optimal’ allocation of workers to tasks is not achieved [72] . Duarte et al . also find that task allocation by response thresholds does not achieve optimal allocation , and they also find no effect of colony size on task allocation performance [75] . Some papers on task allocation in social insects do not examine how group size per se influences task allocation , but look at factors such as the potential for selfish worker motives [76] , which may be affected by group size , and which imply that the task allocation algorithm is not shaped by what maximizes collective outcomes . When interpreting modeling studies on task allocation , it is also important to consider whether the number of inactive workers is an outcome emerging from particular studied task allocation mechanisms , or whether it is an assumption put into the model to study its effect on efficiency of task allocation . In our study , we examined how an assumed level of ‘superfluous’ , thus by definition ‘inactive’ , workers would affect the efficiency of re-allocating workers to tasks after demands had changed . While the above models concern the general situation of several tasks , such as building , guarding , and brood care , being performed in parallel but independently of one another , several published models of task allocation specifically consider the case of task partitioning [77] , defined in the social insect literature as a situation where , in an assembly-line fashion , products of one task have to be directly passed to workers in the next task , such that a tight integration of the activity in different tasks is required . This is , for example , the case in wasp nest building , where water and pulp are collected by different foragers , these then have to be handed to a construction worker ( who mixes the materials and applies them to the nest ) . Very limited buffering is possible because the materials are not stored externally to the workers , and a construction worker cannot proceed with its task until it receives a packet of water and pulp . One would expect different , better-coordinated mechanisms of task allocation to be at work in this case . In task partitioning situations , a higher level of noise ( variation in availability of materials , or in worker success at procuring them ) increases the optimal task switching rate as well as the number of inactive workers , although this might reverse at very high noise levels [78] . Generally larger groups are expected to experience relatively lower levels of noise [79] . In this line of reasoning , inactive workers are seen as serving a function as ‘buffer’ ( or ‘common stomach’ , as they can hold materials awaiting work ) [79 , 80]; this implies that as noise or task switching rate increase , so does the benefit ( and optimal number ) of inactive workers . Is task allocation a difficult problem , and does it matter which algorithm is chosen ? If task allocation is an easy problem , then the match of work to workers can be achieved without significant costs . If task allocation is difficult , on the other hand , the choice of task allocation algorithm matters for system performance; in biological systems where this is the case , we would expect task allocation mechanisms to be under strong selection , and their evolution to reflect the specific ecological context of the system . In social insect colonies , for example , task allocation mechanisms appear to differ between species—this could be the case because they are not under selection , and different species happen to have hit on different , equally good , solutions , or because they are under selection , and different species have different requirements ( e . g . because they differ in the frequency with which demand for work in different tasks changes ) . There is some evidence that even brief mismatches of work to workers , i . e . incorrect task allocation , can be detrimental in certain species ( e . g . because brood do not develop well when briefly not thermoregulated [81] ) . Here we estimate the time to correct allocation for several species and contexts ( Table 4 ) . For example , we estimate that when a honey bee colony is attacked by a large predator , and 5000 ( ±30% ) bees should ideally be allocated to defense , the time to achieve this within our generalized task allocation algorithm would be around 5 − 10 rounds if all bees can directly sense the need for more defenders ( options ( 2 ) or ( 3 ) ) , and 700 rounds if they cannot ( and only arrive in the defense task because they randomly tested different tasks in different rounds , option ( 1 ) ) . Since this particular situation requires a quick collective response , the difference between option ( 1 ) and options ( 2 ) or ( 3 ) appears meaningful , regardless of whether a ‘round’ takes minutes or seconds to complete . In another example , a change in foraging conditions in the case of rock ants ( Temnothorax ) may imply that only five additional workers need to be allocated to the task of foraging; however , in that system it appears likely that individuals need on the order of a minute rather than seconds to assess both the state of their environment and whether their own task performance is successful ( in the sense of fulfilling a demand ) . If that is the case , a delay of 40 rounds may also be a meaningful and costly delay to appropriately exploiting novel food sources , for example . In all cases , the main effect on the difficulty of task allocation is whether or not individuals can assess the demand across different tasks simultaneously ( instead of only in the one task they are working on ) , and what time period a ‘round’ in our model corresponds to ( i . e . how long it takes a worker to assess whether its current work is needed , i . e . whether it is ‘successful’ in the task according to the terms used in our model ) . In addition , the costs as presented in Table 4 have to be paid each time the demands for work in different tasks change , and workers have to be reallocated to match these new demands . Overall , our calculations show that realistic parameter estimates can lead to potentially meaningful costs of slow task allocation . Our calculations are pretty coarse however , as the precise values of many of the parameters are not known ( however see Table 2 for references on parameter estimates ) . More empirical work in this area would be useful . Our work also addresses a more general question . Division of labor is widespread in complex systems from developing embryos to human organizations; it implies a degree of individual specialization , i . e . more or less consistent differences between individuals in the tasks chosen . Division of labor is often associated with ‘progress’ or ‘increase in complexity’ ( e . g . [17] ) . All systems with division of labor must implement some algorithm that lets individuals choose their task . How do these task allocation algorithms evolve , i . e . which external or internal conditions select for which kinds of mechanism ? For example , under which conditions and in which systems do the best task allocation algorithms produce highly specialized workers , insensitive to small changes in demands across tasks ? One might argue that in a system with highly specialized workers , the cost of allocation mismatch is never more than the average allocation minus current demands , because the system can make specialized workers in the correct proportion for the average expected allocation . Any algorithm that allows workers to be fully generalist , i . e . to freely switch between any tasks , must ensure that the mismatch of workers to demands is not on average greater than that . Understanding more about why particular task allocation mechanisms are selected for would thus increase our understanding about the evolution of specialization and division of labor more generally . Does colony size lead to a change in which task allocation algorithms perform well , and does it lead to selection for specialization ? The answers to these questions are not straightforward ( and neither are the empirical results on this [22] ) . Contrary perhaps to conventional wisdom in both biology and computer science , we do not find a direct dependence of the time to solve the task allocation problem on ‘colony size’ or ‘problem size’ , if we assume that the total amount of work scales linearly with the number of workers ( c = |A|/D , the number of workers per work available , is constant across different |A| ) . This holds even if all work only has to be satisfied with a certain probability , and if only close to the total needed work has to be satisfied . This result is perhaps logical because we implemented neither the type of noise that would lead to a benefit of large numbers ( where the relative amount of variation in environments experienced decreases with colony size ) , nor did we implement any economies of scale ( there are no broadcast signals; we did not model any communication explicitly , and if the task feedback is thought of as the result of communication , we did not implement any constant costs with colony size ) . No matter how logical in hindsight however , this was not what we had intuitively expected nor what is sometimes suggested in the literature [22] . If we find empirically that in some systems the level of specialization or the task allocation mechanism implemented change with colony size , some factors not modeled here have to be at play: e . g . fixed costs leading to economies of scale , or non-linear scaling in the effectiveness of communication . For example , it may be that the feedback on whether an individual worker contributes to reducing a deficit depends on social interactions that do not scale linearly with colony size . This is plausible of course ( and has been demonstrated empirically in some cases , e . g . [50] ) . Importantly however , it is not obvious that task allocation will perform better at larger colony sizes in all systems . It is worth noting that even if the time to correct allocation did change with colony size , this does not make obvious predictions for the evolution of division of labor ( the degree to which workers should be specialized ) . If task allocation is difficult ( takes a long time ) , it may be that colonies abandon the attempt to dynamically reallocate workers at all , and instead employ specialized , ‘preprogrammed’ workers in proportions of the average expected demands across tasks . We discover that to understand the dependence of task allocation on the number of workers in the colony ( |A| ) , actually what we really need to know is ( D ) , the total amount of work that needs to be done . Note that D refers to currently open tasks , thus is not likely to be ‘unlimited’; in social insects , if nothing else , in the short term , available work will be limited by the queen’s egg laying rate . This total amount of work available ( or necessary ) has not been studied explicitly either empirically or in models of social insect task allocation , with few exceptions [28] . So , we do not have a good understanding of how D behaves with |A| intra- or inter-specifically . Here we have simply assumed that |A|/D is constant , but this may well not generally be so: previous studies and conceptual papers have suggested either that larger colonies are relatively less productive , perhaps suggesting that less work is available per worker , or that they are more productive ( because they are capitalizing on some economies of scale ) — it is unclear what the latter would imply for the amount of work per worker available . One interesting new hypothesis here is that the evolution of task allocation across social insects may , in part , be driven by the factors that limit productivity -– e . g . is the colony raising brood at near the queen’s maximal egg laying rate ? In this case D may increase less than linearly with increasing colony size , and thus task allocation may become easier , even trivial , at higher colony sizes . Our modeling study thus suggests a new hypothesis ( one for the purposes of modeling more generally , [82] ) , by providing the insight that a previously ignored variable impacts the outcome of a well-studied process . Our results also suggest that c ( the ratio of |A|/D , or the number of workers divided by the amount of work available ) matters , and higher c generally leads to faster allocation time . Thus colonies may benefit from having more workers available than work . This is a novel hypothesis for the existence of ‘inactive’ workers in social insect colonies and other complex systems [14] . That is , colonies may produce more workers than needed to complete available work simply in order to speed up the process of ( re- ) allocating workers to work , and thus potentially reducing costs of temporary mismatches of workers with needed work . In other words , inactive , ‘surplus’ workers in colonies may increase colony flexibility and how close colonies get to an ‘optimal’ task allocation in environments where task demands often change and workers frequently have to be reallocated . The benefit of extra workers ( higher c ) does not depend on colony size ( |A| ) , thus we would expect both large and small colonies to have as many extra workers as they can afford . Although the dependence on c varies with task allocation algorithm ( it is least strong in option ( 2 ) ) , higher c is always beneficial . Apparently inactive workers are common in social insect colonies . While these workers may be selfish [40 , 41] or immature [42] , or temporarily unemployed due to fluctuating total demand [14] , our work here thus implies that they may also be present simply to improve task allocation . That is , colonies may produce extra workers such that some workers are ‘unemployed’ at all times on average , but so that the time to correct reallocation of workers when demands across tasks change is minimal . This is a novel hypothesis for the function of inactive workers in complex systems more generally . It is intuitive that task allocation may be more difficult if workers have to choose among many different possible tasks to perform ( high |T| ) . However , we show that the effects of |T| are mixed and depend both on the information available to workers and the actual combination of parameter values , particularly on the size of |T| . Specifically , in the parameter ranges we explored numerically ( based on empirically plausible parameter values ) , the time to correctly allocated workers to tasks depends linearly on the number of task types for options ( 1 ) and ( 2 ) , and not at all for option ( 3 ) . In option ( 1 ) , where workers effectively have to ‘test’ tasks sequentially to discover where work is needed ( because they only find out through the success mechanism ) , |T| always enters into performance as a linear factor . This would be the case for example if workers have to walk to different locations in the nest , or if they have to invest some other significant effort into assessing demand for each specific task . In options ( 2 ) and ( 3 ) , workers can effectively assess demand across all tasks in parallel; this may be the case if task demand is communicated through global stimuli , such as temperature or volatile pheromone levels . In such a case , the number of task types matters only if it is lower than the second term in the minimum function ( for example , see Corollary C . 6 in S2 Text ) . Thus , whether the number of task types affects task allocation performance depends on the context of other parameter values . What do we know about |T| empirically ? Several authors have attempted to quantify this number ( see Table 2 ) . However , empirically studies have often acknowledged that what are ‘separate tasks’ and what are just elements of the same task is difficult to define , and that this may lead to number estimates that are quite subjective . In our model , workers within the same task are assumed to immediately ( within one round ) correctly distribute the work among themselves , whereas the demand for work in a different task is only assessed via the choice and success feedback mechanisms as defined above . So , one may think , for example , of each item to be worked on as a ‘task’ ( e . g . each larva that needs tending and feeding , or each breach in the wall ) , in which case |T| might be a quite large number; or one may think that all larvae are part of the single task of brood care , and all places in the wall that need repair are part of the task of nest building , in which case |T| is likely to be a small number ( perhaps below 20 , or even below 10 ) . Which is the more appropriate way of counting tasks , in the context of our model , depends on whether , for example , each ant worker dedicated to brood care will be able to immediately assess which particular larvae need care , not loosing time in arriving at a consensus with other brood care workers about who is tending to which exact brood item , or alternatively where each brood care worker can jointly and concurrently contribute to the work in that task without internal coordination required at the timescale of overall task allocation . The results presented in this paper were derived using methods from the field of theoretical distributed computing . The problems considered in this field are very similar to those that are relevant in the biological study of distributed systems—and almost all biological units , from cells with their metabolic and molecular networks to ecosystems , are really distributed systems . We believe that the techniques and results from theoretical distributed computing may lead to many novel approaches and insights in biology in the future , and interdisciplinary work in this area is increasing [29 , 46 , 47 , 83 , 84] . In particular , research in theoretical distributed computing has examined the limitations of distributed algorithms , for example in such contexts as distributed task allocation as we study here . Generally , this field analytically derives results about models that often assume stochastic individual behavior , and in particular quantifies system-level performance given specific individual algorithms ( i . e . behavioral rules ) . Here , we have analyzed how our model , a generalized form of an insect-inspired task allocation algorithm , performs in terms of how long it takes to correctly allocate workers to different task types which need work . We have allowed for approximate solutions , by looking at the time to allocating workers correctly only with a certain minimum probability ( 1 − δ ) , and only to within ϵ of the best allocation . We have also allowed for errors in the demand assessment function , e . g . if workers make mistakes when assessing whether they are needed in a particular task . We have made the assumption that the relevant measure of how well a task allocation mechanism performs is related to the time to correct allocation , that is the time until workers are matched to tasks that need work . Other performance measures are possible , such as assessing how quickly the task-worker match approaches an ideal allocation , or how good the match can ever get; or entirely different parameters may be under selection , such as how much workers have to switch tasks [38] , how well workers prioritize more important tasks over unimportant ones , or how much information workers need to collect in order to allocate correctly . Second , our approach makes another assumption about how the performance of a task allocation mechanism is measured: we only quantify this performance for the worst-case inputs , namely the configuration of task deficits ( i . e . the distribution of unfulfilled demands across tasks ) that leads to the longest possible time to re-allocate . Thus , while stochasticity in worker decisions and information is taken into account and expected results derived , we do not make any assumptions about what configuration of task deficits workers are likely to encounter . If this was known , more precise expectations for performance could be derived . In distributed computing theory , there is a general assumption that such a worst-case scenario ( generally called the upper bound of performance ) is a good measure of algorithm performance; however it does not need to be close to the overall expected case . Finally , we make the crucial assumption that all workers are identical in preferences and skills . Thus , our model represents a system of flexible , homogeneous workers . If workers randomly differed in their ability to perform different tasks , matching them optimally to tasks with changing demands for work becomes an extremely hard problem [12] . On the other hand , worker skills in a task may be linked to their preferences for that task , either because these are innately linked , or because workers learn to prefer the tasks they do well , or learn to do the tasks well they prefer [85] . How much the dynamic ( re- ) allocation of workers in response to changing demands in different tasks is affected by such worker heterogeneity remains to be analyzed . We mathematically derived how the time it takes to correctly allocate workers to work is affected by several factors , such as colony size and the number of ‘extra’ workers . We make only minimal assumptions about the algorithm used , and we explore several ways of measuring performance of task allocation , which means these relationships should hold fairly generally . Our model brings several insights . First , costs or benefits of group size do not arise in task allocation ‘automatically’ , that is from minimal assumptions such as ours . Second , such a result clarifies our thinking and suggests how , for example , colony-size-dependencies may occur ( e . g . if information on work deficits is communicated faster in larger colonies ) , thus guiding future research as well as identifying which variables qualitatively affect system behavior . One such variable is the amount of work available; this has not been considered in previous empirical studies but appears to be a crucial factor affecting the evolution of task allocation algorithms [28] . Third , the model results have generated a novel hypothesis for the existence of inactive workers in social insect colonies [14] , namely that they may serve to speed up the task allocation process . It now can be studied whether this may be the reason for their evolution . All of these results are derived analytically , using approaches from theoretical distributed computing , without the need for parameter estimation such as would be necessary in a simulation study . In summary , our ‘proof of concept’ model sensu [63] helps to identify how limitations and processes at the individual level can affect group level processes in a distributed system .
|
Many complex systems have to allocate their units to different functions: cells in an embryo develop into different tissues , servers in a computer cluster perform different calculations , and insect workers choose particular tasks , such as brood care or foraging . Here we demonstrate that this process does not automatically become easier or harder with system size . If more workers are present than needed to complete the work available , some workers will always be idle; despite this , having surplus workers makes redistributing them across the tasks that need work much faster . Thus , unexpectedly , such surplus , idle workers may potentially significantly improve system performance . Our work suggests that interdisciplinary studies between biology and distributed computing can yield novel insights for both fields .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"honey",
"bees",
"applied",
"mathematics",
"sociology",
"social",
"sciences",
"animals",
"simulation",
"and",
"modeling",
"social",
"systems",
"algorithms",
"systems",
"science",
"mathematics",
"animal",
"behavior",
"probability",
"distribution",
"zoology",
"research",
"and",
"analysis",
"methods",
"bees",
"complex",
"systems",
"foraging",
"computer",
"and",
"information",
"sciences",
"hymenoptera",
"behavior",
"insects",
"probability",
"theory",
"arthropoda",
"eukaryota",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms",
"numerical",
"analysis"
] |
2017
|
Costs of task allocation with local feedback: Effects of colony size and extra workers in social insects and other multi-agent systems
|
Mutations of the Wnt5a gene , encoding a ligand of the non-canonical Wnt pathway , and the Ror2 gene , encoding its receptor , have been found in patients with cardiac outflow tract defects . We found that Wnt5a is expressed in the second heart field ( SHF ) , a population of cardiac progenitor cells destined to populate the cardiac outflow tract and the right ventricle . Because of cardiac phenotype similarities between Wnt5a and Tbx1 mutant mice , we tested potential interactions between the two genes . We found a strong genetic interaction in vivo and determined that the loss of both genes caused severe hypoplasia of SHF–dependent segments of the heart . We demonstrated that Wnt5a is a transcriptional target of Tbx1 and explored the mechanisms of gene regulation . Tbx1 occupies T-box binding elements within the Wnt5a gene and interacts with the Baf60a/Smarcd1 subunit of a chromatin remodeling complex . It also interacts with the Setd7 histone H3K4 monomethyltransferase . Tbx1 enhances Baf60a occupation at the Wnt5a gene and enhances its H3K4 monomethylation status . Finally , we show that Baf60a is required for Tbx1–driven regulation of target genes . These data suggest a model in which Tbx1 interacts with , and probably recruits a specific subunit of , the BAF complex as well as histone methylases to activate or enhance transcription . We speculate that this may be a general mechanism of T-box function and that Baf60a is a key component of the transcriptional control in cardiac progenitors .
The second heart field ( SHF ) provides progenitor cells for the development of several segments of the mature heart , such as the outflow tract , right ventricles and atria [1] , [2] . Mouse models of congenital heart disease suggest that perturbation of SHF development may be the basis of relatively common heart defects in humans , such as conotruncal anomalies , but the transcriptional mechanisms driving SHF development are not well defined . An important example of a congenital heart disease gene that functions within the SHF is TBX1 encoding a T-box transcription factor . This is haploinsufficient in the DiGeorge/Velocardiofacial/22q11 . 2 deletion syndrome , which is associated with various types of cardiac outflow tract ( OFT ) and vascular defects [3] . Tbx1 mouse mutants recapitulate well the human phenotype , which has facilitated a detailed analysis of the role of the gene in heart development . In particular , Tbx1 is expressed in the SHF where it sustains cell proliferation and inhibits differentiation [4] . However , there is considerable less information about the effectors of these developmental roles , and about the mechanisms for target gene regulation . SHF cardiac progenitors , which reside outside the heart , are thought to migrate into the heart and differentiate as they are incorporated into the OFT . Thus , it is reasonable to expect that transcriptional regulation of SHF progenitors should involve cell polarity and cell migration , functions that in many cases are regulated by the non-canonical Wnt pathway [5] . Here we show that Wnt5a , which encodes a ligand of the non-canonical pathway , is expressed in the SHF . It has been shown that Wnt5a−/− animals have OFT abnormalities , though generally not as severe as in Tbx1−/− animals [6] . In addition , OFT defects have also been shown in Ror2−/− mice , Ror2 being a receptor of Wnt5a . Furthermore , mutations of WNT5a or ROR2 in humans are associated with Robinow syndrome [7] , which includes OFT defects , though at a low penetrance . Therefore , we postulated that there may be interaction between Tbx1 and the non-canonical Wnt pathway . We have crossed Tbx1 and Wnt5a mouse mutants and found that there is indeed a genetic interaction and , most interestingly , the loss of both genes caused developmental failure of the SHF-dependent heart segments , indicating that the two genes are required for SHF function . Next , we investigated the transcriptional mechanisms underlying this interaction . We found that Wnt5a is a transcriptional target of Tbx1 . Previous data showing a genetic interaction between Tbx1 and the gene encoding the chromodomain protein Chd7 [8] and physical interaction with the histone methyltransferase Ash2l [9] prompted us to investigate the role of chromatin remodeling and histone modifiers . We found that Tbx1 co-immunoprecipitates with Baf60a , a component of the SWI-SNF-like BAF chromatin remodeling complex , and with the Setd7 histone 3 , Lysin 4 monomethyltransferase . Tbx1 expression increases the occupation of the Wnt5a gene by Baf60a , and enhances the H3K4 monomethylation status of the chromatin in the T-box binding element ( TBE ) -harboring regions of the Wnt5a gene . Finally , we found that Baf60a is required for Tbx1-induced regulation of Wnt5a . Overall , our data support a mechanism by which Tbx1 enhances transcription of Wnt5a ( and possibly other target genes ) by interacting with a chromatin remodeling complex and enhancing H3K4 de novo methylation . This is the first evidence of an impact of Tbx1 on chromatin status and of its interaction with the BAF complex . Our data also raises the intriguing possibility that Baf60a is a critical BAF subunit for the transcriptional program of cardiac progenitors .
We carried out in situ hybridization of Wnt5a in embryos from E8 . 0 to E11 . 5 and found strong expression in the pharyngeal and splanchnic mesoderm including the SHF region ( Figure 1 ) . In addition , we found expression in the OFT , as previously noted [6] , [10] . We next tested whether Wnt5a expression may be altered by Tbx1 mutation . At E8 . 0–8 . 5 , Wnt5a expression was reduced in the SHF of Tbx1−/− embryos ( n = 3; Figure 1 ) , while it was not affected in the OFT or pharyngeal arch core mesoderm ( Figure 1 ) . 3D reconstruction analysis of section images provides an overall view of the expression domains affected by loss of Tbx1 ( Figure 1K–1P ) . Similar results were obtained at later stages ( E9 . 0–11 . 5 , not shown ) . To confirm that Wnt5a is expressed in cardiac progenitors , we carried out qRT-PCR on the P19CL6 mouse embryonal carcinoma cell line that is able to differentiate into cardiomyocyte progenitors upon 5-azacytidine and/or DMSO treatment [11] . We found that Wnt5a is robustly expressed after 3 days of differentiation ( Figure S1 ) . The activation of Wnt5a expression roughly coincides with the up-regulation of Tbx1 expression , suggesting that Tbx1 may have a role in the activation of Wnt5a rather than in its expression maintenance . We also found that loss of Wnt5a in these cells reduces cell migration , as tested by a wound-healing assay ( Figure S2 ) . Next we tested if Tbx1 can regulate Wnt5a expression in these cells . For this , we transfected graded amounts of a Tbx1-expressing vector into P19CL6 undifferentiated cells and evaluated expression of the endogenous Wnt5a gene by qRT-PCR . Results revealed that Wnt5a expression responds positively to Tbx1 in a dosage-dependent manner ( Figure S3 ) . To test whether Wnt5a and Tbx1 interact in vivo , we crossed Tbx1+/−;Wnt5a+/− mice and determined the embryonic phenotype on the progeny at E18 . 5 . Results of the analysis of 162 embryos are summarized in Tab . 1 . Double heterozygous embryos ( Tbx1+/−;Wnt5a+/− , n = 61 ) showed higher penetrance of the typical Tbx1 haploinsufficiency phenotype ( hypoplasia or aplasia of the 4th pharyngeal arch artery ) than did the Tbx1+/− embryos ( n = 17 ) ( 44 . 3% vs . 23 . 5% , p<0 . 05 ) . However , we did not find any additional phenotypic abnormalities in double heterozygous embryos . We did not find any cardiovascular abnormality in Wnt5a+/− embryos ( n = 30 ) consistent with previously reported data [6] . Analysis of Wnt5a−/− embryos ( n = 5 ) revealed ventricular septal defects ( VSD ) and abnormal positioning of the great arteries ( aorta and pulmonary trunk ) , which were side-by-side ( Figure 2B–2D ) . However , in none of these embryos did we observe truncus arteriosus communis ( TAC ) , although this defect was previously reported for most Wnt5a−/− embryos [6] . This difference may be due to different genetic backgrounds . However , the extracardiac phenotype that we found was consistent with previously reported data ( cleft palate , cleft lip , small thymus , truncated tail and limbs ) ( Figure 2A and data not shown ) . Analysis of Tbx1+/−;Wnt5a−/− embryos ( n = 22 ) revealed a more severe phenotype than in Wnt5a−/− embryos from the same crosses . In particular , out of 22 embryos examined , 13 ( 59% ) showed additional abnormalities compared to Wnt5a−/− embryos , namely TAC , small ear and edema ( Figure 2A′–2D′ ) . The other embryos ( 9 out of 22 or 41% ) were phenotypically indistinguishable from Wnt5a−/− embryos . Together , these results indicate a genetic interaction between Tbx1 and Wnt5a . We did not retrieve any double homozygous embryos at E18 . 5 ( n = 162 ) , suggesting that this genotype is lethal during early embryogenesis . Therefore , we harvested embryos at earlier embryonic stages . The latest stage at which we found live Tbx1−/−; Wnt5a−/− embryos was E9 . 5 . At this stage , double homozygous embryos ( n = 10 ) showed a severe cardiac phenotype characterized by severe hypoplasia of the OFT and right ventricle ( RV ) , structures that sometimes appeared to be absent by visual inspection ( Figure 3 ) . Histological sections confirmed that the development of these structures is severely affected ( Figure 3A″–3D″ ) . We carried out in situ hybridization with a probe for CyclinD2 , a marker of the proximal OFT and the right ventricle ( RV ) [12] . Results showed that CyclinD2 was normally expressed in control , Tbx1−/− and Wnt5a−/− E9 . 5 embryos , but it was greatly reduced or undetectable in Tbx1−/−;Wnt5a−/− embryos ( Figure 4 ) , confirming the severe hypoplasia of the OFT and of the RV . We asked whether the hypoplasia of these structures could be due to increased apoptosis . We carried out immunohistochemistry on E9 . 5 embryos using an anti-cleaved Caspase 3 antibody . Results showed very few positive cells in the OFT and RV of WT and Tbx1−/− E9 . 5 embryos , while both Wnt5a−/− and Tbx1−/−;Wnt5a−/− embryos exhibited an increased number of apoptotic cells ( Figure 5A–5D ) . However , we found no obvious difference in apoptosis between Wnt5a−/− and Tbx1−/−;Wnt5a−/− embryos . Therefore , it is unlikely that apoptosis explains the severe heart phenotype of double mutants . It has been reported that there is antagonism between the non-canonical and canonical Wnt pathways ( e . g . [13] , [14] ) . Therefore , we asked whether there could be alterations of β-catenin expression in our mutants . By immunofluorescence , β-catenin expression in the SHF did not change in Tbx1−/− or Wnt5a−/− embryos at E9 . 5 ( Figure 5F–5G and 5F′–5G′ ) , but it was upregulated in the SHF of Tbx1−/−;Wnt5a−/− embryos ( Figure 5H–5K and 5H′–5K′ ) . To understand whether Wnt5a may be a direct target of Tbx1 , we searched for T-Box binding elements ( TBEs ) in a 20 kbp genomic sequence encompassing the mouse Wnt5a gene . We found three putative TBEs , two of which closely located in intron 3–4 ( TBE1: AAGGGGTGAA , TBE2: GTAGGTGCCAGG ) and one in the 3′-UTR ( TBE3: AGAGGTGTTGCA ) ( Figure 6A ) . We next cloned two evolutionarily conserved ( in human and mouse ) DNA segments containing TBE1–2 and TBE3 , into a luciferase reporter plasmid with a basic promoter ( Figure 6B–6C ) . We also generated mutagenized constructs in which the three TBEs were mutated ( the core sequence GTG was changed into AAA ) . We then carried out luciferase assays . Results showed that the two DNA segments responded well to transfected Tbx1 ( Figure 6B–6C ) . However , mutation of the TBEs , individually or combined , all had a significant impact on the ability of Tbx1 to activate the constructs ( Figure 6B–6C ) , although the mutation of TBE1 had a relatively lower effect . Thus , these TBEs are required for Tbx1-induced activation in this assay . Next , we tested whether the endogenous Tbx1 protein occupies these TBE sites in P19Cl6 cells by standard and quantitative chromatin immunoprecipitation ( ChIP ) assays using a Tbx1 antibody . TBE1 and TBE2 were assayed together because they are too close to be assayed independently . Results demonstrated that indeed these sites are occupied by endogenous Tbx1 in P19CL6 cells ( Figure 7A–7B ) . Next , we repeated the same assay using chromatin from E9 . 5 wild type embryos and again we could demonstrate enrichment at the Wnt5a TBE loci ( Figure 7C ) . To begin to dissect the transcriptional machinery within which Tbx1 exerts its transcriptional functions , we tested by co-immunoprecipitation ( co-IP ) or affinity purification chromatin remodeling and histone modification proteins , Baf60a , Baf60c , Baf155 , Setd7 and p300 . Under the experimental conditions tested , we found that Baf60a and Setd7 , but not Baf60c , Baf155 and p300 , co-immunoprecipitated with Tbx1 ( Figure 8 and Figure S4A ) ) . The Baf60a-Tbx1 interaction was also confirmed using a GST pull down assay ( Figure S4B ) . Furthermore , we confirmed Baf60a-Tbx1 co-IP using E9 . 5 embryo tissues ( Figure 9A ) . At this stage , the Baf60a gene is expressed broadly , including the SHF , but is very weakly expressed or absent in the heart ( Figure 9B ) . In contrast , Baf60c , an alternative component of the BAF complex , is mainly expressed in the heart and somites , and it has been previously shown to interact with another T-box transcription factor , Tbx5 [15] . While Tbx1 is mainly required in cardiac progenitors , where it is important to keep cells proliferating and to inhibit their differentiation , Tbx5 is important for cardiomyocyte differentiation . We compared expression of Tbx1 , Baf60a and Baf60c mRNA during P19Cl6 cell differentiation . Results showed that Tbx1 is expressed in the early phases of differentiation . Baf60a is particularly strong in these early phases and then its expression is reduced ( although still robustly expressed ) . In contrast , Baf60c expression is very low in the early phases of differentiation and becomes quite strongly expressed as differentiation proceeds ( Figure 1 ) . If the Baf60a-Tbx1 interaction were a feature of the transcriptional machinery at target genes , we would expect Baf60a to occupy the regions of the Wnt5a gene harboring the TBEs . To test this , we used ChIP with a Baf60a antibody on P19CL6 cells . Results demonstrated that indeed this protein occupies the TBE1/2 region as well as the TBE3 region ( Figure 10A ) . The same results were obtained with a ChIP assay using chromatin from E9 . 5 embryos ( Figure 10B ) . Given the interactions with the BAF complex and with Setd7 , we postulated that Tbx1 would recruit Baf60a to the target gene and would affect histone methylation . Indeed , quantitative ChIP using a Baf60a antibody showed that Tbx1 increases significantly the enrichment of Wnt5a TBE sequences ( Figure 10C ) . Thus , increased dosage of Tbx1 is sufficient to enrich the Wnt5a gene with a chromatin remodeling complex . Next , we asked whether Tbx1 dosage can modulate H3K4me1 at TBE regions of the Wnt5a gene . To this end , we carried out quantitative ChIP with an anti H3K4me1 antibody ( Setd7 is a monomethyltransferase ) on P19Cl6 cells transfected with Tbx1 . Results indicated that there is indeed enrichment of H3K4me1 after increased dosage of Tbx1 ( Figure 11A–11B ) . Next , we tested whether H3K4me1 is also affected by loss of Tbx1 in vivo . To this end , we carried out qChIP assays using chromatin from WT and Tbx1−/− E9 . 5 embryos . Results indicated that there is a significantly higher enrichment of H3K4me1 in WT embryos compared to mutant embryos , but limited to the TBE1/2 locus ( Figure 11C ) . Next , we tested additional histone H3 modifications associated with gene activation . In particular , we carried out qChIP using chromatin from P19Cl6 cells transfected with an empty vector or with a vector over-expressing Tbx1 using antibodies against H3K4me2 , H3K4me3 , or AcH3 ( recognizing acetylation of histone 3 ) . Results indicated that none of these modifications are enriched in correspondence of the TBE loci of the Wnt5a gene , regardless of Tbx1 transfection ( Figure 12 ) . Finally , we asked whether Baf60a is important for Tbx1-induced upregulation of the Wnt5a gene . To this end , we knocked-down Baf60a expression by RNA interference and determined the ability of Tbx1 to regulate Wnt5a in the presence of reduced Baf60a dosage in P19CL6 cells . Results showed that after a knock down of approx . 80% of Baf60a , Tbx1 was unable to regulate Wnt5a ( Figure 13 ) . Interestingly , without Baf60a , Tbx1 was not able to activate other candidate targets ( Fgf8 , Fgf10 and Cyp26a1 ) indicating that Baf60a is required for Tbx1 transcriptional activity in this context . Baf60a knock-down per se did not affect significantly the basal expression of endogenous Tbx1 , Wnt5a , Fgf8 , Fgf10 and Cyp26a1 in control experiments ( Figure S5 ) .
This work identifies Wnt5a as a novel and important transcriptional target of Tbx1 . Combined loss of the two genes caused severe hypoplasia of SHF-derived heart segments , and early embryonic lethality . This is a much more severe phenotype than that caused by loss of the individual genes . The exact cause of the dramatic heart phenotype will require further investigation . However , our data suggest at least two possible mechanisms . Loss of Tbx1 impairs the ability to expand the heart progenitors pool of the SHF [16] , the loss of Wnt5a impairs their ability to migrate ( and/or to be correctly orientated ) into the heart . Thus , a double loss would essentially amount to a double hit upon cardiac progenitors , which would be fewer in number and less capable of contributing to the heart . Another possible cause of the severe phenotype might be an interference with the Wnt canonical signaling pathway in the double mutants , as suggested by the observed upregulation of β-catenin expression . Constitutive activation of Wnt canonical signaling in the SHF leads to severe heart abnormalities that are similar to those observed in the double homozygous mutants [17] . The importance of the non-canonical Wnt signaling for OFT development has already been illustrated by the study of Wnt11 mouse mutants [18] . Wnt5a and Wnt11 are both expressed in the OFT but , in contrast to Wnt5a , Wnt11 is not expressed in the SHF . It would be interesting to cross the two mutants to see whether there is functional redundancy in the OFT proper . Because of the clear in vivo importance of the Tbx1->Wnt5a transcriptional control for SHF function , we decided to focus our attention on the mechanisms regulating this control , as a possible paradigm for transcriptional control in the SHF . We found an interaction between Tbx1 and Baf60a . The latter is a component of the BAF ( Brg1-Brm Associated Factor ) complex , which is related to the yeast and fly SWI-SNF chromatin remodeling complex . The canonical function of the BAF complex is to utilize its ATPase activity to rearrange the nucleosome distribution of chromatin , thus playing a crucial role in regulating accessibility to components of the transcriptional machinery , and thereby gene expression , either positively or negatively [19] . One of the critical issues concerning the biology of the BAF complex , is how it is targeted to genes that need to be regulated , i . e . how is specificity achieved despite the apparent homogeneity of the core components of the complex . One possibility is that specificity is provided by non-core subunits and by the recruitment activity operated by transcription factors that target specific genes . Our ChIP data suggest that this may be the case , as Tbx1 appears to recruit a BAF subunit onto the Wnt5a gene . Interaction with a histone methyltransferase would further help transcription , for example by stabilizing the remodeling machinery on the locus . We show that indeed Tbx1 expression correlates with increased H3K4 monomethylation of the Wnt5a TBE loci in cultured cells and in vivo . Interestingly , Tbx1 co-immunoprecipitates with Baf60a but not Baf60c . These are two alternative subunits of the BAF complex , possibly associated with different target genes and different cellular differentiation states . Baf60a has been associated with undifferentiated/multipotent status [20] , [21] , while Baf60c has been associated with differentiating muscle cells ( cardiac or skeletal ) [22] , [23] . We also show here that Baf60a tends to be downregulated during P19Cl6 cell differentiation , while Baf60c is upregulated . This finding is supported by in vivo expression data indicating that Baf60a is poorly represented in differentiated heart tissue in contrast to Baf60c , which is mostly expressed in heart and somite tissue [15] . Thus , it is possible that an exchange of BAF subunits plays a role in the passage from the progenitor state to the differentiated cardiomyocyte state . Exchange of BAF subunits has already been described during differentiation from neural progenitors to neurons ( Baf53a to Baf53b , and Baf45a to Baf45b ) [24] . Exactly what promoters or enhancers exchange BAF subunits during cardiac differentiation would be an interesting question to address using genome-wide ChIP-seq studies . Setd7 is a H3K4 monomethyltransferase [25]–[27] already shown to interact with another T-box transcription factor named Tbx21 ( also known as Tbet ) [28] . It is thus tempting to speculate that interaction with the BAF complex and histone methyltransferases is a common feature of T-box proteins and at the core of the transcriptional function of these important transcription factors . Interestingly , our tissue culture experiments evidentiated that Tbx1 has a positive effect on H3K4me1 enrichment , but not H3K4me2 , H3K4me3 , or H3 acetylation . This suggests that Tbx1 promotes de novo methylation H3K4 at TBE enhancers perhaps making them a target for additional regulators . Overall , our data support a model by which Tbx1 regulates Wnt5a by interacting with and perhaps recruiting a specific subunit of the BAF complex , along with the histone modifier enzyme Setd7 , resulting in activation or enhancement of transcription of the target gene . While a number of important molecular details remain to be clarified , our data using a tissue culture model indicate the importance of Baf60a for Tbx1-induced regulation of Wnt5a and perhaps other target genes . Hence it is reasonable to speculate that this BAF subunit is a key cofactor for Tbx1 function in cardiac progenitors .
All animal experimentations were carried out according to animal welfare regulations and guidelines of the USA and of the European Union . We used the mouse lines Tbx1+/lacZ ( here cited as Tbx1+/− ) [29] , available through the EMMA repository , and Wnt5a+/− [10] available through the Jackson Laboratories . Both are null alleles . Genotyping was carried out according to instructions provided by the original reports . All crosses were carried out in conventional , clean facilities in a C57Bl6/129SvEv mixed genetic background . Embryos were examined after manual dissection under a stereo microscope . In most cases we also carried histological sectioning of paraffin-embedded specimens . Whole mount in situ hybridization according to standard methods . Embryos were photographed and then sectioned . In some cases we used images of sections for 3D reconstruction using the Amira software . P19CL6 cells were grown in Dulbecco-Modified Minimal Essential Medium supplemented with 10% fetal bovine serum . For cardiomyocytes differentiation the cells were plated at a density of 5 . 0×105 cells/well on a 35-mm tissue culture dish . The next day , when cells reached ∼90% confluence , the medium was replaced with a growth medium containing 10 µM 5-Azacytidine for 24 h [11] . After treatment with 5- Azacytidine , cells were incubated in the growth medium containing 1 . 0% DMSO that was changed daily in order to remove the cell debris resulting from cell death . The experimental days were numbered consecutively beginning from the day of treatment with 5- Azacytidine ( day 0 ) . The generation of the stably transfected cell lines P19-Tbx1-TEV-PA and P19-TEV-PA has been described [30] . These cell lines have been used for affinity purification experiments ( see below ) . For transient transfection , cells were cultured in 10 cm dishes until 60–70% confluent and transfected with Fugene6 ( Roche ) following the manufacturer protocol . Dharmacon ON-TARGETplus SMARTpool Baf60a/Smarcd1 siRNA was used to knockdown Baf60a expression using Fugene6 transfection reagent . ON-TARGETplus siCONTROL Non-targeting pool was used for control transfections . P19Cl6 cells pelleted cells were resuspended in CE buffer ( 10 mM HEPES , 60 mM KCl , 1 mM EDTA , 0 . 075% v/v NP40 , 1 mM DTT and 1X protease inhibitors , pH 7 . 6 ) , and centrifuged . The nuclei were washed with 5× pellet volumes of cold CE buffer without detergent and centrifuged . 2× pellet volume of NE buffer ( 20 mM TrisHCl , 420 mM NaCl , 1 . 5 mM MgCl2 , 0 . 2 mM EDTA , 25% Glycerol and 1X protease inhibitors , pH 8 . 0 ) was added to the nuclear pellet and incubated on ice for 30 min . Nuclear and cytosolic extracts were recovered spinning at maximum speed for 30 min to pellet any nuclei . P19Cl6 cells were cultured to confluent monolayer in 12-well . Two hours before the experiment , we treated cells with mitomycin c ( 10 µg/ml ) and then we made a scratch wound using a standard 200-µl pipette tip . Wounded monolayers were washed with PBS and digitally photographed for the 0 hour timepoint using an inverted microscope equipped with a digital camera ( Leica AF6000LX time-lapse ) . Images of the wound were acquired every 30 min for 24 h . Subsequently , pictures were analyzed using the “Image J” software . The wound healing effect was calculated as area wound closure compared with the area of the initial wound . Briefly , the distance between the wound margins was measured at 0 hours and again every 4 hours post-wounding for 24 h . 12 hours post–wounding the following formula was used to evaluate the area wound closure: distancet = 12 h-distancet = 0 h . Data represent the average of at least 3 independent experiments ( 2 wells/experiment ) . Native affinity purification was performed with strains P19-Tbx1-PA and P19-PA as previously described [30] , [31] . Briefly , nuclear extracts were transferred to Poly-Propylene Chromatography Columns with IgG sepharose beads ( Amersham ) to capture Tbx1-TEV-Protein A-containing complexes . Then , the columns were subjected to protease TEV cleavage to release the Tbx1-containing protein complexes , which were recovered and transferred into PVDF membrane ( Amersham ) for Western blotting analyses . The antibodies used for immunoblotting were the monoclonal anti-Baf60a antibody ( BD Biosciences , #611728 ) and the anti-p300 ( BD Biosciences , #554215 ) . For co-immunoprecipitation experiments , nuclear extracts of P19Cl6 cells were quantified using a modified Bradford procedure ( Bio-Rad Laboratories , Hercules , CA ) . Approximately 100 µg of nuclear extracts were incubated with an anti Tbx1 antibody ( Abcam , #ab18530 ) or Baf60a antibody ( BD Biosciences , #611728 ) or rabbit/mouse IgG ( Santa Cruz Biotechnology , #2027 ) and then incubated with Protein A/G PLUS agarose ( Santa Cruz Biotechnology ) at 4°C ON . The samples were washed 6 times with IPP150 ( 10 mM Tris-HCl , pH 8 . 0; 150 mM NaCl; 0 . 1% NP-40 ) and resuspended in SDS sample buffer . 10 µg of nuclear extracts ( 10% Input ) and immunoprecipitated samples were detected using Western blot analysis . We used anti-Baf60a antibody ( BD Biosciences , #611728 ) , an anti-Setd7 antibody ( Abcam , #ab71214 ) , anti-p300 ( BD Biosciences , #554215 ) and anti-SMARCC1/Baf155 ( Abcam , #ab72503 ) . P19Cl6 cells were cross-linked using 1% formaldehyde at room temperature for 15 min , and the reaction was stopped using glycine at a final concentration of 0 . 125 M for 5 min . Cells were then lysed in 1 ml of lysis buffer ( 10 mM HEPES , 60 mM KCl , 1 mM EDTA , 0 . 075% v/v NP40 , 1 mM DTT and 1X protease inhibitors , pH 7 . 6 ) on ice for 10 min , dounced using a 2 ml B dounce to release nuclei . Isolated nuclei were suspended in Nuclei lysis buffer ( 20 mM TrisHCl , 420 mM NaCl , 1 . 5 mM MgCl2 , 0 . 2 mM EDTA , 25% Glycerol and 1X protease inhibitors , adjusted to pH 8 . 0 ) , incubated on ice for 30 min , washed in nuclei lysis buffer and sonicated to obtain 200–500 bp . Sonicated chromatin was immunoprecipitated 10 µg of a Tbx1 antibody ( Abcam , #ab18530 ) , 2 µg of Baf60a antibody ( BD Biosciences , #611728 ) , 2 µg of H3K4me1 antibody ( Abcam , #ab8895 ) , 2 µg of Dimethyl-Histone H3 ( Lys4 ) ( Millipore , #07-030 ) , 5 µg of Anti-trimethyl-Histone H3 ( Lys4 ) ( Millipore , #07-473 ) , 5 µg of Anti-acetyl-Histone H3 ( Millipore , #06-599 ) , or normal rabbit/mouse IgG ( Santa Cruz Biotechnology , #2027 ) . Samples were then incubated with 20 µl of protein A/G PLUS agarose ( Santa Cruz Biotechnology ) at 4°C ON . The samples were extensively washed and incubated in an elution buffer ( 1% SDS and 0 . 1 M NaHCO3 ) at 30°C for 20 min . Cross-linking of protein-DNA complexes was reversed at 65°C ON , followed by treatment with DNase-fre RNase A for 30 min at 37°C and 100 µg/ml proteinase K for 2 h at 55°C . Phenol/chloroform-extracted , and ethanol-precipitated DNA was resuspended in 50 µl of H2O and subjected to PCR amplification ( see below ) . ChIP experiments using embryo tissue was carried out as follows . WT and Tbx1−/− E9 . 5 mouse embryos were dissected , their heads were cut off and the trunks were stored at −80°C . Pools of 5 embryos were used for each ChIP experiment . Each ChIP experiment was performed using 5 µg of H3K4me1 ( Abcam , #ab8895 ) , 10 µg of TBX1 antibody ( Abcam , #ab18530 ) and 5 µg or 10 µg of normal rabbit IgG ( Santa Cruz Biotechnology , #2027 ) as negative control with the LowCell ChIP Kit ( Diagenode ) . In brief , the embryos were fixed in 1% formaldehyde at room temperature for 15 min and then glycine was added to stop the reaction to a final concentration of 0 . 125 M for 5 min . Subsequently , embryos were washed with PBS , disaggregated by dounce homogenization in cold PBS and then resuspended in cell lysis buffer ( 5 mM PIPES pH 8 . 0 , 85 mM KCl , 0 . 5% Igepal supplemented with protease inhibitors ) . Next , nuclei were resuspended in 130 µl of Buffer B ( LowCell ChIP Kit reagent ) and chromatin was sheared into 200–500 bp long fragments using the Covaris S2 Sample Preparation System ( Duty Cycle: 5% , Cycles: 6 , Intensity: 2 , Temperature: 4°C , Cycles per Burst: 200 , Power mode: Frequency Sweeping , Cycle Time: 60 seconds , Degassing mode: Continuous ) . Following steps included incubation of the sheared chromation with antibody coated beads over night , several washing steps and reverse crosslinking . Next , equal DNA amounts of input and immunoprecipitated DNA were initially used as a template for conventional PCR amplification of the TBE1–2 region of Wnt5a ( Wnt5a TBE1–2_1_F 5′-CTTCCCCTGGTGTGGATATG-3′ , Wnt5a TBE1–2_1_R 5′-AGAGGCTCCTTCCAGTCCTC-3′ ) and TBE3 region ( wnt5a TBE-3_1_F 5′-ACTGCTGGTAGGGCAGAAAA-3′ , wnt5a TBE-3_1_R 5′-TCAGGCACCATTAAACCACA-3′ ) . For quantitative ChIP , we next carried out real-time PCR of the immunoprecipitated DNA and inputs , using the FastStart Universal SYBR Green Master kit ( Roche ) and the 7900HT Fast Real-Time PCR System ( Applied Biosystems ) . ChIP signals were normalized to that of an internal control amplimer that we have selected from an ORF-free region of mouse chromosome 14 ( INT-XIV-1F 5′- TTCTTGTCCACAGCCCTCTT-3′ , INT-XIV-1R 5′-TGGTGGAAGAGGAGACATCC-3′ ) . PCR efficiencies were determined for each primer pair using standard curves . PCR was carried out on 2 µl of 1/100 dilution of the input or 2 µl of immunoprecipitated samples . A GST-Tbx1 expression vector was kindly provided by Dr . Amendt ( IBT , Texas A&M University , Houston , TX USA ) and transformed into E . coli BL21 ( DE3 ) cells . Protein synthesis was induced by the addition of IPTG to a final concentration of 0 . 1 mM . GST pull-down assay was performed using ProFound Pull-Down GST Protein: Protein Interaction Kit ( PIERCE ) . RNA was extracted from P19Cl6 cells using TRI-Reagent ( Ambion/Applied Biosystems ) according to the manufacturer's protocol . Extracted RNA was treated with DNA-free Kit ( Ambion/Applied Biosystems ) . cDNA was synthesized from 2 µg total RNA ( normalized via UV spectroscopy ) using the High Capacity cDNA Reverse Transcription Kit , according to the manufacturer's instructions ( Applied Biosystems ) . Target cDNA levels were compared by Q-RT-PCR in 20-µl reactions containing 1× SYBR green ( Applied Biosystems ) , 100 µm of each primer , and we used the 7900HT Fast Real-Time PCR System ( Applied Biosystems ) . Results were normalized against glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) expression , unless otherwise indicated . Relative expression was evaluated using the delta-delta-cycle threshold method .
|
We have demonstrated a novel interaction between the Tbx1 gene , the mutation of which causes DiGeorge syndrome , and Wnt5a , another human disease gene , which is important for oriented cell migration and cell polarity . We found that , in mice , reduced dosage of each of the two genes enhances the phenotype caused by the mutation of the other . Loss of the two genes in mice has very severe consequences for heart development . Our genetic and biochemical data determined that Tbx1 , a transcription factor of the T-box family , regulates Wnt5a expression . We found that Tbx1 targets the BAF chromatin remodeling complex to the Wnt5a gene and interacts with a histone monomethyltransferase . Tbx1 expression increases Baf60a occupation of the Wnt5a gene and enhances its H3K4 monomethylation status , while Baf60a knockdown abolishes the ability of Tbx1 to regulate Wnt5a and other target genes . Overall , our data identify Wnt5a as an important effector of Tbx1 function in heart development and demonstrate that Tbx1 regulates the gene by interacting with the chromatin remodeling and histone methylation machinery .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"developmental",
"biology",
"model",
"organisms",
"genetics",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
Transcriptional Control in Cardiac Progenitors: Tbx1 Interacts with the BAF Chromatin Remodeling Complex and Regulates Wnt5a
|
Enteropathogenic and enterohemorrhagic E . coli ( EPEC and EHEC ) are closely related extracellular pathogens that reorganize host cell actin into “pedestals” beneath the tightly adherent bacteria . This pedestal-forming activity is both a critical step in pathogenesis , and it makes EPEC and EHEC useful models for studying the actin rearrangements that underlie membrane protrusions . To generate pedestals , EPEC relies on the tyrosine phosphorylated bacterial effector protein Tir to bind host adaptor proteins that recruit N-WASP , a nucleation-promoting factor that activates the Arp2/3 complex to drive actin polymerization . In contrast , EHEC depends on the effector EspFU to multimerize N-WASP and promote Arp2/3 activation . Although these core pathways of pedestal assembly are well-characterized , the contributions of additional actin nucleation factors are unknown . We investigated potential cooperation between the Arp2/3 complex and other classes of nucleators using chemical inhibitors , siRNAs , and knockout cell lines . We found that inhibition of formins impairs actin pedestal assembly , motility , and cellular colonization for bacteria using the EPEC , but not the EHEC , pathway of actin polymerization . We also identified mDia1 as the formin contributing to EPEC pedestal assembly , as its expression level positively correlates with the efficiency of pedestal formation , and it localizes to the base of pedestals both during their initiation and once they have reached steady state . Collectively , our data suggest that mDia1 enhances EPEC pedestal biogenesis and maintenance by generating seed filaments to be used by the N-WASP-Arp2/3-dependent actin nucleation machinery and by sustaining Src-mediated phosphorylation of Tir .
Bacteria and viruses have historically been useful tools for studying the regulation of cytoskeletal dynamics [1] , as several intracellular pathogens rearrange host actin into comet tails , which propel them through the cytosol [2] and/or promote their transmission from cell-to-cell [3] . Pathogen motility is frequently driven by activation of the Arp2/3 complex , a ubiquitous actin nucleator , through either bacterial [4 , 5] or host [6] actin nucleation-promoting factors ( NPFs ) , although how different classes of nucleators cooperate in cells is not well understood . Enteropathogenic Escherichia coli ( EPEC ) and enterohemorrhagic E . coli ( EHEC ) are also capable of reorganizing host actin via the Arp2/3 complex , but these pathogens remain extracellular to form actin-rich protrusions of the plasma membrane called pedestals [7 , 8] . Actin pedestals promote “surfing” motility [9 , 10] , which is important for cell-to-cell spread [11] . Because EPEC and EHEC activate the host actin nucleation machinery from an extracellular location , they represent ideal models for studying the transmembrane signaling mechanisms , cytoskeletal dynamics , and nucleator cooperation that underlie cellular protrusions [12] . To trigger actin pedestal assembly , EPEC and EHEC both translocate effector proteins into the host cell using a type 3 secretion system ( T3SS ) [13] . One effector , Tir ( translocated intimin receptor ) , adopts a hairpin conformation in the plasma membrane and binds to intimin on the surface of the bacterium , enabling tight attachment of EPEC and EHEC to the plasma membrane [14 , 15] . For EPEC , intimin-induced clustering of Tir triggers phosphorylation of tyrosine residue 474 within its cytoplasmic region by host cell kinases from the Abl/Arg , Src , and Tec families [16–21] . Phosphotyrosine 474 binds the adaptor proteins Nck1 and Nck2 [22 , 23] , which in turn recruit the NPF N-WASP , resulting in actin assembly via the Arp2/3 complex [24 , 25] . EHEC-mediated pedestal biogenesis differs from that of EPEC , because it does not rely on tyrosine phosphorylation or Nck1/Nck2 [14 , 22] . Instead , EHEC Tir binds host BAR proteins including IRTKS [26] and IRSp53 [27] to recruit an additional bacterial effector protein called EspFU [28 , 29] , which multimerizes N-WASP to achieve Arp2/3 complex-driven actin assembly [30–32] . EPEC and EHEC pedestals serve several potential pathogenic purposes , ranging from phagocytosis resistance to epithelial colonization [33–36] . Recently , actin pedestals were also shown to allow the formation of large , two-dimensional bacterial aggregates called macrocolonies [11] . A macrocolony encompasses multiple epithelial cells and appears to originate from a single adherent bacterium multiplying and using Arp2/3-mediated actin-based motility to reach and infect neighboring cells . This series of events allows the bacteria to effectively spread infection without dissociating from the epithelium [11] . Although the Arp2/3 complex is a major actin nucleator in cells , it is becoming clear that other types of actin assembly factors , including formins and tandem actin monomer-binding proteins of nucleation , can cooperate [37 , 38] . Further , it is becoming increasingly apparent that the formation of actin-based cellular structures and the ability to undergo actin-driven processes , such as motility , depend on multiple nucleators [39–41] . For instance , while lamellipodial protrusions are known to rely on Rac-mediated stimulation of WAVE-family NPFs and subsequent activation of the Arp2/3 complex [42] , recent work indicates that important contributions come from the formins mDia1 [43] and FMNL2/3 [44] as well . Despite these findings , the precise mechanisms governing nucleator coordination remain poorly understood . While the lamellipodium represents a valuable model for studying nucleator cooperation , pathogens have also been found to employ multiple nucleators [45] , and therefore have the potential to shed light on how such actin assembly factors collaborate . For instance , it is well established that Listeria monocytogenes activates the Arp2/3 complex using the bacterial NPF ActA [5 , 46 , 47] , that Shigella flexneri uses the bacterial N-WASP activator IcsA [6 , 48] , and that vaccinia virus relies on the viral membrane protein A36 to bind the Nck1/2 and Grb2 adaptors [49 , 50] . However , recent studies have uncovered additional roles for formin nucleators in actin tails and pathogen-associated membrane protrusions . Specifically , protrusion formation and cell-to-cell transmission of Listeria and Shigella were observed to be negatively impacted by the knockdown or inhibition of Diaphanous formins [51 , 52] , suggesting that the formin family of nucleators contributes to the force required for bacterial protrusion into neighboring cells . Furthermore , actin comet tails generated by vaccinia virus were found to rely on the formin FHOD1 in addition to N-WASP and Arp2/3 for actin assembly , motility , and cell-to-cell spread [53] . Formin-mediated actin polymerization was also recently shown to be important for the displacement of septins from vaccinia prior to viral egress [54] , although the formin responsible has not yet been identified . Lastly , Rickettsia parkeri was observed to undergo a switch in motility from Arp2/3 complex dependence early in infection to formin-mediated motility late in infection [55] , employing the bacterial NPF RickA [56 , 57] followed by the formin-like nucleator Sca2 [58] . Taken together , these studies reveal that exploitation of several actin nucleators or actin assembly pathways may be necessary for efficient pathogen-driven actin assembly and cell-to-cell transmission . However , the potential contribution of multiple nucleators to EPEC or EHEC pedestal biogenesis , motility , or cell-to-cell spreading has not been addressed . Moreover , since EPEC and EHEC utilize either phosphotyrosine signaling or direct N-WASP multimerization , they represent important model systems for studying how distinct Arp2/3 complex-associated actin assembly pathways may be coordinated with the activities of additional nucleators at the plasma membrane . In the current study , we examined the roles of formins and other actin nucleators in both EPEC- and EHEC-induced actin rearrangements . Our results reveal a phosphotyrosine-specific mechanism of pedestal assembly in which mDia1 contributes to both initiating and maintaining Arp2/3-dependent actin polymerization .
Because EPEC and EHEC initiate actin assembly using different mechanisms , we aimed to evaluate the contributions of the Arp2/3 complex and other nucleators to each of these signaling cascades . However , EPEC and EHEC have distinct repertoires of effectors and different capacities for infecting cultured cell lines [59 , 60] . So in order to directly compare their pedestal assembly pathways , we employed two well-characterized strains , EPEC Y474* ( referred to hereafter as EPEC ) and KC12+EspFU [11] . EPEC differs from the wild type in that it encodes an HA-tagged version of Tir [22] . KC12+EspFU is an EPEC strain that acts as a surrogate for EHEC because it was engineered to express the EHEC version of intimin , HA-tagged EHEC Tir , and myc-tagged EspFU [28] . Thus , the EPEC and KC12+EspFU strains are isogenic except for their pedestal effectors and can be used to examine the differences in actin assembly pathways . The Arp2/3 complex is thought to be critical for all pathways of actin pedestal assembly by EPEC and EHEC . RNAi-mediated knockdown of the Arp2/3 complex or overexpression of the N-WASP WCA domain , which has a dominant negative effect by sequestering and/or ectopically activating Arp2/3 , reduces pedestal formation by both EPEC and EHEC [24 , 61] . N-WASP is essential for EPEC pedestal assembly [25 , 61] , and although some N-WASP-deficient mouse cells do not support EHEC pedestal assembly [62] , others can form pedestals when EHEC Tir and EspFU are either delivered by KC12 or directly expressed in the knockout cells [61] . Therefore , we expect inhibition of either the Arp2/3 complex or N-WASP to completely block or diminish pedestal assembly by EPEC as well as KC12+EspFU . The roles of other nucleators , like formins , are unknown in the context of EPEC or EHEC infections . To initially explore the contributions of the Arp2/3 complex , N-WASP , and formins to actin assembly in pedestals , HeLa cells were pretreated with either DMSO as a control , or the Arp2/3 inhibitors CK666 and CK869 [63] , the N-WASP inhibitor Wiskostatin [64] , and/or the broad formin inhibitor SMIFH2 [65] . The cells were then infected with EPEC or KC12+EspFU for 3 . 5 h and stained to detect HA-Tir , F-actin , and DNA ( Fig 1A ) . The fraction of bacteria that translocated Tir and formed a pedestal was assessed , and the F-actin intensity at the locations of HA-Tir staining was quantified and normalized to an adjacent Tir-free area of the cell to determine the relative F-actin levels beneath the bacteria . Treatment with CK666+CK869 caused a 33% reduction in the average percentage of EPEC associated with pedestals , and a 60% reduction in KC12+EspFU associated with pedestals ( Fig 1B and 1C ) . Furthermore , Arp2/3 complex inhibition resulted in significantly dimmer pedestals than DMSO-treated controls for both strains ( Fig 1D and 1E ) . Wiskostatin had similar effects on EPEC pedestals but did not cause as severe of a reduction in the fraction of KC12+EspFU associated with pedestals ( Fig 1B and 1C ) . Collectively , the results from these pharmacological studies are consistent with previous functional analyses demonstrating that N-WASP-Arp2/3 complex-driven pathways of actin assembly are important for pedestal biogenesis . Interestingly , inhibition of formins using SMIFH2 caused a 20% reduction in the frequency of pedestal formation and significantly dimmer pedestals for EPEC but not for KC12+EspFU , which was generally unaffected by SMIFH2 treatment ( Fig 1B–1E ) . Furthermore , the reduction in EPEC pedestal intensity with Arp2/3 complex inhibition was exacerbated by simultaneous formin inhibition ( Fig 1D ) . Other treatment combinations did not strengthen any of the deficiencies in pedestal formation or intensity . These results provide the first evidence , to our knowledge , that formins may be involved in EPEC pedestal assembly . Actin pedestal-based motility is important for cell-to-cell transmission , and EPEC surfing has been shown to rely heavily on the ability of Tir to become phosphorylated at tyrosine 474 [11] , presumably to trigger a Nck1/2-N-WASP-Arp2/3 complex actin polymerization pathway . To determine if SMIFH2 treatment impacts pedestal motility , cells stably expressing mCherry-actin were infected , treated with inhibitors , and subjected to live imaging . Bacteria associated with actin pedestals were tracked over time , and pedestal speeds were calculated using movies spanning 20–30 min . EPEC pedestals moved on DMSO-treated cells at an average speed of 1 . 02 µm/min , with individual pedestal speeds ranging from 0 . 40–2 . 09 µm/min ( Fig 2A , left ) . Treatment with CK666+CK869 reduced the average speed by more than half , to 0 . 47 µm/min ( range: 0 . 21–0 . 64 µm/min ) , and Wiskostatin resulted in a similar reduction in average speed to 0 . 53 µm/min ( range: 0 . 24–1 . 10 µm/min ) . SMIFH2 treatment also significantly inhibited motility , although to a lesser degree , as the average speed was reduced by 28% to 0 . 73 µm/min ( range: 0 . 34–1 . 13 µm/min ) ( Fig 2A , left ) . Similar to the results in Fig 1 , KC12+EspFU motility was only impacted by inhibition of the Arp2/3 complex or N-WASP , and not by SMIFH2 treatment ( Fig 2A , right ) . These results suggest that formin-mediated actin polymerization contributes to the motility of EPEC pedestals , but not EHEC pedestals . Because motility positively correlates with macrocolony size and epithelial colonization [11] , we next sought to determine if macrocolony development was impacted by treatment with the various Arp2/3 , N-WASP , or formin inhibitors . Polarized Caco2 cell monolayers were thus pretreated with DMSO or inhibitors and infected for 6 h , with hourly washes and media changes to promote colonization only from initially adherent bacteria . Monolayers were then fixed , stained , and imaged at a low magnification to visualize macrocolonies . In line with previous results [11] , EPEC consistently formed smaller macrocolonies than KC12+EspFU ( Fig 2B ) . For EPEC infections , treatment with CK666+CK869 , Wiskostatin , or SMIFH2 individually limited macrocolony size to some extent , while pairwise combinations of CK666+CK869 with SMIFH2 , or Wiskostatin with SMIFH2 resulted in a statistically significant reduction in macrocolony size ( Fig 2C ) . In contrast , KC12+EspFU colonies were unaffected by SMIFH2 treatment , and combining SMIFH2 with either CK666+CK869 or Wiskostatin did not further the deficiencies in colony size beyond what was observed with Arp2/3 or N-WASP inhibition alone ( Fig 2D ) . These data suggest that KC12+EspFU macrocolony size is largely dictated by the N-WASP-Arp2/3 complex pathway of actin assembly , whereas cooperation between the Arp2/3 complex and formins promotes colonization by EPEC . SMIFH2 is a broad inhibitor of actin nucleation by formin FH2 domains [65] , so to determine which specific formins could be contributing to EPEC pedestal assembly , we performed a small scale survey of formin function using pairs of siRNAs to the formins that are expressed in HeLa cells , namely DAAM1 , FHOD1 , FMNL1 , FMNL2 , INF2 , mDia1 , mDia2 , and mDia3 . In addition to targeting these formins , we used siRNAs to the tandem actin monomer-binding proteins of nucleation Cordon-bleu ( Cobl ) , adenomatous polyposis coli ( APC ) , Spire1 , and Spire2 . We also examined additional Arp2/3 complex interacting proteins in our screen , including Cortactin ( CTTN ) , which was previously reported to contribute to EPEC and EHEC pedestal formation [66 , 67] , WISH/SPIN90/DIP1 , which activates the Arp2/3 complex to promote polymerization of unbranched filaments [68] , and JMY , a WASP-family nucleation-promoting factor that can also nucleate actin directly [69] . Lastly , we included two proteins that might influence tyrosine kinase signaling to formins—the GTPase dynamin II ( DynII ) , and the scaffolding protein IQGAP1 . DynII contributes to signaling in EPEC pedestals [70] , and was recently shown to promote a formin-mediated mechanism of septin displacement from vaccinia virus [54] . IQGAP1 localizes to EPEC pedestals , and actin pedestal formation in IQGAP1-deficient MEFs is reduced by about 40% [71] . Further , in vitro experiments suggest that IQGAP1 is capable of binding both to EPEC Tir and to mDia1 [71 , 72] . On control siRNA-treated HeLa cells , 90% of EPEC and 86% of KC12+EspFU generated pedestals and , as expected , siRNAs targeting the Arp2/3 complex or N-WASP significantly diminished pedestal formation by both strains by 32–48% ( Fig 3A–3C ) . In agreement with previous studies , siRNAs to Cortactin negatively impacted the EspFU-dependent pathway of actin polymerization [66] , however EPEC pedestals were unaffected . Targeting of JMY , WISH , APC , Cobl , Spire1 , or Spire2 did not cause any significant defects in pedestal biogenesis by either strain ( Fig 3B and 3C ) . Among the formins , targeting of DAAM1 resulted in a modest ( 10% ) reduction in EPEC pedestal formation , while targeting of mDia1 ( also called DIAPH1 or hDia1 ) caused a more obvious inhibition of pedestal assembly , reflected in an approximately 25% reduction in pedestal formation frequency ( Fig 3B ) . Although DynII and IQGAP1 would be candidates for promoting an interaction between EPEC Tir and mDia1 , we did not observe a measurable defect in pedestal assembly when testing these factors in our screen . Because targeting mDia1 resulted in the same EPEC-specific actin assembly defects that arose with SMIFH2 treatment , we investigated the contributions of mDia1 to EPEC pedestals further . Independent siRNAs targeting mDia1 were each effective at depleting cellular mDia1 levels when assessed by immunofluorescence microscopy ( Fig 4A and 4B ) or western blotting ( Fig 4C ) . Each individual mDia1 siRNA also reduced pedestal formation by EPEC by over 30% , but neither one affected pedestal assembly by KC12+EspFU ( Fig 4A and 4D ) . To more clearly relate cellular mDia1 levels to pedestal formation efficiency , the percentage of EPEC or KC12+EspFU that had successfully formed pedestals on control or mDia1-depleted cells was plotted against the mDia1 staining intensity in those cells ( Fig 4E and 4F ) . For EPEC , the amount of mDia1 present in the cell positively correlated with the percentage of bacteria forming pedestals , but KC12+EspFU formed pedestals more than 60% of the time regardless of mDia1 levels . Finally , to more closely quantify the amount of F-actin that was associated with Tir in mDia1-depleted cells , the phalloidin staining intensity was plotted along a ~3 µm line through the pedestal-forming region , and the brightest pixel in the HA-Tir channel was set to a distance of 0 to compare the intensities across pedestals ( Fig 4G and 4H ) . In control siRNA-treated cells , actin pedestals were strong and peaked immediately adjacent to HA-Tir . However , targeting the Arp2/3 complex or mDia1 diminished this peak in actin intensity , resulting in values that were less than half of those in control cells . These data indicate that EPEC can only assemble pedestals efficiently when mDia1 is present in the host cell . Because mDia1 plays a positive role in EPEC pedestal formation , it would make sense if this protein was present within the pedestal . In fact , endogenous mDia1 could be observed in a subset of pedestals in the control siRNA-treated cells described above ( Fig 4A ) . To more closely assess mDia1 localization , HeLa cells infected with EPEC were fixed and treated with antibodies to detect mDia1 and HA-Tir , as well as phalloidin to visualize F-actin , and examined by confocal microscopy . In parallel , HeLa cells transiently expressing GFP-mDia1 were infected , fixed , and stained for HA-Tir and F-actin . Similar mDia1 recruitment to pedestals was observed with both antibody staining and with the GFP-tagged protein ( Fig 5A ) . To characterize this recruitment more quantitatively , we plotted the pixel intensity profiles of HA-Tir , F-actin , and mDia1 staining along the lengths of several pedestals to determine the position of mDia1 in relation to Tir . On average , F-actin intensity peaked 0 . 19 µm after Tir , while mDia1 staining peaked 0 . 13 µm after the F-actin peak ( Fig 5B ) . This implies that the actin associated with mDia1 in pedestals is further from the bacterium than the actin nucleated by Arp2/3 complex , which is thought to localize throughout the pedestal [61] ( and see below ) . Because mDia1 was not enriched in all EPEC pedestals , we quantified the fraction of pedestals that showed distinguishable mDia1 antibody staining . Additionally , we compared this value to that for KC12+EspFU pedestals , which do not rely on mDia1 during actin polymerization . Pedestals were scored as mDia1 positive or negative , and the average percentage of bacteria in each category was calculated per cell . mDia1 was enriched in 43% of pedestals generated by EPEC and 31% of pedestals formed by KC12+EspFU , and this slight preference for EPEC pedestals was statistically significant ( Fig 5C ) . To determine if mDia1 recruitment was dependent on the known bacterial components that drive pedestal assembly , we employed two pedestal-deficient mutants . For disrupting the main pathway of pedestal biogenesis by EPEC , we used a strain that encodes an HA-tagged version of Tir with a Y474F point mutation that prevents Nck-mediated signaling to the actin assembly machinery ( EPEC Y474F ) [22] . To interrupt the mechanism used by EHEC , we employed a strain of KC12 lacking EspFU ( KC12+vector ) [28] . HeLa cells were infected with these bacteria in addition to their pedestal-proficient counterparts , fixed , and stained to visualize HA-Tir , mDia1 , F-actin , and DNA ( Fig 5D ) . EPEC expressing wild type ( WT ) Tir and KC12+EspFU both formed bright pedestals that had some level of mDia1 enrichment , which was reflected in mDia1 pixel intensity profiles associated with adherent bacteria ( Fig 5D and 5E ) . However , EPEC Y474F and KC12+vector did not assemble pedestals or recruit detectable levels of mDia1 ( Fig 5D and 5E ) . These findings suggest that for EPEC , mDia1 recruitment is reliant on a phosphotyrosine 474-dependent pathway of actin polymerization , while for EHEC , EspFU-mediated actin assembly can also permit some degree of mDia1 localization to pedestals . Because KC12+EspFU pedestals can recruit mDia1 but are unaffected by SMIFH2 or mDia1 siRNAs , we reasoned that EspFU could allow the bacteria to bypass a dependency on mDia1 . To confirm this possibility , we used a strain co-expressing EPEC Tir and EHEC EspFU-myc [11] to infect cells treated with siRNAs targeting mDia1 . Cells that avoided knockdown and retained mDia1 expression were easily discernible from nearby mDia1-depleted cells when stained with mDia1 antibodies . Side-by-side comparisons of these mDia1 expressing versus depleted cells revealed that the actin pedestals formed by EPEC+EspFU-myc in the presence of mDia1 ( Fig 5F , “+” inset; mean fluorescence: 4347 +/-402; n = 15 pedestals , 3 cells ) , were indistinguishable from the pedestals that these bacteria formed in the absence of mDia1 ( Fig 5F , “-” inset; mean fluorescence: 4359 +/-410; n = 15 pedestals , 3 cells ) . Thus , EspFU is sufficient to confer EHEC’s robust mDia1-independent pedestal-forming ability to EPEC . We next sought to determine if the contribution of mDia1 to EPEC pedestals is dependent on or independent of the Arp2/3 complex . Because cells treated with chemical inhibitors or siRNAs to target the Arp2/3 complex were still capable of forming pedestals beneath about 50% of bacteria , it was unclear if mDia1 was responsible for this degree of pedestal formation , or if the ability to make pedestals under these conditions was due to residual Arp2/3 complex activity . To differentiate between these possibilities , we infected cells completely lacking the Arp2/3 complex . ArpC2 knockout ( KO ) mouse fibroblasts [73] were generated by treating ArpC2 Flox cells with 4-hydroxy-tamoxifen for 6 days to delete arpC2 and fully deplete the Arp2/3 complex . These cells were then compared to DMSO-treated ArpC2 Flox control cells during infection . In striking contrast to the >90% of adherent EPEC and KC12+EspFU which formed pedestals in Arp2/3-proficient cells , 0% of bacteria triggered pedestal assembly in the Arp2/3-deficient cells ( Fig 6A , n = 324–344 Tir+ bacteria ) . These results using ArpC2 KO cells indicate that the pedestals which formed during CK666+CK869 or ArpC4 siRNA treatment relied on residual Arp2/3 complex activity , and therefore any effect of mDia1 on EPEC pedestals still requires N-WASP [25 , 61] and the Arp2/3 complex . Although we observed no instances of structures resembling actin pedestals in the KO cells , adherent EPEC were occasionally associated with very small and weak actin puncta ( Fig 6A , arrow ) or intense F-actin staining in basket-like structures around clusters of translocated Tir ( Fig 6A , bottom row ) . While the origin of the filaments comprising these rare actin baskets remains unclear , we tested whether mDia1 might be responsible for generating the Tir-associated actin puncta by treating ArpC2 Flox and KO cells with mDia1 siRNAs prior to infection with EPEC . Immunoblotting confirmed that the mDia1 siRNAs depleted mDia1 in both cell lines , and that ArpC2 levels were undetectable in the KO cells ( Fig 6B ) . Furthermore , RNAi-mediated depletion of mDia1 prevented the formation of actin puncta in the vicinity of Tir in the KO cells ( Fig 6C and 6D ) . These observations are consistent with the idea that even though mDia1 is insufficient to assemble mature actin pedestals in the absence of the Arp2/3 complex , it is nevertheless capable of promoting some degree of actin polymerization in response to signaling from EPEC Tir . Given that the Arp2/3 complex and mDia1 both contribute to EPEC pedestal formation , we next sought to more precisely characterize the localization and timing of their recruitment by EPEC . Therefore , we used immunostaining to visualize Arp3 and mDia1 after HeLa cells were infected with EPEC for 4 h , when most pedestals have reached steady-state in frequency , size , and motility . Consistent with earlier observations ( Fig 5B ) , pixel intensity plots demonstrated that mDia1 was enriched closer to the base of pedestals ( Fig 7A and 7B ) , while Arp3 was found throughout pedestals and was more abundant closer to the bacteria ( Fig 7A and 7B ) . By calculating the distance from the center of the bacteria to the maximum intensities for F-actin , Arp3 , and mDia1 , we found that , on average , F-actin and Arp3 both peak at 0 . 5 µm away from the bacteria , while mDia1 peaks at about 0 . 8 µm from the bacteria ( Fig 7C ) . This pattern of mDia1 recruitment to the less actin-dense distal regions of pedestals was also observed in mouse ArpC2 Flox cells ( Fig 7D , arrowhead ) , demonstrating that the precise localization of this nucleator is conserved in multiple host species and cell types . To test whether mDia1 is recruited by EPEC Tir in the absence of the Arp2/3 complex , we also stained ArpC2 KO cells for mDia1 . In contrast to its strong pedestal base-enriched localization in Flox cells , mDia1 did not localize near Tir in the vast majority of cells ( Fig 7D , top KO row ) . However , in rare instances where weak or unfocused actin was observed near Tir , mDia1 was slightly enriched ( Fig 7D , arrow ) . Such actin recruitment was not found in cells treated with mDia1 siRNAs ( Figs 6C and 6D and 7D ) . These results support a function for mDia1 in the generation of seed filaments that are utilized by the Arp2/3 complex to form bona fide pedestals . Lastly , to examine the temporal arrival of Arp3 and mDia1 during EPEC pedestal assembly , we synchronized the initiation of actin polymerization in infected cells using prime-challenge experiments . HeLa cells were first treated ( “primed” ) with a mutant EPEC strain that translocates Tir but lacks intimin ( EPECΔeae ) , and then infected ( “challenged” ) with a laboratory strain of E . coli expressing intimin ( E . coli+pIntimin ) . This approach allows for rapid pedestal assembly within minutes of the clustering and tyrosine phosphorylation of Tir that is induced by the intimin-expressing strain [21] . After a 10 minute challenge , cells were fixed and examined microscopically . Both mDia1 and Arp3 localized to pedestals that had formed beneath E . coli+pIntimin ( Fig 7E , arrowhead ) , but neither protein was found in association with bacteria in the absence of a pedestal ( Fig 7E , arrow ) . As described above , while both proteins were present in pedestals , they did not colocalize , and Arp3 was interspersed in areas with intense F-actin staining that lacked mDia1 . These results indicate that mDia1 , like the Arp2/3 complex , is recruited to EPEC pedestals very early in their biogenesis , consistent with a role for mDia1 in the initiation of pedestal assembly . When taken together with steady-state localization data , our findings suggest that mDia1 functions in both initiating and maintaining actin assembly in pedestals , but that it does so in a manner that is spatially distinct from the Arp2/3 complex . To better understand the mechanisms by which mDia1 contributes to pedestal formation , beyond simply polymerizing actin filaments that could be incorporated into pedestals , we next examined the influence of mDia1 in the signaling pathway that lies upstream of Arp2/3-mediated actin assembly . Because the most fundamental difference between the mechanisms of actin pedestal formation by EPEC and KC12/EHEC is that EPEC relies specifically on phosphotyrosine 474 in Tir , we evaluated the status of Tir phosphorylation in HeLa cells treated with control or mDia1 siRNAs . As expected based on previous quantifications [21 , 74] , immunofluorescence using antibodies to HA-Tir and phosphotyrosine ( “pY” ) in control cells revealed that virtually all sites of translocated Tir colocalized with bright phosphotyrosine staining ( Fig 8A ) . In contrast , in cells treated with mDia1 siRNAs , phosphotyrosine colocalization with Tir appeared to be somewhat less frequent , and when it did colocalize with Tir , the intensity of phosphotyrosine staining was noticeably lower ( Fig 8A ) . Quantification of the ratio of pY to HA-Tir intensity in pedestal-forming regions indicated that Tir-associated tyrosine phosphorylation was generally 30% lower in cells treated with independent mDia1 siRNAs compared to cells treated with control siRNAs ( Fig 8B ) . Further , due to cell-to-cell variability in mDia1 silencing , we were able to visualize mDia1-expressing ( Fig 8C , “+” ) and mDia1-depleted ( Fig 8C , “-” ) cells in the same field of view . Such images illustrated the fact that bacteria on mDia1-positive cells formed intense actin pedestals containing phospotyrosine and mDia1 ( arrowhead ) , while mDia1-negative cells lacked phosphotyrosine staining and actin pedestal assembly ( arrows ) . These results support an unexpected role for mDia1 in promoting tyrosine phosphorylation of Tir . Because EPEC-associated tyrosine phosphorylation was influenced by the presence or absence of mDia1 , we aimed to test whether mDia1 may also be involved in the activation of the kinases that phosphorylate Tir . Multiple tyrosine kinases can phosphorylate Tir , including Arg , Abl , and Etk [18 , 20] , as well as the Src-family kinase ( SFK ) Fyn [17 , 19] . Additionally , Src-family members have been shown to interact with several Diaphanous-related formins , including DAAM1 , mDia1 , and mDia2 [75–78] , and mDia1 specifically has been found to affect the subcellular targeting of Src [79 , 80] . Therefore , we next probed siRNA-treated cells with an antibody that recognizes active Src-family members , which are characterized by phosphorylation at tyrosine 416 in Src ( “pSrc” ) or equivalent residues in other SFKs . Consistent with the phosphotyrosine and Tir colocalization results described above , the staining of active SFKs in proximity to Tir was noticeably weaker in cells treated with mDia1 siRNAs than in control cells ( Fig 8D ) . Quantification of the ratio of pSrc to HA-Tir intensity revealed that active SFK levels were indeed lower in the pedestal-forming regions of cells treated with mDia1 siRNAs , although data for one of the siRNAs did not reach statistical significance ( Fig 8E ) . To further validate our microscopy-based findings on the effects of mDia1 on Tir phosphorylation and SFK activation , siRNA-treated cells that were infected in parallel to those described above were collected , lysed , and subjected to immunoblot analyses for measuring mDia1 , pTir , HA-Tir , pSrc , Src , actin , and GAPDH levels ( Fig 8F ) . In accordance with our previous quantifications of mDia1 depletion in uninfected cells ( Fig 4C ) , mDia1 protein levels were reduced by 75% , on average , in infected cells ( Fig 8G ) . Strikingly , even though EPEC delivered the mature 90 kDa form of Tir equivalently into cells treated with different siRNAs , the tyrosine phosphorylation of Tir was clearly lower in the mDia1-depleted samples ( Fig 8F ) . In addition , while infection with EPEC caused a major activation of SFKs in control cells ( Fig 8F; compare uninfected lane 1 to infected lane 4; uninfected pSrc:Src ratio = 1 , EPEC WT pSrc:Src ratio = 4 . 99 +/- 0 . 65 , n = 6 blots ) , this activation was markedly dampened in the mDia1-depleted samples ( lanes 5–6 ) . It is important to note that total Src levels were consistent across all conditions and that infections with EPEC expressing the Tir Y474F mutant confirmed both the specificity of Tir tyrosine phosphorylation and its requirement for SFK activation ( Fig 8F , lane 7 ) . Finally , to quantify our biochemical observations , we calculated the relative efficiency of Tir phosphorylation ( i . e . , Tir tyrosine phosphorylation index ) and relative magnitude of Src activation ( i . e . , Src activation index ) across multiple experiments and blots . For assessing Tir phosphorylation , the band intensity of pTir was divided by total HA-Tir . In agreement with our microscopy results , Tir phosphorylation was approximately 50% lower in mDia1-depleted cells than in control cells ( Fig 8H ) . To further evaluate the relationship between cellular levels of mDia1 and phosphorylated Tir , we additionally plotted the mDia1 densitometry data against the Tir phosphorylation index for matched samples , and we found that the two values were positively correlated ( Fig 8I ) . For measuring the relative magnitude of SFK activation in EPEC-infected cells , we calculated the band intensity ratio of pSrc to total Src . We found that SFKs were the most active in infected cells expressing mDia1 , significantly less active in infected mDia1-depleted cells , and basically not activated at all when cells were treated with the Y474F mutant instead of wild type Tir ( Fig 8J ) . Collectively , these results indicate that mDia1 is important for SFK activation , which in turn allows for an efficient and persistent phosphorylation of Tir during actin pedestal assembly .
Pathogens such as Listeria and Shigella are often employed as tools to better understand actin dynamics and uncover new pathways and regulators of actin assembly , yet their utility for modeling actin polymerization at the plasma membrane is limited by the fact that they are cytosolic . By remaining extracellular throughout infection , EPEC and EHEC represent ideal models to study actin rearrangements triggered by transmembrane signaling cascades [12] . While the core pathways of EHEC and EPEC pedestal assembly have been characterized to some degree [81] , the potential contributions of actin nucleation factors outside of the Arp2/3 complex and WASP-family have never been directly assessed . Given that the coordinated actions of multiple nucleators orchestrate a variety of cellular functions , including lamellipodia formation and cell motility [39–44] , and that both the Arp2/3 complex and formins participate in pathogen-induced protrusions [51–53 , 82] , we examined whether some form of nucleator cooperation exists in EPEC and EHEC pedestals . Our results indicate that the formin mDia1 contributes to Arp2/3 complex-mediated actin assembly in the pedestals of EPEC but not EHEC . Our findings also support a model in which mDia1 participates in the biogenesis and maintenance of EPEC pedestals by both providing filaments that can be used by the Arp2/3 complex for branched nucleation ( Fig 9A ) and by promoting tyrosine kinase activation and Tir phosphorylation ( Fig 9B ) . Our first evidence for formin activity in pedestals came from the use of SMIFH2 , which resulted in EPEC-specific pedestal phenotypes , as fewer bacteria formed pedestals , and the pedestals that did form contained less F-actin . These defects are consistent with a role for formins in pedestal biogenesis . Further , SMIFH2 treatment resulted in significantly slower actin-based motility , suggestive of a deficiency in pedestal maintenance and force generation . Lastly , formins also had an apparent role in colonization , as inhibitors of the Arp2/3 complex , N-WASP , or formins did not impact EPEC colony size by themselves , but simultaneously inhibiting both the N-WASP-Arp2/3 machinery and formin nucleators reduced macrocolony size . Taken together , these experiments suggest that collaboration between the Arp2/3 complex and formins is important for pedestal initiation , continuous actin assembly , and EPEC cell-to-cell spreading during colonization . Using a small siRNA screen , we identified mDia1 as the formin most likely responsible for the EPEC pedestal defects that were observed with SMIFH2 . Although targeting DAAM1 also resulted in a decrease in pedestal formation , this phenotype was not as strong or as significant as the one caused by the depletion of mDia1 . Further , we found a positive correlation between the cellular level of mDia1 and the fraction of bacteria generating pedestals . These results parallel the findings that Listeria monocytogenes and Shigella flexneri each rely on Diaphanous formins ( mDia1 , mDia2 , and/or mDia3 ) in addition to the Arp2/3 complex for protrusion formation and cell-to-cell transmission [51 , 52] . Somewhat surprisingly , EPEC did not show any phenotype when FHOD1 was targeted in the siRNA screen . This was unexpected because vaccinia virus , which triggers a similar phosphotyrosine-dependent signaling cascade to EPEC , was found to manipulate Rac1 and FHOD1 for actin tail assembly , motility , and cell-to cell spreading [53] . Canonical vaccinia actin polymerization relies on tyrosine phosphorylation of the viral membrane protein A36 by host cell kinases [49] . Phosphorylated Y112 binds the Nck adaptor proteins , which recruit an N-WASP-WIP complex [50 , 83] . Phosphorylation of a second residue , Y132 , promotes the recruitment of another adaptor , Grb2 , which may contribute to N-WASP activation or stability in the tail [50 , 84] . These mechanisms of actin assembly are strikingly similar to the pathways of actin polymerization promoted by EPEC Tir , which is phosphorylated on two similarly spaced residues , Y474 and Y454 . Although EPEC does not recruit Grb2 [74] , phosphorylated Y474 recruits Nck1 and Nck2 , which bind and activate N-WASP , with or without WIP [22 , 23 , 85] . It is possible that Grb2 somehow promotes FHOD1 recruitment in the case of vaccinia virus , potentially explaining why EPEC does not employ this nucleator . It is also possible that other vaccinia proteins or EPEC effectors positively or negatively influence FHOD1 localization and function . Interestingly , a recent study implicated another formin , which has yet to be identified , in overcoming septin-mediated vaccinia entrapment [54] . Based on our findings , it is plausible that mDia1 is recruited to vaccinia virus and contributes to this ability to evade entrapment and promote egress . Among pathogenic E . coli , the formin-related changes in pedestal assembly were exclusive to EPEC , as neither SMIFH2 treatment nor siRNA targeting of formins decreased pedestal formation by KC12+EspFU . Despite this , KC12+EspFU was still capable of recruiting mDia1 to pedestals , albeit less frequently than EPEC . Therefore , it seems plausible that mDia1 contributes to EHEC pedestal assembly , but that its effects are dwarfed by the activity of EspFU , a multivalent effector protein capable of activating multiple N-WASP molecules to achieve extraordinarily high levels of Arp2/3 complex activation [30–32] . In agreement with this idea , the requirement for mDia1 in EPEC pedestal formation was able to be bypassed by simply introducing EspFU into EPEC . To understand the mechanisms underlying the contributions of mDia1 to EPEC pedestals , we examined some of the bacterial and host factors that could influence , or be influenced by , mDia1 function . On the bacterial side , we employed the pedestal-deficient strains EPEC Y474F and KC12+vector and found that neither of these mutants efficiently recruited mDia1 . This suggests that mDia1 could possibly localize passively to at least a subset of pedestals simply because they are rich in filaments and other actin-associated factors . However , our observation that EPEC Tir-containing pedestals were slightly better than EspFU-derived pedestals at recruiting mDia1 leads us to speculate that Tir phosphotyrosine 474 itself , or some signaling molecule associated with this residue , is actively involved in recruiting mDia1 . In the course of our siRNA screen , we explored whether several such host cell proteins could mediate mDia1 enrichment in EPEC pedestals . One candidate was WISH , which is capable of interacting with Nck1 and Nck2 , as well as mDia1 and the Arp2/3 complex [68 , 86 , 87] . However , our targeting of WISH with siRNAs did not cause any pedestal phenotypes . Another interesting candidate was IQGAP1 , a scaffolding protein that can activate N-WASP and localize to EPEC pedestals [71 , 88] . However , our targeting of IQGAP1 in HeLa cells also did not cause any pedestal phenotypes . Whether other factors are able to physically link EPEC Tir to mDia1 in pedestals remains an open question . In investigating the molecular basis of mDia1 function in EPEC pedestals downstream of Tir Y474 , we focused on its ability to polymerize actin and to affect tyrosine kinase signaling . First , to assess its potential cytoskeletal activities , we examined mDia1 localization and actin filaments in the presence and absence of the Arp2/3 complex . In Arp2/3-proficient cells , mDia1 and Arp3 exhibited clearly distinct localization patterns , with Arp3 found throughout pedestals and more abundant closer to the bacteria , whereas mDia1 was concentrated closer to the base of pedestals . In many instances ( e . g . , Fig 7 ) , F-actin staining was more intense in the Arp3-enriched region , while a weaker F-actin haze was in the more distal mDia1-associated area ( most noticeable in Fig 7D , arrowhead ) . These observations are consistent with the hypothesis that mDia1 may be providing linear seed filaments upon which the Arp2/3 complex can nucleate densely branched actin networks ( Fig 9A ) . In further support of this model are our findings with cells engineered to genetically lack the Arp2/3 complex . In ArpC2 KO cells , EPEC was unable to generate any actin pedestals , confirming the essentiality of the Arp2/3 complex in pedestal assembly . Notably , however , EPEC were sometimes associated with weak F-actin puncta and diffuse mDia1 staining in the KO cells . Such filaments were not observed when mDia1 was depleted , implying that mDia1 can polymerize sparse , unfocused actin filaments near sites of EPEC adherence that could be utilized for branched nucleation if Arp2/3 is present ( Fig 9A ) . Importantly , the coordination of multiple families of nucleators either directly , as in the case of Spire 1 and Formin 2 in vitro [38] , or indirectly , as with the multiple assembly factors that operate in lamellipodia [41–44] , is an emerging theme in the field of cytoskeletal biology . If mDia1 is indeed providing mother filaments that the Arp2/3 complex can use as seeds for branching and nucleation , the EPEC system might be analogous to studies suggesting that mDia1 and the Arp2/3 complex collaborate by acting sequentially in lamellipodia [43] . Given the fact that we cannot temporally separate the arrival of mDia1 and Arp2/3 at sites of EPEC Tir phosphorylation , our model still requires further investigation . Alternatively , in a fashion similar to mDia2 in the lamellipodia and filopodia of melanoma cells [41] , formins may prevent the capping of Arp2/3 complex-nucleated filaments and promote their elongation . However , this seems less likely to take place in the EPEC system due to a lack of mDia1 enrichment at pedestal tips . Lastly , the possibility that both nucleation pathways function independently , in a manner similar to the suggested contributions of FMNL2/3 to Arp2/3-mediated nucleation in lamellipodia [44] , is also unlikely due to the complete absence of pedestals in ArpC2 KO cells . Apart from polymerizing actin filaments that could be incorporated into EPEC pedestals , we also examined whether the molecular mechanism of mDia1 function was related to other activities within pedestals . Tir can be phosphorylated by multiple tyrosine kinases in cells and in vitro [17–20] , and we focused on the Src family , because Y474 phosphorylation by SFKs is known to be induced by Tir clustering [17 , 19] . Diaphanous-related formins have been shown to physically interact with SFKs [75–77] , and mDia1 itself has been implicated in the subcellular targeting of Src to the cell periphery or to focal adhesions [79 , 80] , while another study suggests that Src acts upstream of mDia1 at cell junctions [78] . When we assessed the efficiency of SFK activation and Tir Y474 phosphorylation in control or mDia1-depleted cells , we found that the presence of mDia1 positively correlates with increased SFK activity and Tir phosphorylation , thereby revealing that one of the crucial functions of mDia1 in pedestal formation is to promote SFK activation . Interestingly , EPEC encodes an effector protein , EspJ , that can act as a tyrosine kinase inhibitor [89 , 90] . Thus , deciphering the interplay among Tir , mDia1 , SFKs , and EspJ during the course of infection will be an important topic for future study . Building upon several previously-established signaling pathways [81 , 91] , and in light of our new data , we propose an updated model describing the mechanisms that drive EPEC pedestal assembly ( Fig 9B ) . In the core model: ( 1 ) Intimin-mediated clustering of Tir triggers Y474 phosphorylation [16 , 21] via Src , Abl , and Tec family kinases [17–20] . ( 2 ) This enables Tir binding to the SH2 domains of Nck1 and Nck2 [22 , 23] . ( 3 ) These adaptors in turn cause N-WASP activation directly via their SH3 domains and linker regions [92–95] or through accessory factors . ( 4 ) Active N-WASP then promotes actin branching and nucleation by the Arp2/3 complex to create pedestals [24 , 25] . Our current work places mDia1 at several key positions in this model ( Fig 9B ) . First , Y474 induces mDia1 recruitment to the general vicinity of adherent bacteria by a yet-to-be determined mechanism . Since mDia1 does not colocalize with Tir , mDia1 is likely recruited indirectly through a host cell signaling cascade possibly involving SFKs . The actin nucleation activity of mDia1 then assembles unfocused linear filaments that are unable to be organized into a pedestal in the absence of the Arp2/3 complex . However , Nck-mediated activation of the N-WASP-Arp2/3 branching and nucleation machinery results in the repurposing of those seed filaments into bona fide actin pedestals , wherein Nck and N-WASP are located in contact with Tir , Arp2/3 is enriched in the proximity of Tir but resides at branches throughout the pedestal network , and mDia1 remains at the distal end of the pedestal . Much like the undefined mechanism of initial mDia1 recruitment , how mDia1 is retained at the pedestal base remains an open question . Nevertheless , the pedestal deficiencies that we observed in mDia1-depleted cells imply that the ability of mDia1 to provide an early and consistent supply of seed filaments is important for building and maintaining pedestals . Proper pedestal biogenesis and maintenance clearly also rely on the capacity of mDia1 to promote SFK activation and efficient Y474 phosphorylation . Prior to EPEC infection , a housekeeping function of mDia1 may keep SFKs in a proper subcellular environment or activatable form to enable the initial phosphorylation of Tir . Additionally , during the course of infection , it seems likely that mDia1 participates in a positive feedback loop that reinforces SFK activation and sustains Tir phosphorylation in order to provide continuous signaling to Nck , N-WASP , and Arp2/3 . The fact that mDia1-depleted cells harbor both fewer pedestals and dimmer pedestals is consistent with mDia1 normally supporting the initial and persistent phosphorylation of Tir and perhaps other pedestal components . Although our results have provided new insights into actin nucleator collaboration and the cellular mechanisms underlying plasma membrane protrusions by using EPEC as a tool , it is also imperative to revisit the fact that EPEC and EHEC are human pathogens which cause severe diarrheal diseases . Pedestal formation is an important step in EPEC and EHEC pathogenesis , as their abilities to manipulate actin enhances colonization in animal and cell culture models [11 , 33 , 35 , 36] . Therefore , continuing to improve our understanding of the pathways controlling actin assembly could lead to advances in potential therapies . For example , our current findings may renew interest in deciphering how Src-family kinases operate in pedestals and revitalize investigations in the use of tyrosine kinase inhibitors as anti-infectives [96] . In the future , pathogens like EPEC should continue to shed light on how cells normally control actin assembly and how these mechanisms are altered in the context of disease .
All bacterial and mammalian cells are listed in S1 Table . EPECΔtir+pHA-Tir , EPECΔtir+pHA-TirY474F [22] , KC12+EspFU , KC12+vector [28] , EPEC+pEspFU-myc [11] , EPECΔtirΔeae+pHA-Tir , and E . coli+pIntimin [21] strains were streaked from glycerol stocks onto LB plates containing 35 µg/ml kanamycin and/or 100 µg/ml ampicillin and used within 2 weeks for host cell infections . 24 h prior to infection , single colonies were grown in LB + antibiotics with shaking at 37°C for 8–9 h . Cultures were then diluted 1:500 in Dulbecco’s Modified Eagle’s Medium ( DMEM ) + 100 mM HEPES , pH 7 . 4 , with antibiotics and grown standing overnight at 37°C in 5% CO2 , with the exception of E . coli+pIntimin , which was grown shaking overnight in LB + ampicillin . HeLa cells ( University of Massachusetts Medical School and University of California , Berkeley , Cell Culture Facility ) , NIH3T3 cells ( University of California , Berkeley , Cell Culture Facility ) stably expressing mCherry-βactin , and C2BBe1 ( referred to as Caco-2 ) cells ( American Type Culture Collection ) were cultured and seeded as described previously [11] . Caco2 cells were maintained with half media changes every 48 h for two weeks post confluency to generate polarized monolayers . ArpC2 Flox mouse fibroblasts ( University of North Carolina at Chapel Hill , Bear Lab ) [73] were maintained in DMEM ( with 4 . 5 g/L glucose + L-Glutamine + 110 mg/L sodium pyruvate ) supplemented with 1x GlutaMAX ( Gibco ) , 10% FBS , and 1x antibiotic/antimycotic ( Gibco ) . To obtain knockout ( KO ) and control populations , cells were treated with 2 µM 4-hydroxy-tamoxifen ( 4-OHT ) ( Sigma ) or an equivalent amount of DMSO for 6 days , including a media change to add fresh 4-OHT or DMSO on day 3 . ArpC2 Flox and KO cells were returned to normal media and used within a week . All cells were grown at 37°C in 5% CO2 . Bacterial infections were performed as previously described [11] . Briefly , cells were washed twice with phosphate buffered saline ( PBS ) and infected with bacteria diluted in DMEM + 3 . 5% FBS + 20mM HEPES , pH 7 . 4 to achieve a multiplicity of infection ( MOI ) of 3–10 , depending on the host cell line . For prime-challenge experiments , HeLa cells grown on glass coverslips in 24-well plates were infected with EPECΔtirΔeae+pHA-Tir at an MOI of 6 for 4 h , washed 4 times with PBS , and challenged with ~2 x 107 CFU/ml E . coli+pIntimin . Immediately following the addition of E . coli+pIntimin , plates were centrifuged at 172 x g for 5 min , then incubated for 10 min prior to fixation . For generating cell extracts to be used in immunoblotting , HeLa cells grown in 6-well plates were infected with bacteria at an MOI of 20 for 3 . 5 h , treated with 10 µg/ml gentamicin for 15 min , and washed with PBS prior to collection in PBS + 2mM EDTA and pelleting by centrifugation at 1150 x g for 5 min . HeLa cells and Caco2 monolayers were treated with 50 µM CK666 + 50 µM CK869 ( Calbiochem ) , 25 µM SMIFH2 ( Tocris ) , 4 µM Wiskostatin ( Sigma ) , or equivalent volumes of DMSO for 15 min prior to infection . During infections , media containing bacteria and inhibitors were added to HeLa cells and Caco2 monolayers , and the latter cells were washed with PBS and given fresh inhibitor-containing media every hour during the course of infection . NIH3T3 cells expressing mCherry-βactin [97] were infected 3 . 5–4 . 0 h prior to treatment with inhibitors , and live imaging was completed 15–120 min after the addition of the inhibitors . RNA and DNA transfections were performed using RNAiMAX or Lippofectamine-LTX reagents ( Invitrogen ) . To clone GFP-mDia1 , mDia1 plasmid DNA ( variant BC143413 , Dharmacon ) was PCR amplified as a Kpn1-Not1 fragment using primers ATCATCGGTACCATGGAGCCGCCCGGCGGGAG , and ATCATCGCGGCCGCTTATTAGCTTGCACGGCCAACCAACTC and ligated into the vector pKC-EGFP-C1 [97] . The plasmid was maintained in E . coli XL-1 Blue . For transient expression , 100 ng of GFP-mDia1 plasmid was transfected in 6-well plates . Sigma MISSION siRNAs ( see S2 Table ) were used at 40 nM for RNAi experiments . Targets were selected based on HeLa cell expression data cataloged on the Human Protein Atlas ( https://www . proteinatlas . org/cell ) . Immunofluorescence microscopy was performed as previously described [11] , and all antibodies and molecular probes are listed in S3 Table . Briefly , cells seeded onto glass coverslips were fixed in 3 . 7% PFA for 30 min , washed with PBS , permeabilized with 0 . 1% TritonX-100 , washed , and incubated in blocking buffer ( 1% FBS + 1% BSA in PBS + 0 . 02% NaN3 ) for 30 min . Primary antibodies against HA , LPS , mDia1 , c-Myc , Arp3 , phosphotyrosine , or Src phosphotyrosine-416 were diluted in blocking buffer and cells were probed for 40 min . Cells were washed and treated with Alexa Fluor 488 , 555 , 568 , or 647 conjugated goat anti-rabbit or goat anti-mouse secondary antibodies and/or DAPI and Alexa Fluor 488 or 647 labeled phalloidin for 40 min , followed by washes and mounting in Prolong Gold anti-fade reagent . All fixed and live cells were imaged using a Nikon Eclipse Ti microscope equipped with Plan Apoλ 100x 1 . 45 NA , 60x 1 . 40 NA , and Plan Fluor 20x 0 . 5 NA objectives , an Andor Clara-E camera , and a computer running NIS Elements software . Live phase-contrast imaging as well as mCherry visualization of infected NIH3T3 cells was performed using the 60x objective , and images were captured at 30 s intervals . A Nikon A1R confocal microscope equipped with a Plan Apo 60X 1 . 40 NA objective was used to capture the images in Fig 5A . All image processing was completed in ImageJ , and the mTrackJ and Cell Counter plugins were used for analysis . Pixel intensity plots were generated using the “plot profile” tool , and lines were drawn through pedestals or pedestal-forming regions after random selection in the HA-Tir and DAPI channels . Lines were only excluded or shifted if intense F-actin staining from stress fibers interfered with the profile . Statistical analyses of data sets were performed using Graphpad Prism software , and all statistical tests are noted in the figure legends . To determine the levels of mDia1 , ArpC2 , pTir , Tir , pSrc , Src , GAPDH , actin , and tubulin , cell pellets collected from 6-well plates were resuspended in lysis buffer ( 20 mM HEPES pH 7 . 4 , 100 mM NaCl , 1% TritonX-100 , 1 mM Na3VO4 , 1 mM NaF , 1mM phenylmethylsulfonyl fluoride , and 10 μg/ml each of aprotinin , leupeptin , pepstatin , and chymostatin ) , diluted in Laemmli sample buffer , and loaded into 9% SDS-PAGE gels . Proteins were transferred onto nitrocellulose membranes ( GE Healthcare ) , blocked for 30 min in PBS + 5% milk , and exposed to primary antibodies diluted in blocking buffer overnight at 4°C , plus a further 2 h at room temperature . Membranes were rinsed twice with PBS and washed thrice with PBS + 5% Tween-20 ( PBS-T ) . For detecting most primary antibodies , IRDye-conjugated secondary antibodies were diluted in blocking buffer and incubated with the membrane for 1–2 h . For detecting phospho-specific primary antibodies , horseradish peroxidase ( HRP- ) conjugated secondary antibodies were used . Membranes were again rinsed with PBS and washed with PBS-T . Bands were visualized using a LI-COR Odyssey Fc imaging system . Band intensities were quantified using the Analysis tool in LI-COR Image Studio software . Statistical analyses of data sets were performed using GraphPad Prism software , and all statistical tests are noted in the figure legends .
|
Microbial pathogens that rearrange the host actin cytoskeleton have made valuable contributions to our understanding of cell signaling and movement . The assembly and organization of the actin cytoskeleton is driven by proteins called nucleators , which can be manipulated by bacteria including enteropathogenic Escherichia coli ( EPEC ) , a frequent cause of pediatric diarrhea in developing countries . After ingestion , EPEC adhere tightly to cells of the intestine and hijack the underlying cytoskeleton to create protrusions called actin pedestals . While mechanisms of pedestal assembly involving a nucleator called the Arp2/3 complex have been defined for EPEC , the contribution of additional host nucleators has not been determined . We assessed the roles of several actin nucleators in EPEC pedestals and found that in addition to Arp2/3 complex-mediated nucleation , the formin mDia1 is a key contributor to actin assembly . These findings highlight the importance of nucleator collaboration in pathogenesis , and also advance our understanding of the molecular and cellular basis of EPEC infection , which is ultimately important for the discovery of new drug targets .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"hela",
"cells",
"gene",
"regulation",
"pathogens",
"biological",
"cultures",
"cell",
"processes",
"condensed",
"matter",
"physics",
"cell",
"cultures",
"research",
"and",
"analysis",
"methods",
"contractile",
"proteins",
"small",
"interfering",
"rnas",
"actins",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"actin",
"polymerization",
"proteins",
"gene",
"expression",
"cell",
"lines",
"pathogen",
"motility",
"physics",
"biochemistry",
"cytoskeletal",
"proteins",
"rna",
"cell",
"staining",
"cell",
"biology",
"nucleic",
"acids",
"post-translational",
"modification",
"virulence",
"factors",
"genetics",
"biology",
"and",
"life",
"sciences",
"cultured",
"tumor",
"cells",
"physical",
"sciences",
"non-coding",
"rna",
"nucleation"
] |
2018
|
Enteropathogenic E. coli relies on collaboration between the formin mDia1 and the Arp2/3 complex for actin pedestal biogenesis and maintenance
|
MicroRNAs ( miRNAs ) and trans-acting siRNAs ( ta-siRNAs ) are essential to the establishment of adaxial–abaxial ( dorsoventral ) leaf polarity . Tas3-derived ta-siRNAs define the adaxial side of the leaf by restricting the expression domain of miRNA miR166 , which in turn demarcates the abaxial side of leaves by restricting the expression of adaxial determinants . To investigate the regulatory mechanisms that allow for the precise spatiotemporal accumulation of these polarizing small RNAs , we used laser-microdissection coupled to RT-PCR to determine the expression profiles of their precursor transcripts within the maize shoot apex . Our data reveal that the pattern of mature miR166 accumulation results , in part , from intricate transcriptional regulation of its precursor loci and that only a subset of mir166 family members contribute to the establishment of leaf polarity . We show that miR390 , an upstream determinant in leaf polarity whose activity triggers tas3 ta-siRNA biogenesis , accumulates adaxially in leaves . The polar expression of miR390 is established and maintained independent of the ta-siRNA pathway . The comparison of small RNA localization data with the expression profiles of precursor transcripts suggests that miR166 and miR390 accumulation is also regulated at the level of biogenesis and/or stability . Furthermore , mir390 precursors accumulate exclusively within the epidermal layer of the incipient leaf , whereas mature miR390 accumulates in sub-epidermal layers as well . Regulation of miR390 biogenesis , stability , or even discrete trafficking of miR390 from the epidermis to underlying cell layers provide possible mechanisms that define the extent of miR390 accumulation within the incipient leaf , which patterns this small field of cells into adaxial and abaxial domains via the production of tas3-derived ta-siRNAs .
Small regulatory RNAs play fundamental roles in diverse aspects of animal and plant development [1]–[2] . However , few examples exist in which small RNAs direct early patterning events . Leaves provide a unique developmental system in which multiple small RNAs pattern the adaxial-abaxial ( dorsoventral ) axis [3] . Leaf primordia arise as a group of determinate founder cells on the flank of the shoot apical meristem ( SAM ) , a specialized stem cell niche at the growing tip of the shoot . The establishment of adaxial-abaxial polarity occurs early in leaf development and is concomitant with outgrowth of the primordium . Although polarity in leaves is ultimately specified through positional signals that convey to organ initials their proximity to the meristem tip [4] , in maize the polarized expression of small RNAs , both microRNAs ( miRNAs ) and trans-acting siRNAs ( ta-siRNAs ) , in the incipient leaf directs the patterning of the adaxial-abaxial axis [3] . MiRNA miR166 promotes abaxial/ventral fate in developing primordia by restricting the expression of class III homeodomain leucine zipper ( HD-ZIPIII ) transcription factors , which are necessary and sufficient to specify adaxial/dorsal fate [5]–[7] . Interestingly , the abaxial-restricted accumulation of miR166 is regulated by the tas3 ta-siRNA pathway . Loss-of-function mutations in leafbladeless1 ( lbl1 ) that disrupt ta-siRNA biogenesis give rise to an abaxialized leaf phenotype and loss of hd-zipIII expression , demonstrating that the ta-siRNA pathway is necessary for specifying adaxial fate in maize [3] , [8] . Accordingly , tasiR-ARFs , tas3-derived ta-siRNA species conserved between maize and Arabidopsis , accumulate on the adaxial side of incipient and developing leaf primordia , where they act to spatially restrict the expression domains of abaxial determinants . This includes , although indirectly , miR166 , which is expressed ectopically in lbl1 leaf primordia and accumulates in both the adaxial and abaxial domains [3] . Thus , organ polarity in leaves is ultimately achieved by regulating the spatial accumulation of both tas3-derived ta-siRNAs and miR166 , which divide the small field of cells of the incipient primordium into adaxial and abaxial domains . The functional importance of tasiR-ARF and miR166 activity in specifying the adaxial and abaxial domains of the leaf suggests mechanisms exist to maintain the accuracy of their spatiotemporal localization . However , even though our knowledge regarding small RNA biogenesis and function has increased significantly , and the detailed expression patterns of several mature miRNAs have been described [7] , [9]–[11] , little is known about such regulatory mechanisms . The biogenesis of these different small RNA classes in plants relies on specialized RNAi pathways . MiRNAs are ∼21 nucleotide small RNAs that arise from DICER-LIKE1 ( DCL1 ) -dependent processing of precursor transcripts that contain a stem-loop structure . The mature miRNA forms a complex with ARGONAUTE1 ( AGO1 ) , creating an RNA-induced silencing complex ( RISC ) that silences transcripts with complementary target sites , predominately through cleavage [2] . In contrast , the biogenesis of ta-siRNAs is more complex , relying on both miRNA and siRNA pathway components . Following miRNA cleavage , ta-siRNA precursor ( tas ) transcripts enter into an RNA-DEPENDENT RNA POLYMERASE6 ( RDR6 ) and LEAFBLADELESS1/SUPRESSOR OF GENE SILENCING3 ( LBL1/SGS3 ) -dependent pathway and are processed by DICER-LIKE4 ( DCL4 ) into phased , 21 nucleotide ta-siRNAs [12]–[15] . Although all ta-siRNA biogenesis follows this generic pathway , ta-siRNAs derived from the TAS3 precursor family are processed using a subspecialized pathway dependent on miR390 activity . MiR390 forms a specialized RISC with AGO7/ZIP , the activity of which is required to recognize the 5′ and cleave the 3′ miR390 target sites of TAS3A precursor transcripts in Arabidopsis to produce TAS3-derived ta-siRNAs [16] . Post-transcriptional mechanisms that operate at the level of biogenesis could modulate the spatiotemporal localization of small RNAs , perhaps through the limited availability of necessary biogenesis factors , as has been proposed for animals [17] . In this respect , the diversity of specialized RNAi pathways in plants , as exemplified by the unique association of miR390 with AGO7/ZIP [16] , would be one means to differentially regulate the accumulation of specific small RNA classes . Undoubtedly , transcriptional regulation of small RNA precursor transcripts also provides a mechanism by which the accumulation of mature small RNAs can be specified [18]–[19] . However , the analysis of the regulatory mechanisms that create precisely defined mature miRNA expression patterns in plants is hindered by the extensive redundancy present in miRNA precursor ( mir ) families , in which multiple mir genes produce identical mature miRNA species [2] . Here , we employ laser-microdissection ( LM ) to investigate the complex regulatory mechanisms that allow for the precise spatial accumulation of miR166 and small RNAs in the tas3 ta-siRNA pathway in the maize SAM . The comparison of mature small RNA in situ localization data with the expression profiles of precursor transcripts , as determined using laser-microdissected domains , indicates that the accumulation pattern of mature miR166 results from the co-operative activities of multiple mir gene family members and is further regulated at the level of biogenesis and/or stability . Furthermore , we show that the polarized expression of miR390 , an upstream component in leaf polarity , is established and maintained independently of the ta-siRNA pathway . The discrete adaxial accumulation of miR390 in the incipient leaf is regulated at the post-transcriptional level and might possibly result from limited mobility of mature miR390 over the span of a few cells .
Mature miR166 exhibits a complex spatiotemporal pattern of expression . It is most abundant in a group of cells below the incipient leaf but a gradient of miR166 expression , which establishes organ polarity , extends into the abaxial side of the newly initiating primordium . During primordium development , miR166 expression persists on the abaxial side of the leaf . MiR166 also accumulates in the vascular bundles , specifically in the abaxial phloem [3] , [7] . The maize genome contains at least nine mir166 loci , mir166a through mir166i , which have the potential to produce identical mature miR166 species ( miRbase release 10 . 1 ) . To determine which family members contribute to the spatiotemporal expression pattern of mature miR166 in vegetative tissues , we investigated mir166 gene expression in hand-dissected apices containing the SAM and four to five leaf primordia . Precursor transcripts of all nine mir166 genes accumulate in vegetative apices ( Figure S1 ) . Similarly , all nine mir166 genes are expressed in inflorescence tissues . This result could indicate substantial redundancy in the mir166 gene family , or alternatively , mir166 genes may have distinct , sublocalized expression patterns in these tissues . To distinguish between these possibilities , we sought to localize individual mir166 precursor transcripts in discrete domains of the maize apex . Plant miRNA precursors are typically low in abundance , presumably due to their rapid processing into mature miRNAs [2] . Consequently , expression analysis of mir genes using in situ hybridization is often not successful [3] , [7] . We therefore employed laser-microdissection ( LM ) [20]–[21] coupled to reverse-transcriptase PCR ( LM-RT-PCR ) as a technique to detect mir166 transcripts at cellular resolution [3] . The comparative analysis of mir precursor localization as determined by LM-RT-PCR in combination with localization of mature miRNAs is a potentially powerful technique to reveal novel mechanisms regulating the accumulation of miRNAs . As an initial analysis , we investigated the expression of a subset of mir166 genes in relatively broad domains of the maize apex ( Figure 1 ) . Cells were captured from regions in which miR166 accumulates [7] , including: 1 ) the P2 and P3 developing leaf primordia; 2 ) the incipient leaf ( P0 ) where polarity is established , the P1 , and the region of tissue just below the SAM; 3 ) more developed P4-P6 primordia; and 4 ) stem tissue , which contains extensive vasculature ( Figure 1A ) . The expression profile of selected control genes in the microdissected domains was tested to determine the accuracy of LM . Consistent with their reported in situ hybridization expression patterns , the miR166 target rolled leaf1 ( rld1 ) is expressed in all tissue samples tested [7] , rough sheath1 ( rs1 ) transcripts are limited to those domains that include the subtending regions of leaves [22] , and similar to kanadi2 ( kan2 ) , the expression of kan1 demarks cells in the developing young leaf primordia [23] ( Figure 1B ) . Although multiple mir166 family members are expressed in every tissue sample analyzed , each mir166 gene tested exhibits a unique expression profile ( Figure 1C ) . For instance , mir166a precursor transcripts are detected in the SAM region and stem , whereas transcripts from mir166b accumulate in P2 and older leaf primordia ( Figure 1C ) . mir166c , -d , and -e are expressed more broadly throughout the apex , but each still displays a distinct expression profile . The LM-RT-PCR approach thus allows analysis of maize mir gene expression with both high sensitivity and spatial resolution , without relying on transgenic approaches . The data also indicates that the complex pattern of miR166 accumulation results in part from the distinct transcriptional regulation of individual mir166 precursors . The partially overlapping expression patterns of mir166 precursors suggest some degree of redundancy . However , the expression profiles of individual mir166 genes are unique , indicating a process of functional diversification perhaps not unlike their hd-zipIII targets , which have distinct but overlapping roles in organ polarity , vascular development and meristem activity [24]–[25] . For example , their expression in stem tissues suggests a possible function for mir166a , -c , and -e in vascular differentiation ( Figure 1C ) . Similarly , the accumulation of mir166b transcripts in P2 and older leaf primordia may indicate a role in maintaining , rather than establishing , leaf polarity . To determine which mir166 genes act in the SAM to establish the abaxial-graded pattern of miR166 in the incipient leaf [3] , [7] and therefore contribute to the specification of leaf polarity , we next microdissected precisely defined domains within the SAM . Expression of mir166a through mir166i was evaluated in cells captured from below the incipient leaf , the incipient primordium , as well as the tip of the SAM in which mature miR166 does not accumulate ( Figure 2A ) . The expression profiles of the control genes kan1 , kan2 and rld1 in these domains recapitulate their described expression patterns and thus verify the purity of the LM samples ( Figure 2E ) . Transcripts of only a subset of mir166 family members are detected in the SAM . mir166a , -f and -i are expressed both in and below the incipient primordium , whereas mir166c is expressed at detectable levels exclusively below the initiating leaf ( Figure 2B ) . This data further supports the existence of functional diversification within the mir166 gene family and shows that a subset of mir166 precursor genes contributes to the abaxial accumulation of mature miR166 that is required for establishing abaxial fate . Interestingly , mir166a precursor transcripts also accumulate in the tip of the SAM ( Figure 2A and 2B ) . The detection of mir166a transcripts in this domain is surprising , considering that mature miR166 is not detectable in this region by in situ hybridization using either a riboprobe [3] , [7] or a highly-sensitive LNA-based probe ( Figure 2C ) . The abundance of rld1 and rld2 target transcripts in the meristem tip ( Figure 2D and 2E; [3] , [7] ) further suggests a lack of miR166 activity in this region . The tip of the SAM contains a population of indeterminate stem cells . The lack of observable miR166 accumulation in this region raises the possibility that either the maturation and/or stability of this miRNA is compromised in these cells . Plant stem cells may lack essential components of the miRNA processing machinery . Although we can detect transcripts of dcl1 and se homologs , core components of the miRNA precursor-processing pathway , in the tip of the SAM ( Figure 2F ) , the possibility that other miRNA processing components are lacking cannot be excluded . Alternatively , selected miRNA precursors may fail to engage the processing apparatus in stem cells through a hitherto unknown inhibitory process . A similar accumulation of unprocessed precursor transcripts occurs in mammalian embryonic stem cells in which Lin28 inhibits the Drosha-mediated processing of differentiation-promoting miRNAs [17] , [26]–[27] . Whether plant stem cells utilize an analogous mechanism to block the biogenesis of miRNAs associated with differentiation remains to be seen . The finding that mir166c and mir166i are expressed within and below the incipient leaf ( Figure 2B ) is consistent with previous observations showing that the tas3 ta-siRNA pathway generates the abaxial gradient of miR166 and specifies adaxial fate by restricting the expression domains of these two specific mir166 family members [3] . Maize tasiR-ARFs , which accumulate most pronounced on the adaxial side of developing leaf primordia , are produced from at least four tas3 loci ( tas3a-tas3d ) . To identify which tas3 family members contribute to the localized accumulation of tasiR-ARFs in the incipient leaf , we analyzed the expression of tas3a-d precursors in the SAM . Unlike the mir166 genes , all four tas3 family members are expressed within and below the incipient leaf and thus likely contribute redundantly to the establishment of leaf polarity ( Figure S2 ) . Moreover , tas3b-d transcripts are expressed in the tip of the SAM where tasiR-ARFs do not accumulate . A key component of tasiR-ARF biogenesis , lbl1/SGS3 , is expressed in the meristem tip as well as on the adaxial side of the incipient and developing leaves [3] . The restricted activity of additional tas3 ta-siRNA biogenesis components therefore likely limits the accumulation of tasiR-ARFs to only the adaxial side of developing primordia and prevents their accumulation in the SAM tip . As the biogenesis of tas3 ta-siRNAs is uniquely triggered by miR390 activity [12] , we determined the localization pattern of mature miR390 by in situ hybridization to ascertain whether it might be a restrictive component limiting tasiR-ARF biogenesis . Whereas an LNA probe against a murine miRNA yields no detectable hybridization signal , an LNA probe complementary to miR390 shows it accumulates adaxially within the incipient and developing leaves ( Figure 3A and 3C ) . Additionally , miR390 is expressed in vascular bundles ( Figure 3A ) . The adaxial accumulation of miR390 overlaps with that of tasiR-ARFs and the expression pattern of lbl1/SGS3 in these regions [3] , and is opposite to the abaxial accumulation of miR166 in the incipient leaf ( Figures 2C and 3A ) . Thus , miR390 is a restricting factor in the tas3 ta-siRNA pathway that limits the biogenesis of tasiR-ARFs to within developing leaves . That miR390 restricts the accumulation of tasiR-ARFs places it as one of the upstream factors in the maize leaf polarity network , as genetic and expression data position the tas3 ta-siRNA pathway upstream of miR166 as well as the hd-zipIII and yabby genes [3] , [8] . As a result of the mutually antagonistic relationship between the adaxial and abaxial sides of the leaf , the expression of adaxial and abaxial determinants , such as the HD-ZIPIII , YABBY and KANADI genes , changes in mutants with perturbed leaf polarity [4] . We therefore determined the pattern of miR390 expression in the apex of lbl1-rgd1 mutants , which develop fully abaxialized , radially symmetric leaves [28] . Unexpectedly , miR390 remains polarized within the initiating and developing leaf primordia of lbl1-rgd1 ( Figure 3B ) , even though these leaves are molecularly abaxialized with respect to expression of miR166 and members of the hd-zipIII and yabby families [3] , [8] . This persistent adaxial expression of miR390 in lbl1 mutants further highlights the importance of this small RNA as an upstream component in the maize leaf polarity pathway , whose polarized expression is established independent of the ta-siRNA pathway . To gain insight into the regulatory mechanisms that direct the accurate spatiotemporal localization of mature miR390 in the incipient leaf , we analyzed expression of its precursors using LM-RT-PCR . Previously , we had cloned a single mir390 precursor [3]; however , analysis of the completed draft sequence of the maize genome revealed one additional mir390 locus ( Maize Genome Sequencing Consortium; http://www . maizesequence . org ) . The two mir390 precursors , referred to hereafter as mir390a and mir390b ( Figure S3 ) , map to maize chromosomes one and five to regions of synteny with the rice and sorghum genomes containing a single mir390 locus . Only mir390a is expressed below the incipient primordium , but both mir390a and mir390b are expressed in the initiating leaf and may act redundantly in the biogenesis of miR390 and specification of leaf polarity ( Figure 2B ) . Similar to our observations for mir166a , both mir390 precursor transcripts are detectable in the tip of the SAM , although mature miR390 does not accumulate in this domain ( Figures 3A and 2B ) . This suggests that the post-transcriptional regulatory mechanism that limits the accumulation of miR166 in the stem cell-containing meristem tip also limits the accumulation of other miRNAs , perhaps acting as a general mechanism to limit miRNA-induced differentiation [17] . Surgical experiments separating the initiating leaf from the meristem produce radially symmetric , abaxialized leaves , suggesting the existence of a meristem-borne signal that establishes adaxial fate [29] . More recently , precise microsurgery experiments in tomato have shown that incisions in just the epidermal cell layer ( L1 ) similarly disrupt adaxial-abaxial patterning of leaf primordia [30] , suggesting that the L1 layer is necessary for the adaxial-promoting signal to function . As miR390 constitutes an upstream component in the maize leaf polarity network , we compared the accumulation of mir390 precursor transcripts between the L1 and sub-epidermal L2 layers of the SAM ( Figure 4A ) . Both mir390 primary transcripts accumulate exclusively within the L1 layer ( Figure 4B ) . This is contrary to the accumulation of mature miR390 , which despite the decreased sensitivity of in situ hybridization as compared to LM-RT-PCR , is detectable at least three cell layers internal to the incipient leaf's surface ( Figure 4C ) . The accuracy of LM is demonstrated through the mutually exclusive expression of the tissue specific markers outer cell layer4 ( ocl4 ) in the L1 layer and knotted1 ( kn1 ) in the L2 layer ( Figure 4B; [22] , [31] ) . Furthermore , the inability to detect mir390 precursor transcripts in the L2 samples is unlikely an artifact of sample dilution effects resulting from the fact that the L2 includes more cells than the L1 layer . Microarray analysis of the exact samples used in this study shows comparable expression levels in the L1 and L2 samples for the leaf polarity determinants rld1 , rld2 , Zmyabby9 , and Zmyabby14 , all of which are known to be expressed in both the L1 and L2 layers of the incipient leaf ( [8] ( K . Ohtsu and P . Schnable , unpublished data ) . Likewise , tubulin6 , mir166a , -f , and all tas3 transcripts were detectable in both the L1 and L2 layers ( Figure 4B and 4D ) . As cells in the L1 layer divide almost exclusively anticlinally , accumulation of miR390 in the sub-epidermal layers does not simply reflect inheritance from dividing epidermal cells in the incipient primordium . The non-overlapping localization patterns of mature miR390 and its precursor transcripts could reflect increased processing of miR390 in the L2 layer compared to the L1 layer . However , assuming that the expression of mir390 transcripts before their processing is approximately equal in the L1 and L2 layers , miR390 would be expected to accumulate most abundantly in the L2 , as no precursor transcripts accumulate there . However , in situ hybridization shows mature miR390 is equally or less abundant in the underlying L2 as compared to the L1 ( Figure 4C ) . The observed incongruence between the expression patterns of mature miR390 and the mir390 precursors could conceivably also reflect mobility of the miR390 small RNA from the epidermis into the underlying cells . Although miRNAs are thought to act largely cell autonomously [32]–[34] , mobility over just a few cells or in specific developmental contexts remains a possibility . Given that no maize mutants impairing miRNA biogenesis have been described , we cannot currently distinguish between these possibilities . That miR390 remains polarized in the abaxialized leaves of lbl1-rgd1 mutants ( Figure 3 ) and is responsible for restricting tasiR-ARF accumulation to the adaxial side of leaves places this small RNA upstream in the maize leaf polarity pathway [3] , [8] . The restricted accumulation of miR390 demarcates the adaxial most cells within developing leaves that produce tas3-derived ta-siRNAs , which in turn restrict the accumulation of miR166 to the abaxial side of leaf primordia [3] . As an upstream initiator of a small RNA cascade that ultimately patterns the leaf into adaxial and abaxial domains , the extent of miR390 activity must be precisely defined . Such precision is likely achieved through a variety of mechanisms regulating miRNA biogenesis , stability , and possibly movement . Both mir390 genes are expressed outside the incipient leaf , in the L1 layer of the entire SAM , as their precursors are detected in microdissected samples that comprise just the tip of the shoot apex or the L1 layer , but not in samples comprising the L2 layers ( Figure 4A and 4B ) . mir390a is also expressed below the incipient leaf . Nevertheless , regulation of miR390 biogenesis and/or stability allows this small RNA to accumulate only in the few adaxial cells of the incipient primordium and not elsewhere in the SAM , such as in the tip ( Figure 3A ) . Although the mechanisms for such regulation are not currently understood , these should function independent of the ta-siRNA pathway , as miR390 remains localized to the incipient primordium in lbl1-rgd1 ( Figure 3B ) . The post-transcriptional regulatory mechanisms that limit miR390 accumulation in the SAM tip may similarly regulate other miRNAs such as miR166 , which is not detected in the meristem despite the presence of precursor transcripts ( Figure 2 ) . The restricted adaxial accumulation of miR390 in the incipient leaf might also be achieved through channeling into a subspecialized RNAi pathway . MiR390 is selectively incorporated into an AGO7/ZIP complex to execute the processing of tas3 precursor transcripts [16] . Such an association with AGO7/ZIP , or other components of this unique RNAi branch , might selectively stabilize miR390 on the adaxial side of the leaf . Besides such regulatory mechanisms that limit the biogenesis and/or stability of miR390 , discrete trafficking from the epidermis into underlying cell layers could potentially provide an additional mechanism that defines the extent of miR390 activity within the incipient leaf . The hypothesis of miRNA movement from the L1 epidermal layer is intriguing , as it provides an appealing explanation for miR390's upstream role in defining the boundary between the adaxial and abaxial domains of leaves , and is consistent with an earlier hypothesis suggesting that mobility of ta-siRNA pathway components may generate the abaxial graded pattern of miR166 accumulation [3] . The movement of miR390 could feasibly result in a pattern of graded activity that would translate into a gradient of tas3 ta-siRNA accumulation from the adaxial side of the leaf . As tas3 ta-siRNAs indirectly restrict the expression of mir166c and mir166i , such a ta-siRNA gradient could specify the inverse graded accumulation of miR166 on the abaxial side of the leaf that adaxially restricts hd-zipIII transcripts ( Figure 5 ) . In addition to miR390 , non-cell autonomy of ta-siRNAs remains a possibility , as ta-siRNAs are processed by DCL4 and RDR6 , factors required for the biogenesis of mobile siRNAs during systemic silencing in Arabidopsis [35]–[36] . Our findings indicate that small amounts of RNA from laser-microdissected tissue samples are sufficient to detect mir166 and mir390 precursors , providing a novel approach to study miRNA regulation in specific cell-types and stages of development in plants and , perhaps , animals . Along with in situ hybridization analyses , our results suggest that miR166 and miR390 accumulation is subject to complex tissue and cell-type specific transcriptional and post-transcriptional controls , allowing for functional diversification of miRNA gene family members in development . The potential limited movement of miR390 has significant implications for the possible roles of small RNAs during development in establishing patterning events . Finally , the complex mechanisms regulating the biogenesis , stability , and possible movement of miRNAs should be taken into consideration when designing artificial miRNA studies and interpreting the intricate miRNA accumulation patterns in plants and animals .
Sequences corresponding to the mir166a to mir166i precursors were retrieved from miRbase ( http://microrna . sanger . ac . uk/sequences ) and extended using maize EST and genome sequence databases ( http://magi . plantgenomics . iastate . edu and http://www . maizesequence . org ) . The mir390a precursor has been described previously [3] . Sequences for mir390b were assembled using the above-mentioned databases , and the BAC accession number for mir390b is AC197125 . The maize genome contains at least five dcl genes; two most similar to Arabidopsis DCL3 and one each most similar to DCL1 , DCL2 , and DCL4 ( Figure S4; http://www . chromdb . org ) . The BAC accession number for dcl1 is AC191256 . The maize SERRATE homologs have been described previously [37] . Secondary structures for the mir390a and mir390b precursors were predicted using RNAfold . Total RNA was extracted from vegetative apices including the SAM and four to five leaf primordia and from immature ears ( 0 . 5–1 cm ) using TRIzol reagent ( Invitrogen ) according to the manufacturer's instructions . DNaseI-treated RNA was converted into first-strand cDNA and amplified by PCR according to standard protocols . Primer sequences used in this study are listed in Table S1 . Tissue sections prepared from the shoot apices of two-week-old wild-type or lbl1-rgd1 seedlings were pretreated and hybridized as previously described [38] . LNA probes with sequences complementary to miR390 , miR166 , and murine miR122a were synthesized by Exiqon ( Vedbaek , Denmark ) and digoxigenin-labeled ( Roche ) according to the manufacturer's protocol . Ten picomoles of each probe were used per slide pair and hybridization and washing steps were performed at 50°C . The rld2 probe , comprising nucleotides 625–1677 of the coding sequence , was used at a concentration of 0 . 5 ng/µL/kb as described [3] . Shoot apices from two-week old wild-type seedlings were dissected and immediately fixed in pure , ice-cold acetone . Acetone was gradually replaced with xylene and subsequently paraplast . Cells of interest were captured from 10 µm tissue sections using a PALM MicroBeam system . Detailed embedding and LM protocols have been described previously [20] , [21] . Total RNA was extracted and linearly amplified using PicoPure RNA Isolation and RiboAmp HS RNA Amplification Kits ( Arcturus Bioscience , Inc . ) according to the manufacturer's protocols . RNA was treated with DNaseI and subsequently analyzed by RT-PCR using a one-step method ( Qiagen ) and the primers listed in Table S1 . Three biological replicates with at least two technical replicates were performed .
|
Small RNAs regulate many key developmental processes . Consistent with a prominent role in development , miRNAs exhibit complex and distinctive expression patterns . In this study , we identify regulatory mechanisms that allow for the precise spatial accumulation of developmentally important small RNAs in plants . Plants generate new leaves throughout their lifetime . These arise on the flank of a specialized stem cell niche , termed meristem , at the plant's growing tip . Each newly formed leaf becomes polarized and develops distinct adaxial ( top ) and abaxial ( bottom ) sides . The establishment of adaxial–abaxial polarity requires a complex genetic network , including miRNAs and trans-acting siRNAs . We used a focused laser to microdissect regions of the shoot apical meristem and developing leaves of maize to analyze the expression profiles of the small RNA precursor molecules . By comparing these expression profiles to the accumulation patterns of the mature small RNAs , we show that precursor genes are subject to tissue-specific regulation and exhibit diverse expression patterns during leaf development . Our findings suggest that mechanisms exist to regulate the biogenesis , stability , and possibly even the intercellular movement of small RNAs . Such regulation should be considered when designing artificial miRNAs and has implications for the roles miRNAs play during plant and animal development .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"plant",
"biology/plant",
"genetics",
"and",
"gene",
"expression",
"developmental",
"biology/plant",
"growth",
"and",
"development",
"developmental",
"biology/pattern",
"formation"
] |
2009
|
Regulation of Small RNA Accumulation in the Maize Shoot Apex
|
In systems and computational biology , much effort is devoted to functional identification of systems and networks at the molecular-or cellular scale . However , similarly important networks exist at anatomical scales such as the tendon network of human fingers: the complex array of collagen fibers that transmits and distributes muscle forces to finger joints . This network is critical to the versatility of the human hand , and its function has been debated since at least the 16th century . Here , we experimentally infer the structure ( both topology and parameter values ) of this network through sparse interrogation with force inputs . A population of models representing this structure co-evolves in simulation with a population of informative future force inputs via the predator-prey estimation-exploration algorithm . Model fitness depends on their ability to explain experimental data , while the fitness of future force inputs depends on causing maximal functional discrepancy among current models . We validate our approach by inferring two known synthetic Latex networks , and one anatomical tendon network harvested from a cadaver's middle finger . We find that functionally similar but structurally diverse models can exist within a narrow range of the training set and cross-validation errors . For the Latex networks , models with low training set error [<4%] and resembling the known network have the smallest cross-validation errors [∼5%] . The low training set [<4%] and cross validation [<7 . 2%] errors for models for the cadaveric specimen demonstrate what , to our knowledge , is the first experimental inference of the functional structure of complex anatomical networks . This work expands current bioinformatics inference approaches by demonstrating that sparse , yet informative interrogation of biological specimens holds significant computational advantages in accurate and efficient inference over random testing , or assuming model topology and only inferring parameters values . These findings also hold clues to both our evolutionary history and the development of versatile machines .
Much attention is given to functional networks ( e . g . , scale-free , small world and others ) resulting from the complex interactions between their constituents [e . g . , 1–5] . For example , the mechanisms of module assembly in biological molecular networks [6]–[8] ( with underlying motifs [9] ) exhibit coordinated , complex functionalities; interconnectivity among unreliable elements yields reliable dynamic performance [10]–[12] . Similarly , the study of a complex biological system as a whole can be emphasized to understand how system properties emerge from the interaction of multiple components [13]–[15] . Tendon networks at anatomical scales are intricate and poorly understood componentsof the neuromuscular control of the hand . Understanding their functional characteristics is critical to gaining insight into the brain-body co-evolution that has facilitated dexterous manipulation in modern humans , as well as improving clinical rehabilitation strategies in orthopedic and neurological conditions . The complexity of tendon networks of the fingers is legendary , and thus the so-called Winslow's rhombus is a generic topological approximation that has been widely adopted since the 18th Century as proposed by the famous Danish-born anatomist J . B . Winslow in 1732 [16]—especially as the descriptions popularized by Zancolli [17] and Garcia-Elias et al [18] . It is known that minor variations in its structure can exist across humans [e . g . , 19] , and most work has focused on anatomical/structural descriptions via dissection or imaging and material properties [e . g . , 20–27] , or simplified computational models [e . g . , 28 , 29] . Importantly , critical structural features , e . g . , tendon multiplicity and interconnections are known to remain un-detected with imaging modalities , for example , Ultrasonography ( US ) or Magnetic Resonance ( MR ) [30] . As an alternative to structural descriptions , we have proposed functional descriptions of such systems that underscore their sensitivity to topological details [31] . The purpose of this work is to demonstrate that it is possible to use sparse experimentation to , in practice , extract topologies that capture the dominant functional features of these poorly understood structures , which allows us to begin to understand in detail the anatomical and neural co-adaptations that enable dexterous manipulation in modern humans . This work is enabled and motivated by our earlier work in [32] , where we inferred the topology of tendinous networks in simulation . That work showed not only that network topology matters functionally , but that it was—in principle—possible to infer experimentally the structure of arbitrary networks using the most informative force data . Now , we take the critical enabling experimental step of demonstrating the validity and utility of this approach when applied to actual physical anatomical systems , with the imperfections , nonlinearities and actuation/measurement noise that this implies . We do so by testing networks of “known” topology made of strings of synthetic Latex , as well as biological ( i . e . , cadaveric ) tendinous networks of unknown topology . As mentioned above , these biological specimens are complex sheets of collagen fibers , which for centuries have been approximated as networks of strings [16] , but for which there is no functionally validated string-based approximation .
We evolve a population of models ( both topology and their parametric attributes ) bottom-up from a primordial mesh of strings ( Fig . 2a ) . The strings are arranged such that some are joined at the nodes ( shown as filled circles in Fig . 2a ) , while the rest overlap and slide past each other . This connectivity allows the models to evolve into any topology ( i . e . , number of strings and intermodal connectivity ) that a hidden network system may have . The length and cross sectional area of each string are used as free parameters to evolve the model topology and parameter values simultaneously . String length is a topological as well as a parametric variable: a string can be considered absent from the model if it evolves to a length for which it remains slack for all loading conditions; or present if taut for some or all load setswhere its length influences force transmission and inter-nodal distances . Cross sectional area is a parametric variable that defines the load-bearing and deformation characteristics of a string for given stress-strain relationship ( linear or nonlinear ) . For synthetic networks in this work , we used the linear stress-strain relations for Latex rubber ( part SLR-040-E , 1 mm thickness , 380 mm×305 mm , Small Parts Inc . ) , while for the tendinous extensor mechanism we employed nonlinear tendon properties as reported in [34] . The total number of nodes is maintained constant , but their location is allowed to vary in response to loading—except for the two grounded nodes ( shown as filled squares in Fig . 2a ) where reaction output forces are measured . Stage II assumes access to the most informative database of load sets , and consists of four parts , ( i ) population of models , ( ii ) model analyses , ( iii ) theirfitness evaluation , and ( iv ) evolution . A few additional evolutionary strategies are also used so that the models can evolve better and faster . In the first , very small changes ( <1% ) in the free parameters are performed but with a higher rate ( 5prate ) . In the second , major changes ( >20% ) are accomplished with a smaller rate ( 0 . 2prate ) . Lastly , the worst model in the population is crossed over with a random one if the former does not exhibit improvement over a number of iterations . After the free parameters of a chosen number of models are evolved using the informative input-output data ( Stage II , Fig . 1 ) , only those strings that become taut for at least one set of simulated tests constitute the model topology . Those that remain slack for all simulated tests do not participate in the connectivity . However , they are retained in each model as they may get taut at a later stage in the inference algorithm . Once the population of models has evolved to explain the previous informative datasetsas per the termination criteria ( Stage II , Fig . 1 ) , a new most informative test is generated inStage III , Fig . 1 . Whereas the fitness of the models was to explain available data , the fitness of the tests is to make the models disagree in their prediction . Tests that make the models disagree are likely to be more informative because they uncover functional differences across models . Ideally , conducting the inference process in real time requires the availability of a large parallel computing network to evolve the models and the most informative actuation tests shortly after each new load set is added to the database . Because this is not feasible ( see discussion ) with manual application of loads and the need to test perishable tissue , we did the next best thing: collected the experimental data for the Latex and tendinous networks in dedicated experimental sessions—and ran the algorithm off-line using those data sets . Whenever the algorithm requests the next most informative test ( found as described below ) , we provide the load set corresponding to the nearest neighbor ( in the least squared sense ) to it from the available load sets . In our experiments , the direction of each test ( input force ) is fixed in the global frame , and we therefore only evolve their magnitudes . These magnitudes Fnew are mutated in a similar manner as the string free parameter cnew described above . If Fnew is found to be less than FL ( = 0 N ) or greater than FU ( = 5 N ) , the lower and upper force bounds , Fnew is set to FL or FU respectively . All models in Stage II , Fig . 1 are subject to these testsin simulation and the incongruity in their force-deformation response is quantified by an error e_test given bywhere NDS represents the number of evolved models in the population , NR is the number of reaction forces Rkj , ND is the number of distances dkj between the input and output nodes , is the mean of the reaction forces over the models and is the mean of the distances . Analteredtest replaces the previous test if the corresponding e_test value is higher ( a larger e_test implies greater discrepancy suggesting that the test is more informative ) . Otherwise , the previous testis retained . The termination criteria for the evolution process for the most informative test ( used in the subsequent experiment ) are similar to those used in model evolution ( Stage II , Fig . 1 ) . In Stage I , we first picked a single load set from the experimental data at random to seed the database . This load set formed the initial database used in Stage II ( Fig . 1b ) .
Figures 3c/d show the best eight models inferred for the ‘AFH’ and ‘aWR’ target networks , respectively . These were obtained by inferring a total of 24 models for each network ( three runs , each with population of eightmodels ) . Each run used at most 20 informative load sets . Table 1 shows the training set ( e_training ) and cross-validation ( e_cross ) errors for each model . Model ( iii ) in Fig . 3c has the smallest e_crossand close to the smallest e_training errors . Visual inspection confirms its structural resemblance with the ‘AFH’ network ( Fig . 3a ) . Model ( ii ) in Fig . 3d with the least cross-validation error is also topologically similar to the ‘aWR’ network . Models ( iii ) , ( vi ) and ( viii ) that have e_training errors less than 3 . 6% are also structurally comparable to the ‘aWR’ system . However , the other four models are visually dissimilar though they are functionally similar within the e_trainingerror of 3 . 7% and e_cross error of 6 . 4% ( Table 1 ) . Also , for e_cross errors less than 6 . 1% , models in Fig . 3c are all functionally similar to the ‘AFH’ network though topologically dissimilar . We observe structural diversity among functionally similar models for the ‘AFH’ network to within ±1 . 0% of e_training errors and ±1 . 3% of e_cross errors . Likewise , models functionally similar to the ‘aWR’ network differ in structure within e_training and e_cross errors of ±0 . 8% and ±1 . 1% respectively . The number of functional evaluations required to infer the ‘AFH’ and ‘aWR’ networks is c . 0 . 7 vs . 0 . 5 million , respectively . The CPU times for model evolution and generation of the most informative tests on three different machines for the synthetic targets are presented in Supporting Information , Tables S1 and S2 respectively . To confirm if evolving informative tests improves the inference process , we performed an additional three baseline inference runs for each target network using 20 random data sets . The e_training and e_cross errors are compared ( Fig . 5 ) , with error bars showing the standard error over three runs . The e_training errors with random tests are comparable to , or better than , those found using informative tests ( left plots , Fig . 5 ) . Importantly , however , the e_cross errors using informative tests are significantly lower for both target networks ( right plots , Fig . 5 ) . We also performed simple parametric fitting to infer the target networks in Figs . 3 ( a ) and ( b ) using the all-in-all ( see Supporting Information , Figs . S2 , c–d ) topologies . When performing only parametric vs . parametric and topological inference , the numbers of free parameters used were different . In parametric fitting , only six strings were used to connect the three input nodes to the two output ones . The length and cross section of each string were evolved as in the topological inference which allowed us to use the model evolution algorithm described before . In parametric fitting , all the six strings participated by becoming taut and hence the all-in-all topology did not change . In all cases , inferring the topology of the target network yields significantly better results ( Fig . 5 , green lines ) than only inferring parametric values using respective all-in-alltopologies . The stated primary aim of this study is to infer the structure of hidden tendinous systems . This is important both from the evolutionary perspective where it is the actual topology that changes , and also to plan the rehabilitation and surgical repair/replacement in cases of minor/major injuries . A 3-input , 2-output system for instance the synthetic networks herein can be learnt through say , a neural network . However , such learning processes provide only a mathematical relation but lack the physical structure and insight into the complex biological systems . Thus , in our work we emphasize the need for structural models . By allowing the tendon topology to vary and comparing the data fitting results with those obtained for a presumed , primitive ( all-in-all ) structure , we demonstrate that performing parametric-only fitting with preconceived topologies resulting from , say , a scientist's ingenuity and insight may not necessarily lower the fitting errors . Topological inference is essential in addition to parametric-only fitting . Figure 4 ( e ) shows the ten best models from the 25 total inferred over five runs , with a population of five models per run . The e_training errors are below 7 . 9% and e_cross errors are below 7 . 2% ( Table 2 ) , which are comparable with those for the Latex networks . The mean e_cross error ( Fig . 6 ( b ) , solid line ) converges after the models comply with the experimental information from 16 informative tests . There is no further alteration in their structures with additional informative tests . All ten models differ in topology from each other even though they are functionally similarin that they exhibit comparable e_training and e_cross errors within the limits aforementioned . Models ( i ) , ( iv ) , ( viii ) and ( ix ) closely resemble the Winslow's rhombus ( Fig . 4 ( a ) ) , which is consistent with the findings in [17] that ignores the transverse bands . The cross-validation errors for models ( i ) and ( iv ) in Fig . 4 ( e ) are the lowest , with model ( i ) being marginally higher even though its training set error is significantly lower . Therefore , we chose models ( i ) and ( iv ) from Fig . 4 to further investigate the functional contribution of each member string . We eliminate one string at a time from each network and re-evaluate the e_cross errors . For model ( i ) , that is shown in panel ( a ) in Fig . 7 , the comparison of e_cross ( Fig . 7 ( b ) ) reveals that the intact model invariably yields the lowest e_cross error , demonstrating that all member strings in its topology are functionally relevant . In contrast , e_cross errors are lower ( strings 22 and 24 removed ) or the same ( strings 7 and 10 removed ) when several member strings are removed from model ( iv ) ( shown in panel ( d ) in Fig . 7 ) . The e_cross error decreases when member strings 22 and 24 , which are in series and connect the grounded nodes , are removed . Such a connection does not exist in model ( i ) or the target network Fig . 4c/d , which suggests they are not necessary and in fact only serve to bias the loading on the grounded nodes and thus pollute the reaction forces . In contrast , the e_cross error remains unchanged when member strings that form the right lateral offshoot ( member strings 7 and 10 connected in series ) are removed . Further visual inspection of model ( iv ) indicates that these member strings barely get taut when the model is actuated with the informative tests used during its inference . This suggests that member strings 7 and 10 are not functionally necessary , but do not negatively affect the fitness of the network for the loading tested . Lastly , similar to the Latex networks ( Fig . 5 ) , we report a control case ( Fig . 6 ) with the cadaver network inference where we find that informative tests are better than random tests for both topological and parametric inference . The mean e_training errors are comparable in the three cases ( Fig . 6 ( a ) ) , especially after the 16th data set is introduced . However , the trend in Fig . 6 ( b ) confirms that the use of informative tests lowers the e_cross errors when both topological and parametric inference is performed . Importantly , we also observe that simple parametric inference with informative tests is better than the case of topological and parametric inference with random tests .
We demonstrate , to the best of our knowledge , the first functional inference of a complex anatomical structure using sparse experimentation . Without reverting to exploratory dissection ( which is disruptive ) or structural imaging ( which is expensive and not necessarily informative of mechanical interactions under loading ) , we infer functional structure of a biological tendinous network by co-evolving the models with informative tests . We began by validating and calibrating our novel methodology using two synthetic Latex target networks , and then applied our method to the real-world problem of inference of the functional structure of the tendinous extensor mechanism tissue excised from a cadaver hand . Notwithstanding ( i ) the specific optimization procedure ( we used a stochastic hill climber search , which are hotly debated by its supporters and retractors , but any other optimization that is suitable to the problem at hand can be used ) , or ( ii ) our current inability to run the estimation-exploration algorithm real time ( we collected the experimental data for the Latex and tendinous networks in dedicated experimental sessions—and ran the algorithm off-line using those data sets ) , our results clearly illuminate and demonstrate several important features and concepts about this approach . These include ( i ) the powerful utility of a novel , general purpose predator-prey estimation-exploration algorithm for topologic and parametric inference of physical systems , and ( ii ) the particular functional characteristic of our test system: the extensor mechanism of the fingers whose structure and function have been debated since at least the 16th century . In this first application of the predator-prey estimation-exploration algorithm [33] for topologic and parametric inference toactual biological ( cadaveric ) physical systems , we demonstrate that informative tests perform better than random tests . This is critical when limited to a finite number of tests of the physical specimen , which in this case is costly and can damage the specimen by excessive testing . We define the most informative tests as those that , in simulation ( Fig . 1 ) , are evolved to maximize disagreement among the population of current models . We introduce these tests sequentially in that the population of models evolved explains the informative data available up to that point in time . We show that a small number of informative physical tests produce input-output data sets that significantly lower the cross-validation error of the resulting models . Thus , the predator-prey competition carried out in simulation findsthe most informative tests . These tests , even with minor deviations from those predicted , provide significantly useful experimental data to guide the development of the next generation of models . We remark that , for the tendinous specimen , we needed to extract the experimental data within the first 8 hours of its excision to avoid structural and/or material degradation . Evolving the models with the most informative tests , as predicted in simulation in stage III of the inference process , was not possible with the available computational resources; the overall inference process took much more than 8 hours . Rather , we sought to access the closest possible tests that were informative , if not the most informative , from the experiments performed a priori . We further remark that for both , synthetic and biological networks , generating the most informative test was not possible with the experimental setup used ( Figs . 3–4 ) as it required the input tensions to be achieved by pulling on the tethers manually . As the inputs were strongly interrelated , a slight manipulation of an input tether disturbed the tensions in the others . Achieving the accuracy of the recommended most informative test required much effort and was cumbersome . There were discrepancies ( noise and/or measurement errors ) even when efforts were made to load the network ( s ) with the most informative tests . While structural/material degradation was not a concern with the synthetic networks of known topologies , inference of these was performed with informative tests to verify if the latter , even when they not being the most informative , can evolve models to adequately resemble the target in structure and/or function . We show in all three cases ( synthetic and biological targets ) that informative tests do infer the networks , known or hidden , better than the random tests . This work is , therefore , a successful proof of concept that does demonstrate the utility of our approach and produces results that are valuable to the field of functional inference in biological systems . Based on our earlier work in [32] where we infer the Winslow's Rhombus ( Fig . 4a ) in simulation using the most informative force-displacement data , and supporting information ( Fig . S1 c ) where we infer the structure of the ‘aWR’ and ‘A’ networks in simulation but using only the most informative force data , we further expect that the most informative tests will perform better than the informative ones once better computational resources and experimental setups are available to make the overall inference process more efficient . We also show that simultaneous topological and parametric inference yields better results than the parametric inference alone . Most system identification for biomechanical models is limited to parametric inference [38]–[40]—wherein the structure or topology of the system is chosen a priori based on expert knowledge , and parameter values are tuned to fit the experimental data . Very few studies , ( e . g . [41] ) have performed simultaneous inference of both the topology and parameters of anatomical systems . We show that in our experiments on synthetic and biological physical systems , the tuning of string parameters in networks with fixed topology is insufficient to minimize cross-validation errors . Importantly , cross-validation errors are a better estimate of model accuracy and generalizability because they evaluate fitness with respect to input-output data sets that were not used to train the model in the first place . A clear distinction needs to be made between topology and parameter values . Models are assembled by exploring the space of possible combinations of building blocks ( in this case , strings and nodes ) . A specific model topology is a specific connectivity map among a specific set of building blocks ( i . e . , string connectivity ) . The model parameters are the individual properties of each building block ( i . e . , length of strings ) . In practice , however , it is necessary to “parameterize” the topology to be able to encode ( i . e . , represent ) it so that an algorithm can search the topological space . In our case , we defined our primordial mesh of strings ( Fig . 2a ) and parameterized different topologies by allowing strings to become long enough to , in practice , “disappear” because they can no longer carry tension . While this methodological distinction between topology and parameter values may be considered semantic—and therefore debatable—in practice such implementations can and do produce models with distinctively different physical structures [33] . In our case , we evolve populations of string networks with patently different number and connectivity among load-carrying strings—which assume distinct anatomical structures . On the methodological side , several issues are key to accurate inference . These include accurate assumptions pertaining to ( a ) the material properties and the strain-deformation models , ( b ) informative experiments capturing full network functionality and interaction conditions , ( c ) automated experimental setup , ( d ) the primordial connectivity representation , and ( e ) the choice of the algorithm used for model evolution . If inappropriate assumptions are used , the experimental data may not be within the set of predictions of any feasible model topology or parameter values , or the candidate tests generated by the inference process may not be informative . The inference of the extensor mechanism was also performed assuming constant elastic modulus ( 1 GPa ) as opposed to nonlinear stress-strain relationship for tendons used from [34] to evolve the models in Fig . 4 . The e_cross errors of the resulting models were about 30% ( Fig . 8 ) , much higher than the e_cross errors ( Table 2 ) for the models in Fig . 4 . Inaccurate assumptions for the tendon material properties led to the evolution of models whose functionality was not in agreement with that of the target extensor mechanism . Alternatively , while it is possible to evolve the material properties as well along with the models , approximating the form and range of the stress-strain relation ( e . g . , exponential , transcendental , linear ) may be difficult . Use of a different set of material properties can lead to either ( i ) erroneous structural and parametric predictions with high e-training or e_cross errors for the models , or ( ii ) an alternate set of functionally similar models that explain the network functionality reasonably well . All the string networks were inferred here using the large deformation-small strain assumption . Models may also be evolved using the Green-Lagrange large-strain [35]–[36] theory . In case of the tendinous tissue , the anatomical extensor mechanism did not exhibit significant deformationand thus , the small-strain assumption was suitable . We inferred the hand extensor mechanism by considering only a part of its overall functionality . We were limited to testing the anatomical specimen as it lay flat on a hydrated surface , as opposed to wrapped over finger joints . Even so , all string modules , as described in the classical Winslow's model ( Fig . 4 ( a ) ) , were nearly captured in some models . To reproduce the most informative tests recommended by Stage III of the estimation exploration algorithm , the setup should be fully automated and computer-controlled . Independent motors should be mounted with the respective input tethers wrapped tightly around them to control the tensions . In addition to using digital force scales , force sensors ( e . g . , strain gages ) should also be used to control the motor rotations via a feedback loop . Achieving the most informative test with the automated setup ( as opposed to manual ) will be less painstaking and more accurate . The Latex and biological specimens are inferred using the string model representations shown in Supporting Information , Fig . S2 . Both the number and interconnectivity between the strings in that primordial network can influence model evolution . Here , we allow the strings to be both overlapping and tightly connected ( small circles in meshes in Fig . S2 ) . This initial mesh is chosen to facilitate the model ( s ) inferred from the finger extensor mechanism to assume the form of Winslow's Rhombus . A primordial mesh with no overlapping strings can yield a different set of models . Factually , however , the extensor mechanism is a sheath of collagen fibers . Using two-dimensional parameterizations ( e . g . , rectangular/hexagonal cells to represent the primordial mesh ) canhelp yield more topologically diverse models . All factors mentioned above , and the noise involved in experimental data influence the landscape of the objective function . In view of this , functionally similar but topologically and parametrically diverse models obtained through the Random Mutation Hill Climber ( RMHC , a variant of Genetic Algorithm that employs only mutation ) could all be local optima existing very close to the global one in the design space . Due to the noise present in the data , it is not expected for a global optimum to have significantly lower e_training and e_cross errors . As an aside , we show that the RMHC is capable of finding a close to global optimum for smooth functions ( see supporting information , Fig . S3 ) , even when the design space is infested with numerous local optima , if adequate computational resources are employed . We also show that alternative , classical optimization algorithms often converge to a local minimum . Further , they will not be able to negotiate the discontinuities in the design space such as those in our problem which correspond to cases wherein nonlinear analyses do not converge for candidate models . One of the goals was to confirm whether the Winslow's Rhombus ( Fig . 4 a ) is an accurate string representation of the hand extensor mechanism . On performing parametric only fitting with this representation ( but without transverse bands ) as the primordial mesh and informative data generated using the predator-prey approach , we found that the mean e_training and e_cross errors ( supporting information , Fig . S4 ) were comparable to those of the topologically and parametrically inferred models ( Fig . 4 ( e ) ) in Fig . 6 . This suggests that the Winslow's Rhombus could belong to the same family as these functionally similar models . As detailed later , additional information may be necessary to extract a true global model . The predator-prey approach is an optimal experiment design ( OED ) method wherein through competition , most informative tests ( optimal sample points ) are generated to evolve the best models that explain these tests . However , this approach is unlike other OED methods , e . g . , D-optimal , L-optimal and minimax-optimal wherein model parameters ( or their functions ) are determined by minimizing , for instance , generalized variance . In Bayesian type OEDs , prior information on model parameters is assumed . With regard to the study of complex biomechanical structures by anatomists , biomechanists and biologists , most previous work has naturally focused on inferring the structure of the tendinous networks via dissection or imaging . In contrast , we interrogate biological networks through a non-invasive computational machine learning procedure . Invasive techniques may damage the tissue , while imaging methods may miss critical functional interactions ( e . g . , seen only under specific loading conditions ) . Our non-destructive inference method yields both topological and parametric information . For example , the specific number and lengths of the tendinous members of the extensor mechanism affect the distribution of tension to the finger joints [19] , [41] . We suggested before that the interpretation by Zancolli [17] and Garcia-Elias [18] of the hand extensor mechanism as Winslow's Rhombus is partly correct . We also illustrated that functionally similar models that have different string connectivity can exist to explain the functionality of the extensor mechanism . We reckon that additional information may be necessary to identify the details of structurally diverse models that exhibit similar functionality . This raises the important issues of uniqueness and observability , which are central to computational model inference—and critical in the context of biological populations that naturally exhibit anatomical variability . Some 2D sub-topologies can be equivalent to each other under certain parameter and loading sets [32] . The load transfer patterns in these substructures can be similar despite their structural diversity . Due to these equivalencies , some substructures in a model can be transformed into one another resulting in a number of similar models . Consequently , multiple local optima as opposed to single global optimum may exist in the design space . From the computational perspective , our use of populations of models forrandom mutation based hill climber search is very much conducive to the maintenance of model diversity ( i . e . , alternative hypotheses ) to understand the uniqueness and observability of model topologies . This allows our search in this large dimensional space to proceed along multiple alternative paths that do not favor any particular local minimum . In all of our results with synthetic or anatomical networks , we find families of solutions: multiple different , yet functionally similar , topologies . This suggests that ( i ) additional data are necessary to further constrain the search , ( ii ) that different functional domains ( such as deformation during finger flexion ) are necessary to make the differences across various models observable , or ( iii ) that there are indeed functionally similar implementations for the domain of behavior that we studied ( load transmission in this case ) . The latter idea is quite intriguing from the evolutionary perspective as it agrees with the well-documented natural variability in the gross anatomical structure of the extensor mechanism across humans [19] . It suggests that , for the types of anatomical structures achievable with collagen fibers , anatomical variability in human population may not be functionally detrimental , and may in fact enable a wider variety of adaptations in future generations . Thus , the popular representation of Winslow's Rhombus ( Fig . 4 ( a ) ) may no longer be considered a uniquely valid or accurate representation of the extensor mechanism . Furthermore , we observe that only a few models inferred from the human extensor mechanism concur with its classical description . While Winslow's anatomy book [16 , pub . 1732] has no illustrations , the first graphical string model depiction of the extensor mechanism is by Zancolli [17] and An et al [38] wherein the tendinous network is suggested to have crossover tendons that slide past each other . In most models obtained with our experimental loading conditions , sliding crossover tendons are seldom observed and they do not grant particularly higher fitness—even though the primordial mesh was specifically designed to allow for them . In our detailed dissections of the extensor mechanism as well , crossover tendons were not clearly and independently observed . From the clinical perspective , damage to this network can cause severe dysfunction of manipulation ( e . g . , mallet finger , swan-neck or boutonnière deformity [42]–[48] ) , which can significantly affect a person's quality of life . Both non-operative and surgical methods [49 for a brief overview , 50–55] are reported following which subjects undergo rigorous , rehabilitative tendon gliding exercises [49] . Through accurate structural and parametric prediction of the target biological network with the aim to determine sources of injuries and/or deformities specific to the patients , or the classification of patients into structurally/functionally similar subgroups , our methods and results can help plan surgical/corrective strategies more effectively ( e . g . , multiple trial-and-error procedures can be avoided ) requiring less rigorous and more cost effective rehabilitative follow-ups . By allowing the inference of functional interactions in musculoskeletal systems , they are also relevant to the understanding of the functional adaptations that led to the evolution of the modern human hand and body .
|
In science and medicine alike , one of the critical steps to understand the working of organisms is to identify how a given individual is similar or different from others . Only then can the specific features of an individual be distinguished from the general properties of that species . However , doing enough input-output experiments on a given organism to obtain a reliable description of its function ( i . e . , a model ) can often harm the organism , or require too much time when testing perishable tissues or human subjects . We have met this challenge by demonstrating that our novel algorithm can accelerate the extraction of accurate functional models in complex tissues by continually tailoring each successive experiment to be more informative . We apply this new method to the problem of describing how the tendons of the fingers interact , which has puzzled scientists and clinicians since the time of Da Vinci . This new computational-experimental method now enables fresh research directions in biological and medical research by allowing the experimental extraction of accurate functional models with minimal damage to the organism . For example , it will allow a better understanding of similarities and differences among related species , and the development of personalized medical treatment .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"biotechnology",
"bioengineering",
"computer",
"science",
"computer",
"modeling",
"biology",
"computational",
"biology",
"engineering"
] |
2012
|
Functional Inference of Complex Anatomical Tendinous Networks at a Macroscopic Scale via Sparse Experimentation
|
The multifactorial nature of disease motivates the use of systems-level analyses to understand their pathology . We used a systems biology approach to study tau aggregation , one of the hallmark features of Alzheimer's disease . A mathematical model was constructed to capture the current state of knowledge concerning tau's behavior and interactions in cells . The model was implemented in silico in the form of ordinary differential equations . The identifiability of the model was assessed and parameters were estimated to generate two cellular states: a population of solutions that corresponds to normal tau homeostasis and a population of solutions that displays aggregation-prone behavior . The model of normal tau homeostasis was robust to perturbations , and disturbances in multiple processes were required to achieve an aggregation-prone state . The aggregation-prone state was ultrasensitive to perturbations in diverse subsets of networks . Tau aggregation requires that multiple cellular parameters are set coordinately to a set of values that drive pathological assembly of tau . This model provides a foundation on which to build and increase our understanding of the series of events that lead to tau aggregation and may ultimately be used to identify critical intervention points that can direct the cell away from tau aggregation to aid in the treatment of tau-mediated ( or related ) aggregation diseases including Alzheimer's .
Despite the fidelity of protein folding and the operation of quality control mechanisms to eliminate misfolded and otherwise abnormal proteins , a number of diseases can be traced to defects in these processes [1] . Among them are many neurodegenerative disorders , including the tauopathies , which are characterized by the intraneuronal aggregation of tau protein and of which Alzheimer's disease ( AD ) is an example . Preventing aggregation to halt or reverse cognitive decline is the goal of many drug discovery programs , but effective , long-term treatments have yet to be discovered [2] . A convincing body of evidence implicates defective tau processing and the formation of intraneuronal tau aggregates in cognitive decline . Mutations in the gene encoding tau protein are directly responsible for a number of genetic conditions collectively called primary tauopathies , among which is frontotemporal dementia and Parkinsonism linked to chromosome 17 ( FTDP-17 ) [3] , [4] . Tau pathology is also present in a large number of conditions whose cause cannot be traced to mutations in the gene encoding tau , including traumatic brain injury and repeated head trauma ( dementia pugilista ) from contact sports [5]–[7] as well as Alzheimer's disease , and has been observed with and without amyloid-beta pathology . Post-mortem assessment of the neurofibrillary tangle load in the brains of demented human patients showed that the severity of dementia was well correlated with the presence of tangles , a finding that argues strongly that tau plays a central role in disease progression [8]–[10] . In addition , the deficits in spatial learning and memory observed in mouse models expressing human APP can be ameliorated by reducing endogenous , wild-type tau [11] , which also protects against early mortality and inhibits excitoxicity; this finding is supported by more recent experiments in an AB-forming mouse model [12] . Taken together , these studies point to tau as a key causative factor in neurodegeneration and suggest that the tau pathway itself represents a reasonable therapeutic target for diseases in which the abnormal tau processing pathway is triggered . Tau is a neuronal , microtubule-associated protein ( MAP ) whose physiological function is to regulate microtubule dynamics ( Figure S1 ) . Alternative mRNA splicing yields 6 protein isoforms that are divided into two broad classes according to whether they contain 3 or 4 microtubule binding repeats; they are known as the 3R and 4R isoforms , respectively [13] , [14] . The 4R isoforms have a higher affinity for microtubules and greater tendency to aggregate [15]–[18] . A phospho-protein with nearly 30 phosphorylation sites , tau's biological activity is also governed by its phosphorylation state . In a healthy neuron , tau contains 2–3 moles of phosphate per mole of tau and is found almost entirely bound to microtubules [19] . In degenerating neurons , kinase and phosphatase activity is dysregulated and an abnormal variant containing 5–9 mol phosphate/mol tau is generated . While normal amounts of physiological tau are maintained , high amounts of hyper- and abnormally phosphorylated tau with low affinity for microtubules and resistance to degradation are generated [20] . These tau species dissociate from microtubules and collect in the cytosol , where they subsequently misfold and aggregate . The presence of ubiquitin , a molecular tag that facilitates degradation by the proteasome , in the aggregates suggests a failure of the quality control systems that clear aberrant proteins , contribute to the accumulation of abnormal tau and the neurofibrillary tangles [21] . Experiments demonstrating that the ubiquitous , constitutively expressed chaperone Hsc70 binds tau support this view , as Hsc70 is a chaperone known to mediate a protein triage decision that results in either refolding or degradation [22] . When the cell's quality control systems fail , tau aggregates and eventually neuron death occurs . The long , insoluble filaments that form may serve as a ‘stop-gap’ measure to protect the cell from adverse consequences by sequestering toxic intermediates . However , the actual toxic moiety among various pathological tau states has not been conclusively determined . The multifactorial nature of disease motivates our systems biology approach to understanding tau pathophysiology . We have developed a computational model that represents the network of interactions in which tau is involved as a system of ordinary differential equations that describe the deterministic chemical kinetics . The model was tuned to capture observed behavior in a healthy neuron and an aggregation-prone neuron . Although the class of tauopathies contains several diseases , specific experimental data from Alzheimer's disease studies informed this model . Sensitivity analysis tools were used to interrogate the model and ascertain the relative contributions of each component in the tau pathway from its synthesis to its post-translational modifications , to its degradation . Within both populations of neurons , and particularly the aggregation-prone population , we found ultrasensitive cellular conditions that are likely to be resistant to rescue .
As one of the first attempts at in silico simulation of tau pathophysiology , a mathematical model representative of the known biology was established within the limitations of the available data ( Figure 1 ) . Although this model is necessarily a simplified version of reality , it captures essential features of the known tau network and could be easily extended to incorporate additional detail as new data is generated . Among the key components are the 3R and 4R isoforms of tau . Alternative splicing of other tau exons was not considered in the model; therefore we modeled two species to be representative of the 3R and 4R classes . The isoform classes were divided into a number of phospho-states; although there are likely many disease-relevant phospho-isoforms , for simplicity , each 3R and 4R form was divided into in a minimally phosphorylated , normally phosphorylated , or abnormally phosphorylated/conformationally altered state . Minimally phosphorylated 3R and 4R tau are constitutively produced in a single reaction that captures transcription and translation . Specific tau kinases and phosphatases such as GSK3-β and PP5A were not explicitly included in the model . The kinetics of phospho-isoform conversion were modeled using Michaelis-Menten kinetics and based on in vivo data , from which the bounds on the Michaelis-Menten constants and the dependence of the kinetics on the phospho-state were derived . Tubulin , the building block of microtubules , was included although the total pool of tubulin with which tau interacts was considered constant throughout these analyses . Makrides and colleagues [23] monitored the in vitro reaction kinetics between tau and pre-assembled microtubules and found that a two-step mechanism in which either tau or tubulin underwent a conformational change before binding fit the data best; we employed that two-step mechanism here , assuming the conformational change occurred in the tau protein prior to association . Tau degradation by the proteasome has been shown both in vitro and in vivo in neuronal cell culture [24] , and has also been shown that natively unfolded tau can be degraded by the 20S proteasome in a non-ubiquitin dependent manner [25] . This degradation process was modeled with first order kinetics and a constant pool of proteasomes . Abnormal 3R and 4R tau are bound by the chaperone Hsc70 [22] , which mediates a choice between rescue and ubiquitin-dependent degradation . We assumed a simple , reversible binding reaction that does not involve ATP; although Hsp70 is usually an ATP-dependent chaperone , recent evidence suggests it binds tau independently of ATP [22] . Rescue is facilitated by the chaperone Hsp90 [26] , [27]; although other proteins such as the peptidyl-prolyl isomerase PIN1 are likely to participate in this pathway [28] , we assumed a simple mechanism by which Hsp90 binds abnormal tau . In this simplification , abnormal tau is dephosphorylated and restored to its normal functional form upon Hsp90 binding , and is released to re-bind microtubules . CHIP , an Hsc70-interacting protein and E3 ligase , links the chaperone and degradation machineries and shuttles abnormal tau to the 26S proteasome [29] , [30] . BAG-2 binds with the CHIP-Hsc70-Tau complex and subsequently dissociates with CHIP , restoring the Hsc70-Tau complex B , acting to potentially rescue tau from CHIP-mediated degradation [31] , [32] . Alternatively , CHIP and Hsc70 can release ubiquitinated , abnormal tau in a single-step reaction , after which tau is degraded . Because tau has been shown to be abnormally phosphorylated prior to ubiquitination , we assumed that only the abnormal tau species could be degraded in a ubiquitin-dependent , chaperone-assisted manner [33] . Aggregation is an alternate pathway down which abnormal tau can travel . Tau aggregation was modeled with the nucleation-elongation reaction mechanism and kinetics established by Congdon et . al . They monitored in vitro tau fibrillization and found that a tau dimer acted as the nucleus for the reaction , best fitting the experimental data and providing a good prediction of the length distribution of aggregates through time [34] . We assumed that only abnormal , ubiquitinated tau could polymerize as the presence of ubiquitin in tau aggregates is well-established [35] , [36] and full-length , wild-type tau does not aggregate readily under physiological conditions in vitro in the absence of polymerization promoters because it is hydrophilic and relatively unstructured [18] . Although normal tau may be sequestered by abnormal tau and thus aggregate [37] , this mechanism was excluded from our construction due to a paucity of available data . Furthermore , the paired helical filaments into which abnormal tau aggregates in Alzheimer's disease patients contain 3–4 times more phosphate than physiological tau and the level of phosphorylation observed in soluble amorphous tau is similarly elevated , suggesting that paired helical filaments are primarily comprised of abnormal tau [19] , [38] . In a study of brains from patients diagnosed with the tauopathy FTDP-17 , in whom tau is mutated , the insoluble fraction was observed to have a much greater ratio of mutated tau than normal tau [39] , also supporting this assumption . The effect of macromolecular crowding was also neglected for parsimony . Excluding these mechanisms from our model is likely to have little effect on the qualitative results , resulting in a re-scaling of parameters but not substantially changing the qualitative behavior and overall conclusions . Mass action kinetics described all reactions in the network except the phosphorylation and dephosphorylation reactions , which were described by Michaelis-Menten kinetics . For each species represented by our model , an ordinary differential equation that describes the species time-evolution was constructed as illustrated in Eq . S1 . In total , the network contains 84 reactions , 93 parameters , and 45 states ( i . e . , differential equations ) . A full listing of the states , reactions , parameters , and differential equations can be found in Tables S1 and S2 . Parameter space for the healthy and aggregation-prone identifiability and optimization steps is different , as the chaperone and degradation machinery was considered to be operating homeostatically . As a result , before initiating each stage of the optimization , an a priori identifiability analysis was completed . Correlation matrices were calculated at 1024 quasi-random points in the relevant parameter space , each matrix was weighted based on the objective function value determined at its corresponding location in parameter space , and then the matrices were averaged to establish pseudo-global a priori identifiability . The results of both stages of this analysis confirm that the proposed model is a priori identifiable and , by extension , structurally identifiable ( Figures S2 and S3 ) . To improve the efficiency of the optimizations , we did remove three parameters ( k1 , k84 , k10 ) from the first stage of the procedure as they were highly correlated ( >0 . 95 ) . In the next step , we optimized parameters to achieve steady-state behavior that represents healthy neuron function . Parameters associated with phosphorylation and dephosphorylation , microtubule binding and release , synthesis , and ubiquitin-independent degradation were estimated . We also estimated ATP synthesis and depletion . Parameters were generally assumed isoform-independent , with the exception of the microtubule binding parameters and aggregation parameters . Because evidence suggests that 4R tau has a greater affinity for microtubules [15]–[17] and for aggregation , these parameters were increased relative to the corresponding reactions involving 3R tau . Estimating chaperone and degradation parameters was excluded from the healthy state computations because under normal conditions Hsc70 does not bind microtubule bound tau [22] . Although Hsc70 may bind free normal tau species , these species represent a small portion of total tau and thus the model was simplified to exclude these minor interactions . The objective function that mathematically quantifies the behavior of a healthy neuron was constructed to reflect known quantitative experimental data . It is well-established that aberrant tau species are undetectable in normal neurons; thus we require that free and microtubule-bound aberrant tau is minimized . From measurement of total tau in human brain homogenates [40] , and assuming total protein concentration is 500 mg/ml [41] , the total neuronal concentration of tau protein was estimated to be 5–10 µM , consistent with many reported values . In adult human brain that is not afflicted by Alzheimer's , the ratio of 3R to 4R tau was determined to be 1:1 [14] , [42] . The affinity of normal tau for microtubules is 16 nM [23] and at least 80% of the total neuronal tau is bound to microtubules . These data are quantified in a cost function that sums the squared percent difference between the model result and the experimental results . Several of the objectives in our cost function are “fuzzy” , i . e . they allow states to achieve a range of values without penalty , rather than admitting only a single value without penalty . This construction is a better representation of biological systems than those that force the system to converge to a single value for objectives such as species concentrations , because it captures the intrinsic variability of these systems and it results in a large population of equally feasible parameter sets . A global solver that uses a scatter-search method followed by refinement with a local , gradient-based method handles the flat expanses of the search space . The sample code given in Eq . S2 demonstrates the implementation of this type of multi-objective , fuzzy cost function . Necessarily , the solution in this case is not unique . Therefore , a set of 2500 optimizations was performed in which the model was run to steady-state , then evaluated against these objectives to generate a set of equally valid parameter vectors with which to initialize the model ( Dataset S1 ) ; qualitatively , the number of optimizations does not affect the results . For this stage , the only species for which an initial condition was needed was microtubules; we assume 15 µM tubulin is present in abundance and excess over tau , and therefore do not include synthesis and degradation reactions for them . A total of 31 parameters were estimated . The resulting set of parameter vectors represents a population of neurons that behave in a healthy fashion and provides a way of evaluating the range of possible responses the system can display . The median sensitivity of the population to perturbations in the parameters was calculated at steady-state , to provide insight into the triggers that disturb the system's homeostasis ( Figure 2 ) . The 95% confidence interval for the sensitivities was also calculated ( Figure S4 ) . Because the ratio of 3R to 4R tau is 1:1 in healthy neurons , the results for each are equivalent . The identifiability of the sensitivity coefficients is defined by the span of the confidence interval; if the interval does not contain zero , the coefficient is considered identifiable . Although some small sensitivity coefficients are identifiable , most are not and the converse is true for larger coefficients , particularly those >0 . 5 ( Figure S5 ) . We find that changes in synthesis rates have the greatest positive impact on in silico homeostasis , while the perturbations in ubiquitin-independent degradation strongly and inversely alters the distribution of tau species . The situation for sensitivity to phosphorylation and dephosphorylation is more complex . Strong influences of this part of the network are found , but they do not act in concert . For example , aberrant 3R tau has a positive correlation with perturbations to the rate at with normal 3R tau is phosphorylated but it has an inverse relationship with the Michaelis-Menten constant . A similar situation is seen with bound tau states . The relationship between the microtubule interactions and tau distribution is similarly complex . In Figure 3 , the distribution of sensitivity coefficients within the healthy population is shown . The coefficients for each state were consolidated and transformed by the cube root , to accommodate the large scale and preserve the sign information of the coefficients . For all states , >99 . 9% of the coefficients fall below a value of 10 , but in a few important cases high sensitivity to perturbations is observed . These individuals are relatively more vulnerable and less robust than the bulk of the population . For each model of a healthy neuron , we established a corresponding aggregation-prone model . The two models are coupled through the microtubule binding and release parameters . Synthesis , degradation , and phosphorylation and dephosphorylation were re-estimated because these activities are known to be altered in neurons containing tau aggregates . In addition , parameters associated with the chaperone and degradation machinery were estimated . The objective function that quantifies the behavior of an aggregation-prone neuron is based on the data from several experiments . Quantification of tau in adult human brains affected by Alzheimer's was compared to that in control and showed that normal tau concentration was unaltered , but total tau concentration was 4–8 times normal tau; the increase is in the form of aberrant tau [40] . The critical concentration for aggregation is reported to be 0 . 2 µM [34]; necessarily , ubiquitinated tau approaches this concentration in an aggregation-prone neuron . The results of two silencing experiments were used to finalize the construction of the cost function corresponding to the aggregation-prone population [43] . In these experiments , silencing RNA was used to reduce the levels of Hsp70 and Hsp90 in COS-1 cells over-expressing human tau and the resulting effect on cytosolic ( unbound ) and microtubule-bound tau was assessed . A 50% reduction in Hsp70 resulted in a 5% decrease in unbound tau and a 75% decrease in bound tau , while a 75% reduction in Hsp90 resulted in a 10% decrease in unbound tau and a 70% decrease in bound tau [43] . The objective function was constructed as previously , resulting in the minimization of a function that is the sum of squared percent differences . For the “fuzzy” objectives , no cost was assigned if the model simulated a result in the allowable range of values . Each result from the tuning of a neuron to healthy behavior was used to seed an optimization run designed to generate aggregation-prone behavior . For each run , the model was initialized to the steady-state concentrations achieved by the corresponding model of a healthy neuron . The simulation was run until quasi-steady-state was achieved and evaluated against the objective function to find parameters that instantiate an aggregation-prone model ( Text S1 ) . In general , a single primary route to establish the aggregation-prone behavior was not obvious . Rather , the nature of the changes required to establish aggregation-prone neurons was multifactorial , although definite trends were observed in a small subset of the parameters ( Figure 4 ) . Confidence intervals ( 95% ) were calculated and show just three identifiable trends; synthesis of 4R tau is generally increased while chaperone-independent degradation of normal 4R tau decreased , and the relative rate at which microtubule-bound , normal 4R tau was phosphorylated was elevated . Relative rate is a more meaningful measure of the change in phosphorylation and dephosphorylation processes and thus the metric on which we focus . The consistency with which these effects were observed suggests such behavior is likely to play a key role in initiating the pathological changes seen in vivo . As this result is consistent with the known increase in tau levels and decrease in proteasomal activity , and increased kinase that occurs in affected neurons , it provides a measure of validation for the model and encourages efforts to test the subsequent conclusions drawn from its behavior . In all cases , multiple perturbations in the rates of synthesis , degradation , and phosphorylation and dephosphorylation were required to induce an aggregation prone state . The median sensitivity of the aggregation-prone population was calculated and the 95% confidence interval of the coefficients was used to determine their identifiability ( Figures S6 , S7 , S8 ) . As in the healthy population , synthesis and degradation are important processes with respect to tau distribution . Microtubule binding and phosphorylation and dephosphorylation are relatively less important in this population , although particularly for 3R tau a number of reactions in these processes are sensitive to tau distribution . Chaperone system reactions , on the other hand , do affect the behavior of the aggregation-prone population . Interestingly , the sensitivity to the aggregation reactions is only evident for aggregates; if the toxic moiety is actually soluble , aberrant tau , as it is increasingly thought , and not the aggregates then this has important ramifications for the selection of drug targets as the aggregation reactions have little effect on soluble tau . To compare the aggregation-prone and healthy populations , the ratios between the sensitivity coefficients in each pair of matched individuals was calculated and the medians are shown in Figure 5 . The aggregation-prone population exceeds twice the sensitivity of the healthy population 26% of the time and the magnitude of 46% of the median coefficients it is 2 fold lower . Notably , in nearly 24% of cases , the sign of the median sensitivity coefficient changes . This sign change is a striking and important phenomenon , as it suggests that the fundamental nature of the system's behavior changes during the transition from a healthy to an aggregation-prone state . It also suggests that the effect of changing conditions in the cell , due to drug treatment , for instance , depends on the state of the system . For example , the sensitivity of free and bound abnormal 4R tau species to phosphorylation shows a sign change; therefore , the efficacy of a treatment designed to influence phosphorylation reactions may depend upon the state of the system when treatment is initiated . Evaluation of the distribution of the coefficients revealed subset of individuals with very large magnitude sensitivities to changing parameters , or ultrasensitivity ( Figure 3 ) . As with the healthy population , >99% of individuals were more moderately impacted by parameter perturbations , but the ultrasensitive individuals were of a much higher magnitude in this population . Additionally , this feature of ultrasensitivity was sharp and occurred after the accumulation of tau aggregates began . Systems such as this represent large obstacles to treatment; although sensitivity is required of a suitable drug target , the complex nature of the system's behavior in combination with ultrasensitivity is a challenging control problem and will make it difficult to re-establish homeostasis in these individuals . Disease progression independent of treatment is also significantly impacted by ultrasensitivity; cognitive decline is likely to be faster due to the fragile nature of this kind of network .
The in silico model developed to describe tau pathophysiology displays the very features of robustness and fragility that exist in real biological systems and these concepts are key to our understanding of the tauopathies . Indeed , the concept of robustness provides a framework in which disease can be understood as the inevitable consequence of a breakdown in the systems that normally maintain functionality [44] . Because these systems are complex , highly coupled , and nonlinear , their behavior is difficult to predict and systems-level approaches are required to understand and treat disease [45] . The population of healthy neurons is considered to be robust in several ways . The model generates healthy behavior in a relatively large domain of parameter space , a necessary property to maintain a phenotype given the inherent variation and noise in all biological systems . Likewise , the healthy population is robust and demands a vectorial assault to become pathological , as a multitude of perturbations to synthesis , degradation , and phosphorylation and dephosphorylation are required to generate a corresponding population of aggregation-prone neurons . In contrast , the aggregation-prone population is generally more sensitive to perturbations than the healthy population , as might be expected for a pathological phenotype ( Figure 3 ) . Moreover , the change in sign of a quarter of the sensitivity coefficients suggests that the fundamental behavior of this nonlinear system changes during the transition from healthy to aggregation-prone conditions . This change has implications for the drug discovery process; targeting such parts of the network is likely to be ineffective unless the timing is carefully considered . The case study shown in Figure 6 illustrates this point . In this individual , the binding of normally phosphorylated , 3R tau to microtubules was perturbed 5-fold and the concentration of microtubule bound , unphosphorylated 4R tau monitored in both the healthy and aggregation-prone states; the parameter perturbation is an in silico means of simulating drug treatment . Not only is an inverse response observed in each condition , but the qualitative response of the healthy neuron is in direct opposition to that of the aggregation-prone neuron . As the healthy and aggregation-prone neurons circumscribe the range of behaviors expected as a tauopathy advances , it logically follows that the sensitivity of relevant proteins to parameter perturbations switches at some point during disease progression . Such phenomenon may play an important role in the effectiveness of any particular drug , whose impact may be exactly the opposite of that intended and indeed even validated in experimental models . Therefore the identification of potential drug targets could be guided both by the identification of the perturbations that contribute to generate the diseased state and by the analysis of the parameter sensitivities in the healthy and diseased states . To minimize undesirable system-dependent effects , we suggest to target parameters for which the sign of the sensitivity coefficients does not change between the healthy and aggregation-prone states . Having identified synthesis , degradation , phosphorylation/dephosphorylation as keys to disease progression , the sensitivity coefficient associated synthesis and degradation reactions appeared to have a minimal number of changes of sign compared to the ones of phosphorylation/dephosphorylation reactions ( Figure 5 ) . From that point of view , synthesis and degradation appear to be preferential drug targets within the tau network . A subset of the aggregation-prone population displays extreme fragility ( Figure 3 ) . This ultrasensitivity arises in the models of aggregation-prone neurons , and thus has implications for disease progression; the typical delay in diagnosing neurodegenerative diseases makes this phenomena potentially important with respect to treatment . While it is important to develop drugs that target sensitive points in biological networks , the widespread ultrasensitivity and nonlinearity observed in a subset of the population are likely to make the response of these systems difficult to predict or control , and they are likely to be highly resistant to rescue . The robustness of the tau network and the multifactorial nature of its vulnerability to pathological change presents a challenge to the selection of drug targets , and for a subset of patients the disease is likely to be nearly impossible to reverse after the network becomes ultrasensitive . The model analysis also suggests that stalling or reversing tau pathophysiology will be further complicated by the timing at which the intervention is begun; a treatment may have an opposite effect on the system than is expected due to the sign inversion observed for some sensitivity coefficients . The systems biology approach we have taken here has highlighted the complex , nonlinear behavior that cellular networks can display and suggests the difficulties the pharmaceutical and biotechnology industries will face in attempting to treat diseases associated with their aberrant functioning . By modeling both the physiological and the pathological functioning of the network governing tau function , we have shown that the biological response to a perturbation is dependent on the condition of the network and that , therefore , the time at which a compensatory perturbation is made is potentially significant . This implication is particularly relevant in therapeutic treatment timing and approach . The population-based analyses we have completed also highlights the importance of variability in the study and treatment of disease and the need to characterize the variability of the network components , such as reaction rates , to more fully elucidate its nature . Such variations are distinct from stochastic variation and the extent of the variability is likely dependent on the biological network and the particular network component . From a modeling perspective , in silico populations can be created for any model in a straightforward manner , by retaining not just a single optimization result but a number of results that fit the data almost equally well . As new experimental data is generated , the variation within the in silico populations will become more constrained and approach that seen in vivo . With respect to the optimization results , they suggest an approach that considers fitting matched measurements from the same individuals , for example if data was collected from individual animals over time , rather than taking a conglomerated value over measurements from multiple individuals . The computational , model-based approach to exploring cellular networks demonstrates a new paradigm for understanding disease that is likely to become increasingly effective as high-throughput and sequencing technologies quickly generate large databases of experimental data from which progressively more detailed , accurate models can be built .
Current knowledge about the molecular biology of tau protein was integrated into a deterministic , kinetic model that was realized as a set of 45 ordinary differential equations ( ODE's ) ( Tables S1 and S2 ) and implemented in MATLAB ( Mathworks , Cambridge , UK ) . For each species , a differential equation was constructed from the rate equations for all reactions in which the species is involved; the reactions were modeled with mass action and Michaelis-Menten type kinetics . For example , a representative equation for the time-evolution of unphosphorylated tau is given by Eq . S1 ( Text S1 ) , which describes the change in concentration of unphosphorylated tau due to its synthesis , degradation , a conformational change that precedes microtubule binding , the restoration of the original protein conformation , phosphorylation , and dephosphorylation . To validate the model construction effort , we used the method of Jacquez and Greif [46] to evaluate the a priori identifiability of the model and extended it to develop a suitable substitute for structural identifiability , as direct methods for evaluating structural identifiability are not feasible for large , nonlinear models such as this . In the traditional approach to a priori identifiability analyses , an iterative process of estimation and identifiability analysis is employed , reducing the number of parameters in the model after each iteration until the model is entirely identifiable [47] . We used a pseudo-global extension of this approach to diminish the parameter dependence of the results . First , Sobol' Low Discrepancy Sequences were used to generate 1024 points in parameter space . For each point , an in silico experiment in which tau was allowed to equilibrate for 2 hours after being induced in a tau-free system was simulated . We assumed all states were measurable and measured at 30-minute intervals during the 2 hour experiment . In addition , the local parametric sensitivity of the system was evaluated . From these simulated data , the correlation matrix Mc that establishes a priori identifiability was calculated according to Eq . S3 ( Text S1 ) . Identifiable systems have correlations strictly < |1| . Here , the average correlation matrix is used to ascertain the identifiability of the system . Because the parameter sets were randomly generated , the resulting systems do not necessarily display biologically relevant behavior; therefore , the optimization objective function was calculated at each point in parameter space and used as weighting factors in calculating the average correlation . Given the model structure we established and the bounds on the parameter ranges , we can conclude that the model is a priori identifiable , but the high correlations between some parameters suggest that they might be difficult to estimate and therefore one parameter in each pair with a correlation >|0 . 95| was removed from the optimization and fixed to its nominal value . Using this framework , that reduces parametric dependence and assumes all states are experimentally measurable , a priori identifiability is an acceptable proxy for structural identifiability . However the converse is not true and no conclusions can be drawn from a non-a priori identifiable system , as different experiments could reveal that the system is indeed structurally identifiable . The model parameters were numerically fit using a hybrid stochastic-deterministic global optimization method [48] , [49] that is based on well-established scatter search methods and implemented as a set of MATLAB functions , which are freely available on the authors' website and require only a single function call in MATLAB to implement . In brief , the method iterates between a global scatter search and local refinement of the solution using traditional methods; in this case we used MATLAB's fmincon , which is gradient-based technique , to perform this refinement . Although some experimentally derived kinetic data was available , it originated from heterogeneous sources including in vitro and in vivo platforms and under different experimental conditions . Therefore , generous bounds were used to define and explore parameter space . To assess the effect of parameter perturbations on the steady-state concentrations of protein in the healthy population and quasi-steady-state ( due to the polymerization reaction ) concentration in the aggregation-prone population , the local , relative sensitivity of this system , given by Eq . S4 ( Text S1 ) , was evaluated . The relative sensitivity coefficient gives the dependence of the protein concentration , “xi” , on a parameter , “pj” and is normalized with respect to the parameter and state values to facilitate . The non-normalized coefficients are calculated by applying the chain rule to Eq . S4 ( Text S1 ) , which results in a set of ordinary differential equations that give all the sensitivity coefficients associated with this system ( Eq . S5 , Text S1 ) by simultaneous integration of these sensitivity ODE's and the model ODE's in MATLAB . The sensitivity coefficients at steady-state were collected into a matrix , Sx , of size Nx ( number of states ) by Np ( number of parameters ) .
|
Neurodegenerative disorders , particularly the tauopathy Alzheimer's disease , affect millions of people and cost billions of dollars a year in healthcare costs . Although effective treatments to delay or reverse cognitive decline are still unavailable , several approaches to address this medical need are being pursued . One such strategy involves ameliorating aberrant tau processing , as the characteristic tau tangles associated with the tauopathies are well-correlated with cognitive dysfunction , genetic mutations in tau lead directly to neurodegeneration , and experiments in animal models have yielded promising results . Two avenues are currently being explored: inhibition of kinase activity to reduce the presence of aberrant , hyperphosphorylated tau and means to prevent and reduce tau aggregation . We have taken a systems biology approach to understanding tau pathophysiology , creating a mathematical model to quantitatively explore the vulnerabilities in the tau network and identify effective intervention points . Our analysis of the resulting in silico neuron populations , representing healthy and aggregation-prone neurons , highlights the multifactorial nature of the disease and provides insight into pathological triggers and the timing of treatment , which will be an important element in effectively treating patients .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computational",
"biology/systems",
"biology",
"neurological",
"disorders"
] |
2010
|
Vulnerabilities in the Tau Network and the Role of Ultrasensitive Points in Tau Pathophysiology
|
Leishmania donovani , a protozoan parasite is the major causative agent of visceral leishmaniasis . Increased toxicity and resistance to the existing repertoire of drugs has been reported . Hence , an urgent need exists for identifying newer drugs and drug targets . Previous reports have shown sirtuins ( Silent Information Regulator ) from kinetoplastids as promising drug targets . Leishmania species code for three SIR2 ( Silent Information Regulator ) related proteins . Here , we for the first time report the functional characterization of SIR2 related protein 2 ( SIR2RP2 ) of L . donovani . Recombinant L . donovani SIR2RP2 was expressed in E . coli and purified . The enzymatic functions of SIR2RP2 were determined . The subcellular localization of LdSIR2RP2 was done by constructing C-terminal GFP-tagged full-length LdSIR2RP2 . Deletion mutants of LdSIR2RP2 were generated in Leishmania by double targeted gene replacement methodology . These null mutants were tested for their proliferation , virulence , cell cycle defects , mitochondrial functioning and sensitivity to known SIR2 inhibitors . Our data suggests that LdSIR2RP2 possesses NAD+-dependent ADP-ribosyltransferase activity . However , NAD+-dependent deacetylase and desuccinylase activities were not detected . The protein localises to the mitochondrion of the promastigotes . Gene deletion studies showed that ΔLdSIR2RP2 null mutants had restrictive growth phenotype associated with accumulation of cells in the G2/M phase and compromised mitochondrial functioning . The null mutants had attenuated infectivity . Deletion of LdSIR2RP2 resulted in increased sensitivity of the parasites to the known SIR2 inhibitors . The sirtuin inhibitors inhibited the ADP-ribosyltransferase activity of recombinant LdSIR2RP2 . In conclusion , sirtuins could be used as potential new drug targets for visceral leishmaniasis .
Acetylation and deacetylation of proteins have recently emerged as a major post-translational modification [1] . Initially described for N-terminal tails of histones [2] , reversible acetylation of proteins in the cytoplasm and in the mitochondria , suggest a central role for acetylation in regulatory mechanisms within and outside the nucleus of the cells [3] . The Silent Information Regulator ( SIR2 ) , the founding member of the family of sirtuins , was originally described as a regulator of transcriptional silencing of mating-type loci , telomeres and ribosomal DNA [4] , and also involved in the lifespan extension of yeast [5] . Since their discovery , SIR2-like genes , known as sirtuins , have been studied in most organisms , including plants , bacteria , and animals , where they play a vital role in promoting an organism’s health and survival [6] . All sirtuins are characterized by a core domain of ~250 amino acids that is highly conserved among different organisms [7] . They were first described to have NAD+-dependent deacetylase activity , that consumes the cofactor nicotinamide adenine dinucleotide ( NAD+ ) , yielding nicotinamide , O-acetyl ADP ribose ( AADPR ) , and the deacetylated substrate [8 , 9] . Bacterial and archaeal genomes express one or two sirtuins , but eukaryotes usually have multiple sirtuins , like yeast has Hst1–4 in addition to Sir2p , and humans have seven sirtuins named SIRT1–7 [10] , showing a discrete pattern of subcellular localization . However , being present in large number in different cellular compartments , other novel enzymatic activities for sirtuins have also been described like ADP-ribosyltransferase activity involving the transfer of a single ADP-ribosyl group from NAD+ to proteins [11] and NAD+-dependent desuccinylase and demalonylase activity [12] . ADP-ribosylation activity is known to be present in yeast SIR2 , the human SIRT2 , the mouse SIRT6 , the Trypanosoma brucei SIR2RP1 , Leishmania infantum SIR2RP1 and the Plasmodium falciparum SIR2 [11 , 13–15] . Desuccinylation and demalonylation activity has only been described for SIRT5 of human [12] . In summary , the evidence emerging from the literature is that the SIR2 proteins regulate the structure/function of multiple targets within the cell , by deacetylation , ADP-ribosylation and desuccinylation/demalonylation of specific lysine residues . Parasite sirtuins are distributed in all the phylogenetically defined sirtuin classes [16] . Sirtuins of protozoan parasites have both the canonical and atypical activities that contribute to both conserved and apparently unique functions . T . brucei SIR2RP1 , a nuclear protein that co-localizes with telomeric sequences and mini-chromosomes and has both deacetylase and ADP ribosylase activity , is known to be involved in DNA repair [13] . T . cruzi lacks SIR2RP2 but expresses SIR2RP1 and SIR2RP3 that have been found to be essential for the proliferation of the parasite , host- parasite interplay and differentiation among life cycle stages [17] . Besides this , TcSIR2RP3 has been studied as a possible target for the treatment of Chagas’ disease [18] . The parasite Leishmania donovani is a protozoan parasite , the major causative agent of visceral leishmaniasis [19] . The disease is fatal if left untreated . The parasite has a digenetic life cycle which alternates between mammalian immune cells and gut of insect vector , phlebotomine sand flies [20] . The current therapies are inadequate because of the increasing resistance to the currently used drugs and their serious side effects . Hence , an urgent need exists to develop new chemotherapeutic targets and agents against Leishmaniasis . Leishmania parasites are known to express three sirtuins; SIR2RP1 , SIR2RP2 , and SIR2RP3 . Out of the three , only SIR2RP1 has been characterized in L . major and L . infantum wherein it was found to be present in the cytoplasmic granules and indispensable for parasite survival [21 , 22] . It was found to have both NAD+-dependent deacetylase and ADP- ribosyltransferase activities unrelated to epigenetic silencing . The other two sirtuins , SIR2RP2 and SIR2RP3 , have not yet been characterized . Here , we for the first time report the functional characterization of an SIR2RP2 protein from L . donovani . SIR2RP2 is an NAD+-dependent ADP-ribosyltransferase . SIR2RP2 localizes to the mitochondrion of the promastigotes . Gene deletion mutations were attempted via targeted gene replacement methodology in order to elucidate the physiological role of LdSIR2RP2 . Deletion of LdSIR2RP2 from the parasite caused compromised mitochondrial functioning and accumulation of cells in the G2/M phase . The null mutants also had attenuated infectivity . Deletion of LdSIR2RP2 resulted in increased sensitivity of the parasites to the known sirtuin inhibitors .
All restriction enzymes , DNA-modifying enzymes , and DNA ladders were obtained from New England Biolabs . Plasmid pETM-41 and TEV protease were kindly provided by Dr Amit Sharma ( ICGEB , New Delhi ) . Protein markers were obtained from Thermo Fisher Scientific ( USA ) . Nicotinamide Adenine Dinucleotide [Adenylate-32P] ( 800Ci/mmol ) was purchased from American Radiolabeled Chemicals ( USA ) . Ni2+-NTA agarose and amylose resin were purchased from Qiagen and New England Biolabs , respectively . Sirtinol , nicotinamide , cambinol , and Ex-527 were obtained from Sigma-Aldrich ( USA ) . SIRT1 Fluorometric Drug Discovery Kit and SIRT5 Fluorometric Drug Discovery Kit were procured from Enzo Life Sciences ( USA ) . Other materials used in this study were of analytical grade and were commercially available . L . donovani Bob ( LdBob strain/MHOM/SD/62/1SCL2D ) was originally obtained from Dr Stephen Beverley ( Washington University , St . Louis , MO ) . Wild-type promastigotes were cultured at 22°C in M199 medium ( Sigma-Aldrich , USA ) supplemented with 100 units/ml penicillin ( Sigma-Aldrich , USA ) , 100 μg/ml streptomycin ( Sigma-Aldrich , USA ) and 5% heat-inactivated fetal bovine serum ( Biowest ) . Wild-type ( WT ) parasites were routinely cultured in media with no drug supplementations , whereas the genetically manipulated LdSIR2RP2 heterozygotes , in which one allele of LdSIR2RP2 gene has been replaced either with hygromycin phosphotransferase gene ( LdSIR2RP2/HYG ) or with neomycin phosphotransferase gene ( LdSIR2RP2/NEO ) and null mutants ( ΔLdSIR2RP2 ) were maintained in either 200 μg/ml hygromycin or 300 μg/ml paromomycin or both respectively . The ΔLdSIR2RP2 strain containing pSP72a-zeo-a-LdSIR2RP2 episome ( ‘add-back’ mutant cell line ) was maintained in 800 μg/ml zeocin , 200 μg/ml hygromycin and 300 μg/ml paromomycin . pSP72a-neo-a-GFP-LdSIR2RP2 transfected parasites were maintained in 40 μg/ml G418 . For characterising the mutant parasites phenotypically , cells were sub-cultured without selection antibiotics prior to experiments . The mouse monocyte-macrophage-like cell line J774A . 1 obtained from ATCC was cultured in RPMI 1640 ( Sigma-Aldrich , USA ) supplemented with 10% FBS and 100 units/ml penicillin and 100 μg/ml streptomycin at 37°C in humidified CO2 incubator . Sirtuin sequences of Leishmania and other kinetoplastids were retrieved from TriTrypDB database [23] and used for sequence analysis . The human sirtuin sequences were obtained from UniProt [24] . Subcellular localization prediction was made using a web version of WoLF PSORT [25] . Phylogenetic analysis was performed using MUSCLE [26] and Unrooted software . Multiple sequence alignment of these sequences was generated using a standalone version of CLUSTALW [27] using default parameters . For analyzing the conserved motif patterns and subfamily classification of the kinetoplastid sequences , multiple sequence alignment of only the kinetoplastid sequences was generated . The gene for LdSIR2RP2 was amplified by PCR using a forward primer with a flanking NcoI restriction site ( Forward: 5′ AACCATGGCTATGAGGCCGGCGGGGACGATC 3′ ) and a reverse primer with a flanking KpnI restriction site ( Reverse: 5′ AAGGTACCCTAGAGTTGAATCGTCTTGCGGCGGAAG 3′ ) from L . donovani genomic DNA . The ~ 963 bp amplicon encompassing the complete ORF of LdSIR2RP2 gene was cloned into a pETM41 expression vector . The recombinant vector pETM41-LdSIR2RP2 was transformed into Artic-Express DE3 strain . Expression of recombinant LdSIR2RP2 was induced in mid-exponential phase with 0 . 1 mM IPTG ( isopropyl ß D-thiogalactoside ) for 24 h at 10°C . Cells were harvested by centrifugation at 5 , 000 rpm for 15 min . The bacterial pellet was resuspended in lysis buffer ( 20 mM Tris-HCl pH 7 . 2 , 200 mM NaCl , 10% glycerol , 10 mM beta-mercaptoethanol , 0 . 1 mg/ml lysozyme , 2 mM phenylmethylsulfonyl fluoride and protease inhibitor cocktail ) . The cells were lysed by sonication and cleared by centrifugation at 6 , 000 rpm for 20 min . The cleared supernatant was applied to pre-equilibrated amylose beads ( NEB ) , and protein was eluted with buffer containing 20 mM Tris–HCl pH 7 . 2 , 200 mM NaCl , 10 mM beta-mercaptoethanol and 2 mM maltose . The tag ( MBP with His tag ) was removed by incubating with TEV protease at 4°C for 24 h . The NAD+-dependent deacetylase activity was estimated by using a commercially available SIRT1 Fluorometric Drug Discovery Kit ( Enzo Life Sciences ) . The enzymatic reaction containing rLdSIR2RP2 was carried out according to the manufacturer’s protocol . Briefly , the recombinant protein was incubated with 50 μM to 1 mM NAD+ and 64 μM fluorogenic peptide for 20 min at 37°C followed by incubation in the developer for 45 min at 37°C . The fluorescence was measured with excitation at 360 nm and emission at 460 nm using Varioskan Flash Multimode Reader ( Thermo Fisher Scientific ) . The enzymatic activities were calculated by plotting a standard curve using deacetylated standard available with the kit . Protein ADP-ribosylation assays were performed as described in [13] . Briefly , the reaction was carried out in a volume of 20 μl containing 2 . 5 μg of rLdSIR2RP2 , 2 . 5 μCi of [32P]NAD+ and 5 μg of calf thymus histones ( Sigma-Aldrich , USA ) or BSA as indicated . The reaction buffer contained 150 mM NaCl , 10 mM dithiothreitol ( DTT ) , 50 mM Tris-HCl pH 8 . 8 . The samples were incubated for 2 h at room temperature . The reactions were terminated by the addition of Laemmli gel loading buffer . The proteins were resolved on a 12% SDS-PAGE gel and visualized with the help of a PhosphorImager ( Fujifilm ) . For quantitative experiments , the reaction products were precipitated with 20% ( w/v ) trichloroacetic acid ( TCA ) , washed and counted for radioactivity after the addition of scintillation cocktail . The NAD+-dependent lysyl desuccinylase activity was done using a commercially available SIRT5 Fluorometric Drug Discovery Kit ( Enzo Life Sciences ) . The enzymatic reaction having 2 . 5 μg of rLdSIR2RP2 was carried out according to the manufacturer’s protocol . Briefly , recombinant protein was incubated with 50 μM to 1 mM NAD+ and 50 μM fluorogenic peptide for 60 min at 37°C followed by incubation in the developer for 15 min at 37°C . The fluorescence was measured with excitation at 360 nm and emission at 460 nm using Varioskan Flash Multimode Reader ( Thermo Fisher Scientific ) . The enzymatic activities were calculated by plotting a standard curve using desuccinylated standard available with the kit . Intracellular localisation of LdSIR2RP2 in L . donovani was detected using pSP72-α-neo-α-GFP-LdSIR2RP2 transfected promastigotes . For the construction of LdSIR2RP2-GFP fusion construct , the 963 bp ORF of LdSIR2RP2 was amplified from LdBob genomic DNA using sense primer: 5′ AACCATGGCTATGAGGCCGGCGGGGACGATC 3′ and antisense primer 5′ CTCTAGACTAGAGTTGAATCGTCTTGCGGCGGAAG 3′ . The restriction sites incorporated in the primers are underlined . The amplicon was cloned into BamH1 and Xba1 restriction sites of the vector pSP72-α-neo-α-GFP . Correct orientation and sequence fidelity of the inserts was verified by nucleotide sequence analysis . The recombinant vector was transfected by electroporation in wild-type L . donovani promastigotes according to the standard protocol [28] , and the transfectants were selected in the presence of 40 μg/ml G418 ( Sigma-Aldrich , USA ) . L . donovani , pSP72-α-neo-α-GFP-LdSIR2RP2 transfected promastigotes were used to detect the cellular distribution of LdSIR2RP2 . Log phase promastigotes were incubated with 1 nM MitoTracker red CMXRos ( Molecular Probes ) diluted in M199 medium for 20 min at 22°C in the dark . The cells were then washed with 1 X PBS and immobilised on poly-L-lysine-coated glass coverslips . Subsequently , the cells were fixed with 4% paraformaldehyde for 30 min , washed and permeabilized in 0 . 5% Triton X-100-PBS for 5 min . The cellular DNA was then stained with 1 μg/ml of DAPI ( Sigma ) for 30 min at RT . The coverslips were mounted on glass slides for visualisation . The cells were imaged by Andor Spinning Disk Confocal Microscope equipped with iXon Ultra 897 EMCCD camera at the required fluorescence excitation and emission wavelengths . The raw images were processed using FV10-ASW 1 . 7 viewer or Image J software . The co-localization analysis was done using JACoP ( Image J ) . Targeted gene replacement strategy was utilized for the inactivation of LdSIR2RP2 gene in L . donovani . A fusion PCR-based strategy was employed as reported earlier [29] . Briefly , LdSIR2RP2 flanking regions were amplified from LdBob genomic DNA and fused to antibiotic resistance cassettes: hygromycin phosphotransferase gene ( HYG ) or neomycin phosphotransferase gene ( NEO ) . The 5′UTR ( 907 bp ) of L . donovani LdSIR2RP2 was amplified with primers A & BHyg or primers A & BNeo ( Table 1 ) . The NEO gene ( 795 bp ) was amplified from pX63-NEO with primers CNeo & DNeo . The HYG gene ( 1012 bp ) was amplified from pX63HYG with primers CHyg & DHyg ( Table 1 ) . The 3′UTR ( 967 bp ) of L . donovani LdSIR2RP2 was PCR amplified from wild-type LdBob genomic DNA using primers EHyg or ENeo & antisense primer F ( Table 1 ) . The 5′UTR of L . donovani LdSIR2RP2 was then ligated to the antibiotic resistance marker genes by PCR using primers A & DHyg or A & DNeo . Finally , this fragment ( 5′UTR-marker gene ) was fused with 3′UTR of LdSIR2RP2 using primers A & F , yielding the fragment , 5′UTR-Hyg-3′UTR or 5′UTR-Neo-3′UTR . To generate the episomal ‘add back’ construct , the full-length LdSIR2RP2 coding sequence was amplified with primers; Forward: 5′ CCCTCTAGAATGAGGCCGGCGGGGACGATC 3′ and Reverse: 5′CCCAAGCTTCTAGAGTTGAATCGTCTTGCGGCGGAAG 3′ . This amplified product was then cloned into the Xba1 and HindIII restriction sites of the pSP72α-zeo-α vector to get pSP72α-zeo-α-LdSIR2RP2 complementation construct . All the fragments and constructs were sequenced for confirmation of their correct orientation and sequence fidelity . 5′UTR-Hyg- 3′UTR or 5′UTR-Neo-3′UTR linear fragments were generated through PCR amplification . The fragments were gel purified , and about 1–2 micrograms of each fragment were individually transfected by electroporation in wild-type L . donovani promastigotes according to the standard protocol [28] Depending on the marker gene , transfectants were selected either in the presence of 200 μg/ml hygromycin ( Sigma-Aldrich , USA ) or 300 μg/ml paromomycin ( Sigma-Aldrich , USA ) . The cells resistant to antibiotic selection were checked by PCR-based analysis for the correct integration of the replacement cassettes using primers shown in ( Table 2 ) . Thereafter , the second round of transfection was done to knock-out the other copy of LdSIR2RP2 gene . The genotypes of the LdSIR2RP2 mutants were confirmed by Southern blotting analysis using standard protocols [30] . The ‘add-back’ line ΔLdSIR2RP2/+ was created from the ΔLdSIR2RP2 null mutants by transfecting these parasites with pSP72α-zeo-α-LdSIR2RP2 episome . After transfection , these parasites ΔLdSIR2RP2/+ were selected in 800 μg/ml zeocin , 200 μg/ml hygromycin and 300 μg/ml paromomycin . Further , the genotype of the ‘add-back’ line ΔLdSIR2RP2/+ was confirmed by PCR analysis using primers mentioned in ( Table 2 ) . Growth rate experiments were conducted by inoculating stationary-phase parasites at a density of 1 × 106 cells/ml in standard M199 medium with 5% FBS in 25 cm2 flasks without particular selection drug and culturing at 22°C . The growth rate of each of the cultures was determined at 24 h intervals by using a Neubauer hemocytometer . Growth studies with each individual cell line were performed at least three times , and similar results were obtained consistently . J774A . 1 murine macrophage cell line was plated on poly-L-lysine-coated glass coverslips at a density of 5 × 105 cells per well in a 6-well flat bottom plate . The adherent cells were infected with stationary-phase promastigotes , at a ratio of 20:1 for 6 h . Excess non-adherent promastigotes were removed by incubation of the cells for 30 s in 1X phosphate buffer saline ( 1 X PBS ) . These were subsequently maintained in RPMI1640 containing 10% FBS at 37°C with 5% CO2 . Intracellular parasite load was visualised by Giemsa staining . 2 x 107 log phase promastigotes of WT , ΔLdSIR2RP2 and ΔLdSIR2RP2/+ were collected , washed twice with 1X PBS and then fixed in ice-cold 30% PBS/70% ( v/v ) methanol for 1 h at 4°C . The fixed cells were washed twice with ice-cold 1X PBS and then resuspended in 1 ml 1X PBS containing 100 μg/ml RNase and 20 μg/ml propidium iodide ( Sigma-Aldrich , USA ) . The cells were incubated for 45 min at 37°C in the dark . The samples were then analysed using BD Biosciences FACS Calibur system using BD Biosciences CellQuest software . For each sample , data for at least 20 , 000 events were collected . The resulting distribution of cells was analysed by the Modfit Lt . Software to determine the percentage of cells in G0/G1 , S , or G2/M phases of the cell cycle . The mitochondrial transmembrane potential was investigated using MitoTracker Red CMXRos ( Invitrogen ) . Logarithmically growing promastigotes ( 1 × 106 cells ) were incubated with Mito Tracker Red CMXRos ( 100 nM ) for 30 minutes . Wild-type cells treated with protonophore carbonyl cyanide m-chlorophenyl hydrazone ( CCCP ) ( 50 μM ) ( Sigma-Aldrich , USA ) , a mitochondrial membrane depolarization compound was used as a control . Subsequently , the cells were washed with 1 X PBS and fixed as mentioned in the above protocol . The samples were then analysed using FACS , BD Biosciences FACS Calibur system using BD Biosciences CellQuest software . The mean fluorescence intensities ( MFI ) of FL2 channel were used for analysis . For each sample , data for at least 20 , 000 events were collected . The cellular ATP levels of the parasites were measured using a bioluminescence-based ATP detection assay kit ( BioVision ) as per the manufacturer’s protocol . Briefly , 1 × 107 log phase WT , ΔLdSIR2RP2 and ΔLdSIR2RP2/+ cells were seeded in a 96-well plate and treated with indicated compounds ( 5 mM , 2-deoxyD-glucose ( 2DG ) ( Sigma-Aldrich , USA ) or 10 μM Oligomycin ( Oligo ) ( Sigma-Aldrich , USA ) . The cells were lysed , substrate solution was added , and the luminescence intensity was measured in a luminometer ( Turners Design , TD 20/20 ) . Percentage cellular ATP levels were plotted in reference to control . The susceptibility profile of L . donovani wild-type and mutant promastigotes for sirtinol , nicotinamide ( NAM ) , Ex-527 , and cambinol was determined using MTT [3- ( 4 , 5- dimethylthiazol-2-yl ) -2 , 5 diphenyltetrazolium bromide] assay [31] . Briefly , log-phase promastigotes ( 5 × 105 cells/well ) were seeded in a 96-well flat-bottomed plate and incubated with different drug concentrations at 22°C . DMSO was used as a vehicle control for inhibitors which were dissolved in DMSO . After 72 h of incubation , 20 μL of MTT ( Sigma-Aldrich , USA ) ( 5 mg/ml ) was added to each well , and the plates were further incubated at 37°C for 3 h . The reaction was terminated by the addition of 50 μL of stopping solution ( 50% isopropanol and 20% SDS ) followed by gentle shaking at 37°C for 30 min to 1 h . The absorbance was measured at 570 nm in a microplate reader ( SpectraMax M2 from Molecular Devices ) . The susceptibility of wild-type and mutant amastigotes to the above-mentioned inhibitors was determined by visualisation of intracellular parasite load using Giemsa staining of the infected J774A . 1 murine macrophages , 48 h after treatment with different concentrations of the drug . The cytotoxicity of inhibitors on J774A . 1 murine macrophage cell line was determined by MTT assay . Briefly , 1 x 104 cells per well were seeded in a 96-well flat-bottomed plate and incubated with different drug concentrations at 37°C , 5% CO2 . DMSO was used as a vehicle control for inhibitors which were dissolved in DMSO . After 48 h of incubation , the assay was terminated by adding MTT as mentioned above and the absorbance was measured at 570 nm . The effect of inhibitors on the ADP-ribosyltransferase activity of recombinant LdSIR2RP2 was assessed . Briefly , the assays were carried out with 2 . 5 μg of recombinant LdSIR2RP2 , 5 μg of calf thymus histones ( Sigma ) , and the respective inhibitors at indicated concentrations . The reaction mixtures were incubated at 37°C for 1 h . Thereafter , 2 . 5 μCi of [32P]NAD+ was added to the above mixture , and the reactions were allowed to proceed further at room temperature for 2 h . The reactions were terminated , resolved on 12% SDS gel and visualized as stated above . Statistical analysis was done using Graph Pad Prism Version 5 . 0 . Data shown are representative of at least three independent experiments unless otherwise stated as n values given in the legend . All the experiments were set in triplicate , and the results are expressed as the mean ± S . D . Student’s t test was employed to assess the statistical significance of differences between a pair of data sets with a p-value of < 0 . 05 considered to be significant .
The eukaryotic sirtuins are classified into four classes: I , II , III and IV , based on the conserved sequence motif patterns [7] . Humans have seven sirtuins distributed in all the four classes [32] while kinetoplastids have Class I , Class II , and Class III sirtuins ( Fig 1A ) . There are three SIR2 related proteins in L . donovani . The SIR2 related proteins of L . donovani ( LdBPK_260200 . 1 , LdBPK_231450 . 1 , and LdBPK_341900 . 1 ) are termed as LdSIR2RP1 , LdSIR2RP2 , and LdSIR2RP3 . The phylogenetic analysis of kinetoplastid sequences along with the homologs from human ( Fig 1A ) suggests a clear branching of the individual sirtuin subfamilies . LdSIR2RP1 belongs to Class I; LdSIR2RP2 belongs to Class II , and LdSIR2RP3 belongs to Class III , of sirtuins . Class I , kinetoplastid sequences are related to HsSIRT1 , HsSIRT2 , and HsSIRT3 . Class II kinetoplastid sequences are closer to HsSIRT4 . Further , it is also evident from the phylogeny that class III kinetoplastid sirtuins are closely related to HsSIRT5 sirtuin that have been shown to possess both NAD+-dependent deacetylase and the novel desuccinylase/ demalonylase activities [12] . To date , the Class I homologs of Leishmania [15 , 21] and Trypanosoma [13 , 17] parasites have been characterized , and nothing is known about the sirtuins belonging to Class III and Class II subfamilies . The LdSIR2RP1 , LdSIR2RP2 , and LdSIR2RP3 encode putative polypeptides of amino acids 373 , 320 and 243 , respectively . The predicted molecular mass of LdSIR2RP1 , LdSIR2RP2 and LdSIR2RP3 is 41 , 35 and 27 kDa , respectively . Of these , the SIR2 Leishmania homolog LdBPK_260200 . 1 was predicted to be localized in the cytosol . The other two SIR2 copies on Chromosome 34 ( LdBPK_341900 . 1 ) and Chromosome 23 ( LdBPK_231450 . 1 ) are predicted to be localized in the mitochondria . All the three Leishmania sirtuins lack the N-terminal extension present in HsSIRT1 and HsSIRT2 that are required for nucleolar localization [33] but contain a full catalytic SIR2 domain ( Fig 1B ) . These proteins also have the conserved Zn2+ binding motif ( CX2CX20CX2C ) , although one of the Cys residues is lacking in LdSIR2RP3 ( Fig 1B ) . Multiple sequence alignment of the catalytic region of Leishmania sirtuins with that of human sirtuins homologs shows the conserved sequence patterns characteristic of the individual subfamilies of sirtuins ( Fig 2 ) . GAG , TQNID and HG motifs , as well as other residues essential for enzymatic catalysis , are conserved in Leishmania sirtuins . Among the conserved motifs , the “HG” motif is of interest as mutation of “HG” to “YG” has been shown to convert the yeast sirtuin into a dominant negative gene with loss in function [7] while it is conserved as “QG” in microbial sirtuins denoted as SirTM subfamily [34] . Thus , the “HG” motif essential for sirtuin-mediated ADP-ribosylation and deacetylation is conserved in all the kinetoplastid sequences suggestive of active sirtuins . Both LdSIR2RP1 and LdSIR2RP2 contain a zinc binding motif . Overall , LdSIR2RP1 sequence shares 46% sequence identity with its human homolog HsSIRT2 , while LdSIR2RP2 and LdSIR2RP3 share 39% and 37% sequence identity with HsSIRT4 and HsSIRT5 , respectively . In order to overexpress the recombinant LdSIR2RP2 , the coding sequence of LdSIR2RP2 was cloned into a pETM41 expression vector possessing an N-terminal maltose binding protein ( MBP ) tag . The construct was transformed into Artic-Express DE3 strain and induced as explained in the Methods section , resulting in expression of MBP-tagged recombinant LdSIR2RP2 with an estimated molecular size of ~ 77 kDa ( Fig 3A ) . The size of the recombinant protein correlated with the amino acid composition of the LdSIR2RP2 protein ( ~ 35 kDa ) and MBP tag ( ~ 42 kDa ) . The recombinant MBP-LdSIR2RP2 was affinity purified on a pre-equilibrated amylose resin column yielding ∼1 mg of pure protein from 1 litre of bacterial culture . The MBP tag cleavage of the recombinant protein was done by using TEV protease at a w/w ratio of 1% the amount of fusion protein at 4°C . The pure recombinant LdSIR2RP2 was obtained after passing the reaction mixture , first through pre-equilibrated Ni2+-NTA column to remove TEV protease and then through pre-equilibrated amylose resin column to remove the MBP tag protein ( Fig 3A ) . Proteins belonging to the SIR2 family exhibit NAD+-dependent deacetylase activity due to the presence of a well-conserved enzymatic core SIR domain of ∼250 amino acids [7 , 35] . Some members of the SIR2 family are also known to catalyze the transfer of ribose 5′-phosphate from nicotinic acid mononucleotide to amino acid residues of bovine serum albumin ( BSA ) , histones or SIR2 proteins themselves [11 , 35] . Recently , class III sirtuins have been reported to have novel enzymatic activities like desuccinylase and demalonylase [12 , 36] . Using recombinant LdSIR2RP2 , we checked NAD+-dependent deacetylase and/or NAD+-dependent ADP-ribosyltransferase activity and/or desuccinylase activities . However , the recombinant protein did not show any detectable deacetylase and desuccinylase activities . Next , to assess whether LdSIR2RP2 is an ADP-ribosyltransferase , [32P]NAD+ as the donor and bovine serum albumin ( BSA ) or calf thymus histone as acceptor substrates were used . rLdSIR2RP2 protein was able to catalyze the ADP-ribosylation of calf thymus histones ( Fig 3B ) . However , there was no transfer of ADP-ribose on BSA , suggesting its strong specificity towards histones . Quantitative analysis of rLdSIR2RP2 done using histones as acceptor protein showed that a two-fold increase in the number of histones and recombinant protein did not have any effect on the extent of ADP-ribosylation . However , when the amount of NAD+ was doubled , there was ~1 . 5-fold increase in ADP-ribosylation ( Fig 3B ( ii ) ) , suggesting that NAD+ is the only limiting factor for ADP-ribosylation activity of rLdSIR2RP2 . Phylogenetically LdSIR2RP2 belongs to class II of sirtuin family , of which HsSIRT4 is a prominent member . HsSIRT4 only has NAD+-dependent ADP-ribosyltransferase activity [37] . Thus , it can be concluded that like its human counterpart , LdSIR2RP2 possessed an NAD+-dependent ADP-ribosyltransferase activity and lacked measurable NAD+-dependent deacetylase and desuccinylase activities . However , this needs to be further validated using stringent purification methods . In-silico analysis performed with PSORTII software indicated a higher probability for a mitochondrial localisation of LdSIR2RP2 . The subcellular localisation of LdSIR2RP2 was confirmed using confocal microscopy to evaluate the localisation of C-terminal GFP-tagged full-length LdSIR2RP2 . L . donovani parasites transfected with the LdSIR2RP2-GFP fusion constructs and parasites transfected with GFP vector alone ( without insert ) were fixed and analysed by fluorescence microscopy . The GFP fluorescence was visible in approximately 80% of the cells ( Fig 4A ( i ) ) . The kinetoplast and nuclear DNA in these cells were readily identified by their bright staining with DAPI . The parasites transfected with the GFP vector alone showed GFP fluorescence in the entire promastigotes ( Fig 4A ) . However , the LdSIR2RP2-GFP fusion protein was found to localise in the mitochondria as seen by the co-localization of the LdSIR2RP2−GFP with the fluorescence associated to the MitoTracker Red CMXRos that reveals the position of mitochondria within the promastigotes ( Fig 4B ) . In order to determine the essentiality and biological function ( s ) of the mitochondrial sirtuin in L . donovani , we replaced both the alleles of LdSIR2RP2 gene using classical gene replacement experiments . Two successive rounds of gene targeting with two dominant selectable markers were undertaken to inactivate the LdSIR2RP2 gene completely . This was done by the generation of inactivation cassettes having hygromycin phosphotransferase ( HYG ) or neomycin phosphotransferase ( NEO ) as selection markers along with 5′UTR and 3′UTR of LdSIR2RP2 gene , as described in the Methods . Linear replacement cassette fragments were transfected into wild-type L . donovani promastigotes leading to the generation of heterozygous parasites in which one copy of LdSIR2RP2 gene was replaced with either the hygromycin or neomycin drug resistance gene . Subsequently , another round of gene targeting was done to generate LdSIR2RP2 homozygous null mutant parasites . PCR analysis was done to confirm the recombination events ( Fig 5A and 5B ) . Genomic DNA from the WT parasites was used as a positive control ( Fig 5C ) . Bands corresponding only to the LdSIR2RP2 gene were obtained indicating the specificity of HYG and NEO primers . The genotype of the heterozygous ( LdSIR2RP2/Hyg and LdSIR2RP2/Neo ) and homozygous ( ΔLdSIR2RP2 ) null mutant parasites was further confirmed by Southern blot analysis ( Fig 5D ) . The ‘add-back’ mutant line ( ΔLdSIR2RP2/+ ) was prepared by transfecting pSP72a-zeo-a-LdSIR2RP2 episome into homozygous null mutant parasites ( ΔLdSIR2RP2 ) . PCR analysis was done to confirm the presence of the episomal plasmid ( Fig 5A and 5E ) . A band of ~ 360 bp was obtained upon amplification with ZeoF and ZeoR primers , which correspond to Sh ble gene , that confers Zeocin antibiotic resistance [38] , thus confirming the presence of episomal pSP72a-zeo-a-LdSIR2RP2 in the ‘add-back’ line ( ΔLdSIR2RP2/+ ) . The growth rate of each of the cell line was determined in order to verify phenotypic alterations in the wild-type and genetically manipulated parasites . This was done by counting promastigote cells using a hemocytometer for a period of 12 days . The absence of LdSIR2RP2 in the promastigotes led to a significant decrease in the growth rate of the parasite as compared to the WT cells ( Fig 6A ) . The doubling time of the ΔLdSIR2RP2 ( ~ 32 h ) was ~ 2 . 5 fold higher than the WT cells ( ~ 12 h ) . This restrictive growth phenotype was rescued in the promastigotes expressing episomal LdSIR2RP2 ( ΔLdSIR2RP2/+ ) with doubling time of ~ 14 h . Next , we assessed whether the genetic deficiency of LdSIR2RP2 in L . donovani has an impact on its ability to infect host cells by performing infectivity assays with stationary-phase promastigotes in the J774A . 1 murine macrophage . Microscopic observation of murine macrophages stained with Giemsa showed that , while WT parasites were capable of ~ 50% infection in murine macrophages , ΔLdSIR2RP2 parasites had reduced infectivity with only ~ 25% of the macrophages infected ( Fig 6B ) . The add-back line ( ΔLdSIR2RP2/+ ) showed infection comparable to that of WT ( Fig 6B ) . Upon comparing the parasite numbers of the WT , ΔLdSIR2RP2 , and ΔLdSIR2RP2/+ in the murine macrophages , it was observed that ΔLdSIR2RP2 mutant parasites had ~ 50% reduction in the number of amastigotes per macrophages relative to the wild-type parasites 48 h p . i ( Fig 6C ) . In the ‘add-back’ line ( ΔLdSIR2RP2/+ ) the parasite numbers were restored to the levels comparable to that of WT . Thus , these observations imply that loss of LdSIR2RP2 affects the ability of the parasite to infect and sustain robust infection within murine macrophages . This could be partially attributed to the slow growth phenotype of the null mutants compared to their wild-type counterparts . Since the ΔLdSIR2RP2 parasites exhibited a restrictive growth phenotype , the possibility of any cell cycle-related defects , which could have lowered the growth rate of mutant parasites , was examined . For this , log phase cells of WT , ΔLdSIR2RP2 , and ΔLdSIR2RP2/+ parasites were taken and examined for their DNA content . The null mutants showed an increased G2/M population of cells ( ~ 40 . 33% ) ( p = 0 . 0001 ) compared to the WT ( 23 . 68% ) and ‘add-back lines ( 17 . 79% ) ( Fig 6D ) . It is possible that the G2/M block in the null mutants may be indirectly responsible for the slow growth kinetics of ΔLdSIR2RP2 parasites . Since LdSIR2RP2 had mitochondrial localization , the effect of LdSIR2RP2 deletion on the functioning of the parasite mitochondria was investigated . Mitochondria utilise oxidation of substrates to produce membrane potential in the form of a proton gradient across the inner mitochondrial membrane . Maintenance of this potential is necessary for the generation of ATP by mitochondria [39] . Hence , the effect of LdSIR2RP2 gene deletion on the mitochondrial transmembrane potential ( ΔΨm ) was evaluated by using MitoTracker Red . MitoTracker Red is known to accumulate in energised mitochondria [40] . Relative to wild-type control , ΔLdSIR2RP2 mutants showed decreased MitoTracker red fluorescence by ~ 48 . 33% thus indicating reduced ΔΨm ( Fig 7A ) . ΔLdSIR2RP2 mutants expressing episomal LdSIR2RP2 showed ΔΨm comparable to that of the WT control . WT cells treated with 50 μM CCCP were used as positive controls . These cells showed a decrease in the mean fluorescence intensity values ( 32 . 67% of reduction ) as compared to the untreated WT cells . The alterations in the mitochondrial membrane potential are known to affect the levels of cellular ATP . Since the null mutants exhibited reduced ΔΨm , intracellular levels of ATP were determined . ΔLdSIR2RP2 cells exhibited a significant decline ( 47% ) in cellular ATP levels when compared to the wild-type controls ( Fig 7B ) . The ATP levels of rescue mutants ( ΔLdSIR2RP2/+ ) cell were restored to that of the WT cells . The present data indicates the effect of LdSIR2RP2 deletion on mitochondrial membrane potential and hence , ATP synthesis in the cell . Furthermore , we delineated whether this decline in total cellular ATP levels was due to a decrease in the mitochondrial or glycolytic ATP synthesis . For this , WT and ΔLdSIR2RP2 parasites were treated either with oligomycin , a classical inhibitor of F0-F1-ATP synthase or 2-deoxyD-glucose ( 2DG ) , a competing substrate for hexokinase . It was observed that the WT cells treated with 2DG , exhibited higher ATP levels ( 85% ) in comparison to ΔLdSIR2RP2 ( 20% ) when treated with 2DG ( Fig 7C ) . This implied that the decline in total cellular ATP pool in ΔLdSIR2RP2 was due to a reduction in the mitochondrial generated ATP . On the contrary , when the cells were treated with 10 μM of oligomycin , WT parasites had only 5% ATP levels as opposed to 22 . 5% ATP levels in ΔLdSIR2RP2 parasites ( Fig 7C ) , suggesting that the glycolytic ATP generation is higher in ΔLdSIR2RP2 parasites than the WT parasites . This is in accordance with the observation that when the ability of parasites to generate ATP through mitochondrial oxidative phosphorylation is compromised , parasites increase their glycolytic metabolism to maintain the energy supply [41] . The structural and biochemical differences between the human and the parasitic sirtuins have led to exploring sirtuins as potential anti-parasitic therapeutic targets [42] . The effect of sirtuin inhibitors: sirtinol , nicotinamide ( NAM ) , 6-chloro-2 , 3 , 4 , 9-tetrahydro-1H-carbazole-1-carboxamide ( Ex-527 ) and cambinol , on the growth of WT , ΔLdSIR2RP2 , and ΔLdSIR2RP2/+ parasites , was determined . Sirtinol is a naphthol derivative compound and is a known inhibitor of NAD+-dependent deacetylase activity of sirtuins . Treatment of WT , ΔLdSIR2RP2 , and ΔLdSIR2RP2/+ with sirtinol inhibited the growth of both the promastigote and the amastigote stage of the parasites in a concentration-dependent manner . The IC50 of sirtinol for the promastigotes of WT was not significantly different from that of ΔLdSIR2RP2 ( p = 0 . 08 ) . Similar observation was made in the case of intracellular amastigotes ( p = 0 . 09 ) ( Fig 8A , Table 3 ) . Nicotinamide ( NAM ) is a known physiological inhibitor of SIR2 deacetylase activity of HsSIRT1 and ScSIR2 [43] . Concentrations of NAM as high as 10 mM did not inhibit the growth of the promastigote stage of WT , ΔLdSIR2RP2 and ΔLdSIR2RP2/+ parasites ( Table 3 ) . This could be attributed to the presence of thick lipophosphoglycan layer around the promastigotes that could interfere with the entry of the inhibitor inside the promastigotes . However , NAM inhibited the proliferation of intracellular amastigotes of WT , ΔLdSIR2RP2 , and ΔLdSIR2RP2/+ . The IC50 value of NAM was significantly lower in the case of ΔLdSIR2RP2 than that of WT parasites . The IC50 value of NAM in the rescue mutants ( ΔLdSIR2RP2/+ ) was comparable to that of the WT parasites ( Fig 8B ) . NAM was also tested for its cytotoxicity to J774A . 1 and was found to inhibit its growth at an IC50 value of 29 . 21 ± 5 . 3 mM , which was ~3 fold higher than that observed for the intracellular amastigotes ( Table 3 ) . EX-527 is an indole-based sirtuin inhibitor . WT , ΔLdSIR2RP2 , and ΔLdSIR2RP2/+ , promastigotes were susceptible to EX-527 inhibition . The IC50 value of EX-527 for null mutants was ~2 fold lower than that of the WT parasites . The IC50 of EX-527 for ΔLdSIR2RP2/+ was comparable to that of the WT parasites ( Fig 8C , Table 3 ) . EX-527 was found to be a more effective inhibitor in the case of intracellular amastigotes of all parasitic cell lines ( Table 3 ) . The IC50 value of EX-527 for ΔLdSIR2RP2 was ~2 fold lower than that for the WT intracellular amastigotes . EX-527 inhibited J774A . 1 mouse macrophages at higher concentrations ( IC50: 157 ± 4 . 4 μM ) when compared to that observed in the case of the intracellular amastigotes . Cambinol , a potent human SIRT1 and SIRT2 inhibitor [44] , inhibited the growth of WT , ΔLdSIR2RP2 , and ΔLdSIR2RP2/+ promastigotes ( Fig 8D , Table 3 ) . The IC50 value of cambinol for ΔLdSIR2RP2 was ~4 fold lower than that of the WT promastigotes . Cambinol inhibited the growth of WT , ΔLdSIR2RP2 and ΔLdSIR2RP2/+ intracellular amastigotes at the IC50 value of 1 . 7 ± 0 . 02 μM , 0 . 8 ± 0 . 1 μM and 2 . 4 ± 0 . 2 μM , respectively , ( Fig 8D , Table 3 ) . The compound was found to be cytotoxic to the J774A . 1 mouse macrophages ( IC50: 188 ± 11 . 6 μM ) . Our data indicates increased susceptibility of null mutants to all the tested compounds except sirtinol . This would indicate the possible pleiotropic effect of these inhibitors on the other two parasitic sirtuins , LdSIR2RP1 , and LdSIR2RP3 . Earlier studies have also shown that NAM inhibits recombinant LiSIR2RP1 ( 15 ) . Furthermore , overexpression of either of these sirtuins , TcSIR2RP1 and TcSIR2RP3 in T . cruzi protected the parasite from the effect of cambinol and NAM ( 17 ) . Assay for the ADP-ribosyltransferase activity of LdSIR2RP2 was further performed in the presence of sirtinol , NAM , EX-527 and cambinol to test their specificity and investigate the ability to inhibit the activity of recombinant LdSIR2RP2 . A dose-dependent inhibition of the ADP-ribosyltransferase activity of LdSIR2RP2 was assessed . ADP-ribosylation assays were performed as mentioned in “materials and methods” section with varying concentrations of the inhibitors . Sirtinol did not result in inhibition of ribosylation of histones at concentrations as high as 40 μM ( Fig 9A ) . NAM and EX-527 inhibited LdSIR2RP2 activity at concentrations as low as 50 μM ( Fig 9A ) . Out of all the four inhibitors , cambinol was the most effective in inhibiting the activity of LdSIR2RP2 ( Fig 9B ) . Concentration as low as 2 . 5 μM of cambinol inhibited the ADP-ribosyltransferase activity of LdSIR2RP2 .
The silent information regulator 2 ( SIR2 ) -like family of NAD+ dependent protein deacetylases are highly conserved proteins from archaea to higher eukaryotes . These proteins are involved in the regulation of several functions in eukaryotic cells , including transcriptional repression , recombination , cell cycle , cellular responses to DNA-damaging agents , microtubule organisation and longevity . Sirtuins , being central to proper cellular functioning and proliferation , have been studied in protozoan parasites like Plasmodium and Trypanosomes . The parasitic sirtuins have been found to have both conserved and unique functions , which regulate a broad diversity of cellular processes , thus making them suitable drug targets for anti- parasitic therapy [16] . Three SIR2 homologs were identified in the in silico analysis of Leishmania genome; SIR2RP1 , SIR2RP2 , and SIR2RP3 . SIR2RP1 , a cytosolic sirtuin , is known to be essential for the infectivity and survival of the parasite and hence an attractive drug target for antileishmanial chemotherapy [21] . The other two mitochondrial sirtuins of Leishmania have not yet been characterised . In the present study , we describe the functional role of LdSIR2RP2 in L . donovani . Phylogenetic and sequence analysis reveals that LdSIR2RP2 is closer to the human homolog HsSIRT4 , which belongs to class II of the sirtuin family . In the present study , we demonstrate that LdSIR2RP2 like HsSIRT4 has only NAD+-dependent ADP-ribosyltransferase activity . The protein was found to localise in the mitochondria of the parasite , similar to HsSIRT4 . Although LdSIR2RP2 was not found to be essential for the survival of the parasite , the null mutants exhibited delayed growth rate and attenuated infectivity . Cell cycle analysis of the null mutant parasites revealed a G2/M block which could be a possible reason for the growth defects observed in the mutant lines . Since , LdSIR2RP2 is a mitochondrial protein , analysis of the mitochondrial parameters revealed compromised mitochondria with lowered ΔΨm and hence , lesser mitochondrial ATP content of the cell . These phenotypic alterations in the ΔLdSIR2RP2 parasites were relieved by ectopic expression of LdSIR2RP2 in ΔLdSIR2RP2/+ . Thus , deletion of the mitochondrial sirtuin LdSIR2RP2 in Leishmania affects mitochondrial functioning leading to lowered ATP content of the cells and hence delayed growth kinetics . HsSIRT4 is a mitochondrial ADP-ribosyltransferase that inhibits mitochondrial glutamate dehydrogenase 1 activity [37] . Overexpression of SIRT4 in mammalian cells causes an increase in mitochondrial respiration , glycolysis , and glucose oxidation , but with no change in growth rate or in steady-state ATP concentrations [45] . Mitochondria , the energy provider of the cell , depends on the universal coenzyme ( NAD+ ) or its phosphorylated counterpart NADP , to maintain homoeostasis within the cell [46] . In addition to participation in redox reactions , NAD+ acts as a versatile cellular signalling molecule through the generation of ADP-ribose ( ADPR ) [47] . The mitochondrial sirtuins SIRT3 , SIRT4 , and SIRT5 utilise mitochondrial NAD+ pool to regulate the activity of their targets implicated in the regulation of both glycolysis and cellular oxidative stress [46 , 48] via deacetylation and mono- ADP-ribosylation . Mono-ADP-ribosylation of proteins is a phylogenetically ancient , reversible , and covalent posttranslational modification of proteins . Both mono- and poly-ADP-ribosylation of nuclear and cytosolic proteins are known to regulate various physiological processes , such as mitosis , cellular differentiation , and proliferation , telomere dynamics , and ageing programmed necrosis and apoptosis via signalling , chromatin modification and remodelling of chromatin structure [49] . Apart from this , both , poly and mono-ADP-ribosylation modification of mitochondrial proteins is reported to have an effect on the metabolism of this organelle as well [47] . Recently , the role of mitochondrial sirtuins in parasites has been reported . It was observed that the overexpression of mitochondrial SIR2RP3 in T . cruzi led to an increase in parasite proliferation , movement , and differentiation . This was due to increased deacetylation of mitochondrial targets within the parasite [17] . Similarly , in our study , we observed a slow growth pattern and reduced infectivity upon deletion of the mitochondrial sirtuin , LdSIR2RP2 . The null mutants also exhibited compromised mitochondrial functioning and delayed growth kinetics . In kinetoplasts , actively respiring mitochondria are required for survival of both , the promastigotes and amastigotes [50 , 51] . Thus , it can be speculated that deletion of LdSIR2RP2 might have affected the activity of some of the main mitochondrial proteins which could be regulated by ADP-ribosylation , thereby affecting mitochondrial functioning in the null mutants . However , further studies are necessary to identify the specific substrates targeted by LdSIR2RP2 , in order to gain insight into the role of this mitochondrial sirtuin in parasite biology . Sirtuins are known to be involved in regulation of vital cellular processes . Hence , they have been proposed as promising targets for the development of anti-parasitic drugs [42 , 52 , 53] . Here , the efficacy of known sirtuin inhibitors; sirtinol , nicotinamide , Ex-527 , and cambinol , on the growth of the WT and genetically manipulated parasites was determined . Except for sirtinol , all the other three inhibitors were more effective in inhibiting the growth of ΔLdSIR2RP2 parasites than the WT . This increased susceptibility of the inhibitors was relieved by ectopic expression of LdSIR2RP2 . The concentrations tested had no significant effect on the host cells , indicating a selectivity of these inhibitors for the parasitic sirtuins than for the host sirtuins . While these inhibitors show specificity towards the ADP-ribosyltransferase activity of LdSIR2RP2 , we cannot rule out the off-target effect of these inhibitors on the other two sirtuins . Earlier studies have demonstrated that NAM inhibits recombinant LiSIR2RP1 which is both a deacetylase and ADP-ribosyltransferase ( 15 ) . Furthermore , overexpression of either of these sirtuins , TcSIR2RP1 and TcSIR2RP3 protected the parasite from the effect of cambinol and NAM ( 17 ) . With increasing drug resistance , toxicity and the cost of the available chemotherapeutic agents for the treatment of Leishmaniasis , the development of new leishmanicidal drugs and the search for new targets is required . The peculiar differences between the parasite and mammalian mitochondria , as well as unique characteristics of parasite mitochondria , makes mitochondrial proteins as good drug targets . Several new studies involving inhibitors like; Benzophenone-derived bisphosphonium salts [54] , artemisinin [55] , chalcones , including licochalcone A [56] , Tafenoquine [57] , luteolin and quercetin [58]; indicate the essential role of mitochondrial biology in the survival of the parasite . Here , we have attempted to characterise a mitochondrial sirtuin , which is involved in maintaining the mitochondrial homoeostasis . LdSIR2RP2 deletion resulted in reduced growth and virulence of the parasite . Known sirtuin inhibitors were able to inhibit the growth of the parasite . The inhibitors also showed inhibitory effect on the enzymatic activity of recombinant LdSIR2RP2 . However , the pleiotropic effect of these inhibitors on the other two parasitic sirtuins , LdSIR2RP1 , and LdSIR2RP3 , cannot be ruled out . Thus , developing a specific inhibitor to target LdSIR2RP2 alone or in combination with the available chemotherapeutic agents could provide a better rationale for the treatment of Leishmaniasis .
|
Sirtuins are present in most organisms , including plants , bacteria , and animals . They play a vital role in promoting an organism’s health and survival . These proteins are involved in the regulation of several functions in eukaryotic cells , including transcriptional repression , recombination , cell cycle , cellular responses to DNA-damaging agents , and longevity . Sirtuins are known to be involved in regulation of vital cellular processes . Hence , they have been proposed as promising targets for the development of antiparasitic drugs . Leishmania donovani , a protozoan parasite that causes visceral leishmaniasis is known to express three sirtuins; SIR2RP1 , SIR2RP2 , and SIR2RP3 . We have worked on the functional characterization of the SIR2RP2 protein from L . donovani in this study . We report that the SIR2RP2 is an NAD+-dependent ADP-ribosyltransferase . This protein is present in the mitochondrion of the promastigotes and deletion of both copies of the gene caused reduced growth , compromised mitochondrial functioning and cell cycle arrest in the transgenic parasites . The transgenic parasites also had reduced infectivity . Deletion of LdSIR2RP2 resulted in increased sensitivity of the parasites to the known sirtuin inhibitors . Furthermore , the sirtuin inhibitors were found to inhibit the ADP-ribosyltransferase activity of LdSIR2RP2 thus , indicating that parasitic sirtuins can be exploited as drug targets for antileishmanial chemotherapy .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"parasitic",
"cell",
"cycles",
"split-decomposition",
"method",
"microbiology",
"parasitic",
"diseases",
"protozoan",
"life",
"cycles",
"parasitic",
"protozoans",
"parasitology",
"multiple",
"alignment",
"calculation",
"developmental",
"biology",
"protozoans",
"leishmania",
"mitochondria",
"sequence",
"motif",
"analysis",
"bioenergetics",
"promastigotes",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"bioinformatics",
"proteins",
"life",
"cycles",
"recombinant",
"proteins",
"leishmania",
"donovani",
"biochemistry",
"cell",
"biology",
"computational",
"techniques",
"database",
"and",
"informatics",
"methods",
"biology",
"and",
"life",
"sciences",
"protozoology",
"energy-producing",
"organelles",
"organisms",
"parasitic",
"life",
"cycles"
] |
2017
|
The mitochondrial SIR2 related protein 2 (SIR2RP2) impacts Leishmania donovani growth and infectivity
|
Cotrimoxazole prevents opportunistic infections including falciparum malaria in HIV-infected individuals but there are concerns of cross-resistance to other antifolate drugs such as sulphadoxine-pyrimethamine ( SP ) . In this study , we investigated the prevalence of antifolate-resistance mutations in Plasmodium falciparum that are associated with SP resistance in HIV-infected individuals on antiretroviral treatment randomized to discontinue ( STOP-CTX ) , or continue ( CTX ) cotrimoxazole in Western Kenya . Samples were obtained from an unblinded , non-inferiority randomized controlled trial where participants were recruited on a rolling basis for the first six months of the study , then followed-up for 12 months with samples collected at enrollment , quarterly , and during sick visits . Plasmodium DNA was extracted from blood specimens . Initial screening to determine the presence of Plasmodium spp . was performed by quantitative reverse transcriptase real-time PCR , followed by genotyping for the presence of SP-resistance associated mutations by Sanger sequencing . The prevalence of mutant haplotypes associated with SP-resistant parasites in pfdhfr ( 51I/59R/108N ) and pfdhps ( 437G/540E ) genes were significantly higher ( P = 0 . 0006 and P = 0 . 027 , respectively ) in STOP-CTX compared to CTX arm . The prevalence of quintuple haplotype ( 51I/59R/108N/437G/540E ) was 51 . 8% in STOP-CTX vs . 6 . 3% ( P = 0 . 0007 ) in CTX arm . There was a steady increase in mutant haplotypes in both genes in STOP-CTX arm overtime through the study period , reaching statistical significance ( P < 0 . 0001 ) . The frequencies of mutations in pfdhfr and pfdhps genes were higher in STOP-CTX arm compared to CTX arm , suggesting cotrimoxazole effectively controls and selects against SP-resistant parasites . ClinicalTrials . gov NCT01425073
Despite the changes in the epidemiology and improvement in the control of HIV-infection and malaria , both remain important infectious diseases and global health priorities . Through immunosuppression , HIV infection affects the acquisition and persistence of immune response to malaria , causing substantial increase in the malaria prevalence and malaria-related morbidity and mortality [1] . Antiretroviral therapy ( ART ) and cotrimoxazole , a fixed-dose trimethoprim-sulfamethoxazole ( an antifolate ) widely used to prevent opportunistic infections in HIV-infected individuals , including falciparum malaria significantly reduces mortality and morbidity in HIV-infected individual . In countries with adequate health infrastructure , the World Health Organization ( WHO ) recommends daily cotrimoxazole prophylaxis for HIV-infected individuals with low CD4 cell count levels ( < 350 cells/mm3 ) , whereas in countries with high prevalence of HIV and limited health infrastructure , cotrimoxazole prophylaxis is recommended for all HIV-infected individuals regardless of the CD4 cell count levels [2] . However , there are concerns that widespread use of cotrimoxazole prophylaxis may result in selection of Plasmodium falciparum parasites with cross-resistance to closely related antifolate antimalarials such as sulphadoxine-pyrimethamine ( SP ) [1] . Although artemisinin based combination therapy is the mainstay for treatment of uncomplicated malaria in most malaria endemic countries , SP is widely used as intermittent preventive treatment of malaria in pregnancy ( IPTp ) and in infants ( IPTi ) in sub-Saharan Africa ( SSA ) [3–5] . Some of the important mutant alleles that confer P . falciparum parasite resistance to SP are in P . falciparum dihydrofolate reductase ( pfdhfr ) gene at codons 51 , 59 and 108 , and P . falciparum dihydropteroate synthase ( pfdhps ) gene at codons 437 and 540 . Recent studies have shown high prevalence of these mutant alleles and haplotypes in Western Kenya , including mutant allele at codon 164 in the pfdhfr gene which is associated with high-grade resistance to SP [6–9] . Despite the high prevalence of SP-resistant mutations in parasite population in Western Kenya , there is limited clinical evidence associating these mutations with compromised efficacy of cotrimoxazole prophylaxis and IPTp/i [10] . Recent studies have indicated there is fixation of some of the key SP-resistant mutations in the parasite population despite discontinuation of SP as the first-line treatment for more than a decade [6 , 7 , 9] . Based on the malaria risk map and the eco-epidemiology of malaria , Kenya is stratified into four malaria ecological regions [11] , with the lake region in Western Kenya having the highest , stable transmission of malaria with an estimated prevalence of 27% based on microscopy [12 , 13] and 37% based on PCR [14] . The HIV-1 prevalence in Kenya is estimated at 5 . 9% , with Homa Bay County , one of the eight counties in the lake region of Western Kenya having the highest prevalence estimated at 26% [15] . With such high HIV and malaria prevalence , the selective pressure due to cotrimoxazole prophylaxis and the risk of developing antifolate resistance in P . falciparum warrants further investigation . From February 2012 to August 2013 , we conducted a randomized controlled trial ( RCT ) among adults on ART with evidence of immune recovery to determine whether discontinuation of cotrimoxazole was non-inferior to continuation of cotrimoxazole prophylaxis in decreasing morbidity in Homa Bay County [16] . Study participants were recruited in the first six months of the study on a rolling basis and randomized to discontinue or continue cotrimoxazole , then followed-up for 12 months with the primary endpoint a composite of malaria , pneumonia , diarrhea and non-trauma mortality events . Samples were collected at enrollment , quarterly , and at sick visits which the participants were encouraged to visit the clinic to see study providers for any illness . Malaria was defined as a fever , measured or self-reported , and either a positive rapid diagnostic test or thick smear showing the presence of parasites . Patients who were diagnosed with malaria were treated following the Kenyan Ministry of Health national guidelines . In the RCT study , we found increased incidence of malaria ( 13 . 0 in discontinuation of cotrimoxazole arm [STOP-CTX] vs . 0 . 4 in continuation of cotrimoxazole arm [CTX] per 100 person-years ) [16] . In a follow-up study which we characterized the risk associated with stopping CTX therapy by determining parasite density , multiplicity of infecting parasites , and rates of new cases of parasitemia by PCR , malaria incidence was 42 . 0 in STOP-CTX vs . 9 . 9 in CTX per 100 person-years [17] . In this study , we determined and compared the prevalence of P . falciparum parasites with mutations associated with SP-resistance in HIV-infected individuals in the two study population arms , STOP-CTX and CTX .
The study protocol was approved by the ethical review committee of the Kenya Medical Research Institute and the institutional review boards of the University of Washington and the Walter Reed Army Institute of Research . All participants gave informed consent . Consent was written if literate and fingerprint if illiterate , with the signature of an independent witness . For the clinical study , Vestergaard Frandsen donated insecticide-treated bednets and water filters . Alere donated cartridges for the Pima machines used for CD4 count measurements . Samples used in this study were collected between February 2012 and September 2013 in an unblinded , two-arm randomized non-inferiority clinical trial ( clinical trials registration NCT01425073 ) . The details of the study and sample collection have been described elsewhere [16] . Briefly , a total of 500 participants ≥18 years old , HIV seropositive , and taking first-line ART and cotrimoxazole with evidence of immune recovery ( ART for ≥18 months and CD4 count > 350 cells/mm3 ) were enrolled in the study , and randomized to discontinue with cotrimoxazole prophylaxis ( STOP-CTX; 250 individuals ) or continue ( CTX; 250 individuals ) . The study took place in Homa Bay County , Western Kenya , a malaria holoendemic lake endemic region where transmission is intense through-out the year with high annual entomological inoculation rates [12] . Generally , a bimodal pattern of rainfall is observed with the long rainy season from March to June and the short rainy season from November to December , but the periods vary each year with malaria prevalence peaking 1–2 months after the rainy season . Annual rainfall ranges from 700 mm to 1 , 200 mm with mean temperature of 25°C , with relatively high humidity [14] . This study lasted 18 months , enrollment taking place during the first six months ( 01 February 2012 to 27 August 2012 ) with participants randomized to STOP-CTX or CTX . This strategy ensured that participants were enrolled and followed over different malaria seasons . Blood samples were collected from the participants during the scheduled visits at months 0 , 3 , 6 , 9 and 12 ( M0 , M3 , M6 , M9 and M12 respectively ) , and sick visits . Participants were encouraged to come to the clinic to see study providers for any illness as a sick visit . At each sick visit , a standardized questionnaire was provided to assess participants’ symptoms and a clinician performed a physical exam . Additionally , available and clinically relevant basic diagnostic tests were performed ( e . g . malaria smear , chest radiograph , stool ova and parasite exam ) to assist with diagnosis as per routine clinic practice . Additionally , pertinent microbiological samples were taken in order to better evaluate cause of illness . If further evaluation was necessary , patients were referred for hospitalization at the nearest facility . Clinical and laboratory records from any hospitalization during participation were reviewed . Participants with malaria were treated following Kenyan national guidelines . In the RCT , dried blood spots samples were collected from the participants at enrollment , every 3 months and during sick visits ( whether or not they were diagnosed with malaria ) for the duration of the study which was 12 months . DNA was extracted from the FTA filter papers using the QIAamp DNA mini kit ( Qiagen , Valencia , CA ) . The detection of P . falciparum positive samples was performed by quantitative reverse transcriptase real-time PCR ( qRT-PCR ) as previously described [17 , 18] . The presence of mutations in dihydrofolate reductase ( pfdhfr: codons 16 , 50 , 51 , 59 , 108 , and 164 ) and dihydropteroate synthase ( pfdhps: codons 436 , 437 , 540 , 581 , and 613 ) genes which are associated with antifolate resistance in P . falciparum samples were assessed by Sanger sequencing as previously described [6] . Briefly , after successfully amplifying the target regions , the PCR amplicons were purified using Exosap-it ( Affymetrix , Santa Clara , CA ) per the manufacturer’s protocol . Sequencing of the target regions was done on the ABI 3500 xL genetic analyzer using version 3 . 1 of the big dye terminator method ( Applied Biosystems , Foster City , CA ) . Bioinformatics analysis of the sequence data was done on the CLC Main Work Bench v6 software ( Qiagen , Redwood City , California , USA ) . All sequences were compared against the pfdhfr ( Accession Number; XM_001351443 ) or pfdhps ( Accession Number; XM_001349382 ) 3D7 reference sequence published at the NCBI database . The different Plasmodium species and genotype polymorphisms within pfdhfr and pfdhps genes of P . falciparum were analyzed as proportions showing frequency rates . The differences in frequencies were determined by the Chi-square test . All statistical analyses were performed at the 5% significance level . Graph pad Prism 4 . 0 software ( Graph pad Software , San Diego , California , USA ) was used for the analyses .
A total of 2 , 625 samples were initially screened for presence of malaria parasites by qRT-PCR [17 , 18] . Of these , 183 samples were positive for Plasmodium genus , 131 ( 71 . 6% ) were P . falciparum , and 101 ( 55 . 2% ) were successfully genotyped in pfdhfr ( at codons 16 , 50 , 51 , 59 , 108 , and 164 ) and pfdhps ( at codons 436 , 437 , 540 , 581 , and 613 ) genes; 30 samples had non-falciparum parasites . The difference in the number of samples that were successfully genotyped by sequencing ( n = 101 ) and those that were Plasmodium spp . positive as detected by qRT-PCR ( n = 183 ) is due to the difference in sensitivities of the amplification assays used . For the detection of Plasmodium genus used for the initial screening for the presence of the parasite , Ottichilo et al . ( 2016 ) used previously described qRT-PCR assay ( probe based assay ) which amplifies total nucleic acids ( RNA and DNA ) of the 18S rRNA genes , increasing sensitivity several fold [17 , 18] . For genotypic analysis , nested PCRs that target DNA only [6] were used in the sequencing reactions . Table 1 shows the prevalence of the mutant alleles in pfdhfr and pfdhps genes . The prevalence was based on the total number of samples that were P . falciparum positive in each of the arms ( total n = 101; STOP-CTX = 85 and CTX = 16 ) . Single nucleotide polymorphisms ( SNPs ) were designated as wild , mutant or mixed alleles [6 , 7] . None of the parasite samples contained mutations in pfdhfr codons 16 and 50 or in pfdhps codons 436 and 613 . Three samples , two in STOP-CTX arm and one in the CTX arm had the pfdhfr 164L mutation which confers high grade resistance to antifolate [19 , 20] . The two samples in the STOP-CTX arm were collected in M3 and M12 whereas the one sample in the CTX arm was collected at enrollment . Mutations in the STOP-CTX arm were present at a higher frequency compared to CTX arm in pfdhfr codons 51 ( 65 . 9% [n = 56/85] vs . 25 . 0% [n = 4/16]; P = 0 . 0043 ) , 59 ( 60% [n = 51/85] vs . 12 . 5% [n = 2/16]; P = 0 . 0007 ) and 108 ( 65 . 9% [n = 56/85] vs . 31 . 3% [n = 5/16]; P = 0 . 0126 ) . In the pfdhps gene , mutations were only present in codons 437 and 540 with frequencies higher in the STOP-CTX arm compared to the CTX arm ( 68 . 2% [n = 58/85] vs . 37 . 5% [n = 6/16] and 70 . 6% [n = 60/85] vs . 37 . 5% [n = 6/16] respectively ) . Fig 1 shows the prevalence of the different mutation haplotypes in pfdhfr and/or pfdhps genes . The prevalence of the triple mutant haplotype ( pfdhfr 51I/59R/108N ) was 52 . 9% ( n = 45/85 ) in the STOP-CTX arm versus 6 . 3% ( n = 1/16; P = 0 . 0006 ) in the CTX arm , and the pfdhps double mutant ( 437G/540E ) was 57 . 6% ( n = 49/85 ) in the STOP-CTX arm versus 25 . 0% ( n = 4/16; P = 0 . 027 ) in the CTX arm . The prevalence of quintuple haplotype ( 51I/59R/108N/437G/540E ) was 51 . 8% ( n = 44/85 ) in the STOP-CTX arm versus 6 . 3% ( n = 1/16; P = 0 . 0007 ) in the CTX arm . To determine change in the prevalence of the mutations over the study period for each study arm , we analyzed samples carrying mutations at each time-point , starting M0 –M12 , and the sick visits . However , the sample sizes were small in the CTX arm . In the STOP-CTX arm , the percent prevalence of point mutations in both genes increased over time with marked increase occurring in M9 followed by a slight drop in M12 ( Table 2 ) . The difference in prevalence of mutations was less pronounced in the sick visits between the two arms . Fig 2 shows the prevalence of mutations at the different time-points of the triple mutant haplotype ( 51I/59R/108N ) in the pfdhfr gene , the double mutant haplotype ( 437G/540E ) in pfdhps gene and the quintuple haplotype ( 51I/59R/108N/437G/540E ) in the STOP-CTX arm . All the changes ( increases ) over time reached statistical significance ( P = 0 . 0069 , 95% CI = 19 . 99–67 . 45; P = 0 . 008 , 95% CI = 32 . 90–61 . 02; and P = 0 . 0044 , 95% CI 18 . 35–52 . 09 , respectively ) .
The clinical benefits of using cotrimoxazole prophylaxis in controlling opportunistic infections including falciparum malaria in HIV-infected individuals are clear [21] . The widespread use of cotrimoxazole prophylaxis for individuals with HIV infection in malaria endemic countries has been a concern because of the risk of developing cross resistance to other antifolate drugs [22–24] . However , concern for cross resistance was based on in vitro data [3 , 25 , 26] , and has not been substantiated by field clinical data ( reviewed by [27] ) . Despite the high prevalence of SP-resistant mutations , cotrimoxazole continues to provide important benefits in reducing morbidity and mortality particularly in the setting of HIV infection [28–30] , and does not lead to increased resistance [22 , 31–33] . Current field data clearly supports the continued use of cotrimoxazole as a prophylactic drug in HIV-infected populations [21 , 27] . Additional studies to investigate the use of cotrimoxazole as an alternative to SP in IPTp/i , malaria treatment and prophylaxis , and as a combined anti-malarial therapy with artemisinin are warranted . In this study , we demonstrated cotrimoxazole prophylaxis did not results in increased risk of developing resistance , corroborating previous studies [22 , 28–32 , 34] . Interestingly , we found the prevalence of SP-resistant alleles increased steadily over the study period in the STOP-CTX arm for the first 9 months . Further , although the sample size was small , the prevalence of SP-resistant alleles in the CTX arm did not change over the study period . Taken together , cotrimoxazole prophylaxis lowered the overall incidence of SP-resistant parasites , consistent with previous studies [22 , 33] . Key SP-resistant mutations have become fixed in parasite populations despite discontinuation of SP as the first-line treatment for more than a decade [6 , 7 , 9] , indicating these mutations might be providing benefit to the parasite population without fitness cost . It is possible that cotrimoxazole selects against parasites carrying SP-resistant alleles in the population , and removal of cotrimoxazole pressure allows the SP-resistant parasite population , which seems to be more fit than SP-susceptible population to dominate . As studies are underway to investigate expanded role of cotrimoxazole in developing countries [21] , the use of cotrimoxazole in the prevention of malaria HIV-infected and HIV uninfected populations , especially as a tool for malaria elimination and as travelers’ prophylactic drug , needs to be further investigated . Mutation in pfdhfr 164L codon is associated with high grade resistance to SP [35–38] . While some studies have shown evidence that cotrimoxazole prophylaxis might be associated with presence of pfdhfr 164L codon [31 , 39] , other studies do not support this observation [40] . In Kenya , only low prevalence of pfdhfr 164L has been reported [6 , 7 , 36 , 41] . In our study , three parasite isolates had pfdhfr 164L , one in CTX arm collected at M0 and two in STOP-CTX collected in M9 and M12 , indicating the presence of this mutation is unlikely due to cotrimoxazole drug pressure . Although cotrimoxazole has been speculated to contribute to antifolate selective pressure [40] , additional studies are required to support this observation . This study had several limitations . First , the study was unblinded clinical trial without a placebo or concurrent control group of HIV-uninfected individuals [16] . Second , the number of infections in the CTX arm was small , limiting statistical analysis or might have resulted in bias . This made it especially difficult in interpreting the data when evaluating prevalence at each time-point . Also , the use of highly sensitive qRT-PCR assay for the initial screening followed by multiple sequencing reactions , which uses multiple PCR steps that are not as sensitive was a limitation . In conclusion , this study demonstrated cotrimoxazole does not select for SP-resistance P . falciparum parasites but instead , lowers the overall incidence of SP-resistant parasites . Since cotrimoxazole is available in malaria-HIV co-endemic regions with infrastructure in place and is effective against SP-resistant parasites , additional studies are required to validate these findings and to further explore the possibility of expanding the use of this drug for IPTp/i , prophylaxis for non-HIV population and travelers .
|
Cotrimoxazole , an antifolate , is a fixed-dose trimethoprim-sulfamethoxazole used to prevent opportunistic infections including malaria in HIV-infected individuals . There are concerns that widespread use of cotrimoxazole for prophylaxis may result in selection of P . falciparum parasites with cross-resistance to other antifolate drugs such as sulphadoxine-pyrimethamine ( SP ) , which is used as intermittent preventive treatment of malaria in pregnancy ( IPTp ) and in infants ( IPTi ) in Africa . This sub-study used samples from a clinical trial in which HIV-infected individuals on antiretroviral treatment were randomized to discontinue ( STOP-CTX ) or continue ( CTX ) cotrimoxazole prophylaxis for 12 months . The sub-study was designed to assess whether taking cotrimoxazole increased the risk of selecting for parasites with SP-resistant mutations in HIV-infected individuals . Samples were genotyped by sequencing to assess the prevalence of mutations associated with SP-resistance . We found there was no risk of selecting for parasites with SP-resistance mutations while on cotrimoxazole . In fact , the opposite was true; cotrimoxazole controlled parasites carrying SP-resistance mutations as evident by the gradual increase in the prevalence of parasites with mutant alleles in the STOP-CTX arm and not in the CTX-arm . We concluded that cotrimoxazole remains effective in controlling malaria infection despite of the high prevalence of SP-resistant parasites , and its use does not select for SP mutations .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"parasite",
"groups",
"plasmodium",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"genetic",
"mapping",
"parasitology",
"preventive",
"medicine",
"mutation",
"apicomplexa",
"protozoans",
"public",
"and",
"occupational",
"health",
"malarial",
"parasites",
"haplotypes",
"point",
"mutation",
"eukaryota",
"prophylaxis",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"malaria",
"organisms"
] |
2019
|
The prevalence and antifolate drug resistance profiles of Plasmodium falciparum in study participants randomized to discontinue or continue cotrimoxazole prophylaxis
|
Bovine tuberculosis ( bTB ) caused by Mycobacterium bovis is an important re-emerging disease affecting livestock , wildlife and humans . Epidemiological studies are crucial to identifying the source of bTB infection , and its transmission dynamics and host preference , and thus to the implementation of effective strategies to contain it . In this study , we typed M . bovis isolates from livestock , and investigated their genetic diversity and distribution . A total of 204 M . bovis isolates were collected from cattle ( n = 164 ) and Sicilian black pigs ( n = 40 ) reared in a limited area of the province of Messina , northeastern Sicily , an area that had previously been identified as having the highest incidence of bTB in livestock on the island . All M . bovis isolates were typed by both spoligotyping and 12-loci MIRU-VNTR analysis . Results from both methods were then combined in order to improve the discriminatory power of M . bovis typing . We identified 73 combined genetic profiles . Thirty-five point six percent of the profiles were common to at least two animals , whereas 64 . 4% of profiles occurred in only one animal . A number of genetic profiles were predominant in either cattle or black pigs . We identified common genetic patterns in M . bovis isolates originating not only from neighboring districts , but also from non-neighboring districts . Our findings suggest that bTB is widespread in our setting , and is caused by a large number of genetically diverse M . bovis strains . The ecology and farming practices characteristic of the area may explain the substantial M . bovis heterogeneity observed , and could represent obstacles to bTB eradication .
Bovine tuberculosis ( bTB ) is an infectious disease of worldwide distribution caused mainly by Mycobacterium bovis , one of the members of the Mycobacterium tuberculosis complex ( MTC ) . M . bovis has the broadest host range of any member of the MTC , and an intricate epidemiological pattern of infection . The pathogen may infect a wide range of domestic animal species , with negative impacts on both animal productivity and the international trade of animal products . M . bovis can spread via aerosols , suckling , and the sharing of water and feed [1 , 2] . M . bovis may also spread to wildlife species which can act as reservoir hosts , contributing to the transmission and persistence of the disease [3 , 4] . Finally , it may infect humans , causing a disease–zoonotic tuberculosis—which is indistinguishable from that caused by M . tuberculosis [5 , 6] . Direct contact with infected animals and the consumption of unpasteurized dairy products have been indicated as the most likely routes of zoonotic transmission [7] . In most industrialized countries , animal test-and-slaughter schemes have successfully reduced the occurrence of bTB , and only occasional cases of M . bovis infection in humans are documented [8] . M . bovis represents a serious public health issue in non-industrialized countries , however , where factors such as absent or inadequate bTB control programs , immunodeficiency , close contact with infected animals , consumption of infected animal products and malnutrition , contribute to increase the risk of zoonotic tuberculosis [8 , 9] . The design of intervention strategies in animals is informed mainly by the epidemiology of the disease . Comprehensive epidemiological studies of bTB can provide valuable insights into the sources of infection , routes of transmission , geographical localization , host preference , disease dynamics and risk factors for the maintenance and spread of the disease , thus contributing to contain the disease in animals and reduce the risk to humans [2] . Various molecular-based techniques have been developed to study the epidemiology of M . bovis infections [2] . Spoligotyping [10] and mycobacterial interspersed repetitive unit-variable number tandem repeats ( MIRU-VNTR ) typing [11 , 12] are the most commonly used methods for M . bovis genotyping . The use of spoligotyping in combination with MIRU-VNTR typing has been shown to improve the discriminatory power of M . bovis typing [13–15] . Combined genetic profiles have recently been used to analyze the transmission of M . bovis in France [16] , Midwest Brazil [17] , Cameroon [18] and Mozambique [19] . In Italy , a national eradication program has resulted in a gradual reduction of bTB prevalence in most regions . On the island of Sicily , bTB remains a major concern , however . Our group has previously investigated the role played by the Sicilian black pig—an autochthonous free- or semi-free-ranging breed of domestic pig—in the maintenance of bTB in two neighboring areas in northeastern Sicily: the Nebrodi and Madonie Natural Parks [20] . The characteristics of the lesions , their localization and the genetic profiles of M . bovis isolates indicated that Sicilian black pigs may act as a reservoir of bTB in the ecological setting studied [20] . More recently , our group has genotyped MTC isolates from livestock and wild animals throughout Sicily [21] . We found evidence of considerable diversity of M . bovis spoligotypes and MIRU-VNTR profiles in domestic animals , with the province of Messina recording the highest number of outbreaks in the period 2004–2014 in livestock in Sicily [21] . The latest Epidemiological Veterinary Bulletin of the province of Messina , published in December of 2018 ( https://docs . google . com/viewer ? a=v&pid=sites&srcid=aXpzc2ljaWxpYS5pdHxpenN8Z3g6NzIzNjhmOTYwYWJjMDEzMw ) , reported bTB prevalence ( 1 . 75% ) and incidence ( 1 . 47% ) data for the province , in terms of the percentage of farms affected . During recent bTB control campaigns in the province of Messina , Caronia has emerged as one of the districts with the highest incidence rates of bTB in livestock . In-depth epidemiological studies aimed at investigating genetic relationships among M . bovis isolates in animals in this high-risk area , especially in the district of Caronia , may contribute to more effective bTB eradication interventions . In this study , we typed M . bovis isolates from livestock—cattle and Sicilian black pigs—reared in the province of Messina in Sicily , by both spoligotyping and MIRU-VNTR typing , for epidemiological purposes .
Tissue samples from slaughtered cattle and Sicilian black pigs revealing tubercular lesions on postmortem examination were collected at abattoirs . Inspections at abattoirs were carried out in accordance with Italian law , and no permission from abattoir owners was needed . Over a period of two years , from 2015 to 2016 , tissue samples from slaughtered cattle and Sicilian black pigs revealing tuberculous-like lesions on postmortem examination were collected at the only three abattoirs allowed to slaughter bTB-infected animals in the province of Messina: Pascoli dei Nebrodi soc . coop 1503 M Mirto , Caruso Impex 1344 M Barcellona Pozzo di Gotto , and Si . L . Car . 940 M CEE Merì . Slaughtered animals included both those for consumption purposes and those with positive response to bovine tuberculin within the frame of bTB control and eradication programs . Tissue samples were cultured and isolates were typed as previously described [20] . Of all isolates identified as M . bovis over the two years ( N = 765 ) studied , a total of 204 M . bovis isolates were randomly selected for analysis from slaughtered cattle ( n = 164 ) and Sicilian black pigs ( n = 40 ) . The province of Messina ( 3 , 247 km2 ) covers approximately 11% of the total area of the island . This largely mountainous province , divided into 108 districts , is home to the Nebrodi Park , a rural nature reserve extending over an area of nearly 860 km2 . This mostly mountainous park is covered by wide pastures and woods that give shelter to numerous species of wild mammals , birds , reptiles , and invertebrates . Wild boar are absent from the Nebrodi Park . The Sicilian black pig lives mostly in the woods of the Nebrodi Park , where it is reared in free or semi-free roaming conditions , frequently sharing pastures with cattle . Approximately 1 , 800 cattle and 1 , 200 black pig herds are present in the province of Messina . Sample data concerning the name of the farmer , farm location ( district of the Messina province ) and main type of livestock reared on the farm , were recorded for every slaughtered animal with a culture-confirmed diagnosis of bTB . All M . bovis isolates were typed by both spoligotyping [10] and 12-loci MIRU-VNTR analysis . In spoligotyping , the spacer sequences contained in the direct repeat locus were detected by hybridization onto a spoligotyping membrane ( Ocimum Biosolutions , Hyderabad , India ) . The spoligotypes obtained were checked against an international spoligodatabase ( Mbovis . org Mycobacterium bovis Spoligotype Database https://www . mbovis . org/ ) . For MIRU-VNTR typing , 12 genomic loci were selected , according to Boniotti and colleagues [13] , and amplified individually: VNTR loci 2165 , 2461 , 0577 , 0580 and 3192 ( i . e . , ETR-A to–E ) [22] , VNTR locus 2996 ( i . e . , MIRU26 ) [23] , VNTR loci 2163a , 2163b , 3155 and 4052 [24] , and VNTR loci 1895 and 3232 [25] . M . tuberculosis H37Rv was used as reference strain . Allele assignment was performed on the basis of PCR fragment size as compared to a 50-bp molecular weight marker . The resulting genetic profiles , obtained by combining spoligotypes and MIRU-VNTR results , were used for the epidemiological investigation . Minimum spanning tree analysis based on 12-locus MIRU-VNTR results was used to infer relationships between the isolates sharing the same spoligotype . The number of isolates assigned to each combined genetic profile ( spoligotype and MIRU-VNTR type ) , the host ( cow or Sicilian black pig ) and the Messina district of pathogen isolation , are provided in the graphs . Data analysis was performed using BioNumerics Seven platform ( Applied Maths , Sint-Martens-Latem , Belgium ) . Applied Maths has granted us a temporary BioNumerics evaluation license and authorized the publication of the results .
A total of 204 M . bovis isolates from slaughtered cattle ( n = 164 ) and Sicilian black pigs ( n = 40 ) with bTB were anaysed . Bovine tuberculosis cases originated from a total of 114 small-scale herds in 21 out of the province's 108 districts ( Fig 1 ) . Of the infected farms , 96 were cattle farms , 16 were black pig farms and two farms bred both cattle and black pigs . In 76 of 114 small-scale farms ( 66 . 5% ) , a single isolate was found; in 18 farms ( 16% ) , two isolates , were identified , with two different genotypes in most of these herds ( 14/18 ) ; and in 20 farms ( 17 . 5% ) , more than two isolates were found , with at least two different genotypes in most herds ( 12/20 ) . All M . bovis isolates were genotyped by both spoligotyping and 12-loci MIRU-VNTR analysis . Results are shown in Table 1 . A total of 15 spoligotypes were found . All 15 genotypes were present in M . bovis cattle isolates , with SB0120 , SB0134 and SB0841 being the most prevalent profiles ( 85 , 32 and 21 isolates , respectively ) . M . bovis pig isolates , on the other hand , belonged to only two of the 15 genotypes–SB0841 and SB0120 and ( 23 and 17 isolates , respectively ) . Overall , SB0120 , SB0841 and SB0134 were the most frequent spoligotypes circulating in livestock , accounting for 50% ( 102/204 ) , 21 . 6% ( 44/204 ) and 15 . 7% ( 32/204 ) of M . bovis isolates , respectively . MIRU-VNTR analysis yielded a total of 62 MIRU-VNTR types: 50 types in cattle only , three types in pigs only , and nine types common to both animal sources . The profiles MIRU-VNTR-a ( 22 isolates ) , MIRU-VNTR-b ( 15 isolates ) and MIRU-VNTR-s ( 11 isolates ) were the most frequently identified in cattle . MIRU-VNTR-m ( 12 isolates ) and MIRU-VNTR-n ( 8 isolates ) were predominant in pigs . Overall , MIRU-VNTR-a , MIRU-VNTR-b , MIRU-VNTR-m and MIRU-VNTR-n were the most frequent MIRU-VNTR types found in livestock , accounting for 13 . 2% ( 27/204 ) , 7 . 8% ( 16/204 ) , 7 . 8% ( 16/204 ) and 6 . 9% ( 14/204 ) of M . bovis isolates , respectively . The combination of both typing methods increased the number of genetic profiles to 73 ( S1 Table ) . Specifically , 60 profiles affected cattle only , four profiles affected black pigs only and nine profiles affected both populations . In addition , while most profiles ( 47/73 , 64 . 4% ) affected a single animal , the remaining profiles ( 26/73 , 35 . 6% ) were common to at least two bTB-infected animals . The combined genetic profiles obtained by subtyping the predominant spoligotypes SB0120 , SB0841 and SB0134 with 12-loci MIRU-VNTR analysis , were used for the epidemiological investigation . Results , together with information concerning the animal host and the district of isolation , are shown in Figs 2–4 . For a complete list of identified profiles , see supporting information ( S1 Fig , S2 Fig and S3 Fig ) . Thirty-three genetic profiles were found by subtyping the SB0120 spoligotype with MIRU-VNTR ( Fig 2 ) . Common profiles identified in at least two M . bovis isolates were marked ( SB0120/a-l ) . The most frequent combined genetic profiles were SB0120/a ( 27 isolates: 22 from cows and 5 from pigs ) , SB0120/b ( 16 isolates: 15 from cows and 1 from a pig ) , SB0120/c ( 7 isolates: 5 from cows and 2 from pigs ) and SB0120/d ( 7 isolates: all from bovine hosts ) , accounting for 13 . 2% ( 27/204 ) , 7 . 8% ( 16/204 ) , 3 . 4% ( 7/204 ) and 3 . 4% ( 7/204 ) of M . bovis isolates , respectively ( Fig 2 , panel A ) . SB0120/a occurred in seven districts , SB0120/b in three districts , SB0120/c in two districts , and SB0120/d in four districts ( Fig 2 , panel B ) . Sixteen genetic profiles were obtained by combining the SB0841 spoligotype and MIRU-VNTR results ( Fig 3 ) . Profiles found in two animals or more were marked ( SB0841/m-r ) . The most frequent combined profiles were SB0841/m ( 15 isolates: 4 from cows and 11 from pigs ) and SB0841/n ( 11 isolates: 3 from cows and 8 from pigs ) , accounting for 7 . 3% ( 15/204 ) and 5 . 4% ( 11/204 ) of M . bovis isolates , respectively ( Fig 3 , panel A ) . Profiles SB0841/m and SB0841/n occurred in four districts , and five districts , respectively ( Fig 3 , panel B ) . Ten genetic profiles were obtained by combining the SB0134 spoligotype and MIRU-VNTR results ( Fig 4 ) . Profiles found in two animals or more were similarly marked ( SB0134/s-v ) . Unlike SB0120/MIRU-VNTR and SB0841/MIRU-VNTR types , SB0134/s-v profiles were found only in cattle . The most frequent combined genetic profiles were SB0134/s ( 11 isolates ) and SB0134/t ( 9 isolates ) , accounting for 5 . 4% ( 11/204 ) and 4 . 4% ( 9/204 ) of isolates , respectively ( Fig 4 , panel A ) . Profiles SB0134/s and SB0134/t occurred in two districts and one district , respectively ( Fig 4 , panel B ) . The distribution and frequency of the SB0120/a-l , SB0841/m-r and SB0131/s-v combined profiles across the province of Messina are shown in Fig 1 . The fact that the highest number of bTB cases , in both cattle and pigs , was recorded in Caronia ( 116/204; 56 . 7% of M . bovis isolates ) , was probably due to the fact that the highest number of slaughtered animals were sampled from this district . Identical genetic profiles were detected in livestock originating from neighboring districts ( Fig 1 ) such as those of Caronia , Mistretta and Cesarò ( SB0120/a ) , Caronia and Cesarò ( SB0120/f; SB0841/m ) , Caronia and San Fratello ( SB0120/b; SB0120/c ) , Mirto and Naso ( SB0841/n—in pigs only ) , Galati Mamertino and Tortorici ( SB0841/p—in cows only ) , and Caronia and Acquedolci ( SB0134/u—in cows only ) . Interestingly , non-neighboring districts—some of which are quite distant from one another—exhibited identical profiles as well ( Fig 1 ) . For example , SB0120/a was detected in the districts of Caronia , Castelmola , Tortorici , Mirto , Militello Rosmarino and Cesarò , SB0841/m in the districts of Caronia , Mirto and Messina , and SB0134/v in the districts of Venetico and Mistretta . Different source animals were also found to carry identical profiles . Profiles SB0120/a , SB0120/b , SB0120/e , SB0120/j , SB0841/m , AB0841/n and SB0841/o were detected in both cattle and black pigs; profiles SB0120/d , SB0120/i , SB0134/s and SB0134/v were instead recorded in cattle only . As mentioned above , profiles affecting a single animal were also detected: SB0120/g , SB0120/h , SB0120/k , SB0841/q and SB0841/t in cattle , SB0120/l and SB0841/r in pigs .
The present study aimed at investigating the genetic diversity and distribution of M . bovis in livestock ( cattle and black pigs ) in an area , the province of Messina , identified as having the highest bTB incidence in Sicily . We collected and typed 204 M . bovis isolates from slaughtered animals with tuberculous-like lesions . We then combined spoligotyping and 12-loci MIRU-VNTR analysis results in order to increase the discriminatory power of M . bovis typing . Studies have shown that combined genotyping , such as spoligotyping with MIRU-VNTR , affords greater discriminatory power of M . bovis typing than the use of either method alone [13–15] . Combined genetic profiles have recently been used to study the molecular epidemiology of M . bovis in vast areas in France [16] , Midwest Brazil [17] , Cameroon [18] and Mozambique [19] . The present study is a detailed examination of the genetic diversity of M . bovis in a small , high-risk area . We combined spoligotypes and MIRU-VNTR typing , and used the resulting genetic profiles to perform a detailed assessment of M . bovis epidemiology in a small geographical area . Our results showed high genetic diversity among M . bovis isolates from across the province of Messina . Numerous genetic profiles were common to cattle and/or black pigs from both neighboring and distant districts , suggesting intense intra- and inter-species M . bovis transmission in the area under study . SB0120—the most common M . bovis spoligotype worldwide [26] , and one that has been shown by our group to be the most prevalent in Sicily as a whole [20 , 21]—was confirmed here as the predominant spoligotype in the province of Messina as well . While 15 spoligotypes were identified in M . bovis cattle isolates , only two—SB0120 and SB0841—were revealed in M . bovis black pig isolates . Based on the social behavior and eating habits ( hunting and rooting ) of pigs in general , and on the fact that free- or semi-free-range farming is the norm for Sicilian black pigs , one might have expected to find as large a variety of spoligotypes among pigs as that observed in cattle , or more . Instead , we characterized a much smaller number of spoligotypes in pigs than in cattle . These findings are in agreement with our previous studies , where a large variety of spoligotypes were identified in cattle [21] , while in Sicilian black pigs , the spoligotypes found were almost exclusively SB0120 and SB0841 [20 , 21] . All the above findings seem to indicate that cattle and black pigs in our setting vary in their susceptibility to infection with specific M . bovis spoligotypes , although a larger sample of black pigs should be studied to confirm or refute this hypothesis . We identified 73 different combined genetic profiles in the province of Messina . These profiles were distributed across 114 small-scale herds and 21 different districts . Most of the M . bovis isolates included in our study were from Caronia . This is consistent with the high bTB incidence rate found in the course of recent bTB control campaigns in Sicily , which resulted in large-scale slaughtering . The greater M . bovis variability identified in Caronia , as compared to other districts , is probably due to the fact that the number of bTB-infected animals sampled from this district was the highest . Of the profiles identified , 35 . 6% were common to at least two animals , whereas 64 . 4% occurred in only one animal . Some genetic profiles were predominant in either the cattle or the black pig population ( e . g . SB0120/a , SB0120/b and SB0134/s in cattle and SB0841/m and SB0841/n in black pigs ) . We discovered that animals sharing M . bovis genetic profiles could originate not only from the same farm , but also from farms located in neighboring and even non-neighbouing districts . Taken together , the above findings suggest that bTB is widespread in our setting and is caused by a large number of genetically diverse M . bovis strains distributed across the province . Compared to our previous study showing the distribution of different spoligotypes throughout Sicily over a period of 10 years [21] , here we have provided a close-up look at the genotype distribution of combined genetic profiles in the province of Messina , hoping to contribute to the improvement of bTB surveillance and control programs in this high risk area . Circumstances known to favor bTB transmission include the presence of undiagnosed infected animals , fence-line contact with other herds and animals in close living quarters [27–29] . In addition , the geography and ecology of the province of Messina , as well as local farming practices may contribute to explain the spread of genetically diverse M . bovis strains across the province . The area is largely mountainous , and numerous small-scale farms are scattered throughout the province . Several of the districts studied are situated in the Nebrodi Park ( Acquedolci , Alcara Li Fusi , Caronia , Cesarò , Galati Mamertino , Militello Rosmarino , Mistretta , Sant’Agata Militello , San Fratello and Tortorici ) . Cattle and sheep are also bred in the park , and frequently share pastures with black pigs . Moreover , herds of cattle and sheep are moved between summer and winter pastures , a traditional animal husbandry practice known as transhumance . These farming methods may be expected to increase the probability of both inter-species and intra-species contagion , even between distant farms . Bovine tuberculosis-affected animals may shed mycobacteria in the ecosystem , contaminate food , water and pastures , and thus transmit the disease to both livestock and wildlife . Direct and/or indirect contact between infected and uninfected animals ( livestock-livestock , livestock-wildlife , wildlife-livestock ) may occur at shared pastures , waterholes and feeding sites , especially in the dry season . Local livestock trade and the exchange of stud animals are also common in Sicily , and may facilitate the spread of infection to other farms and districts . In conclusion , bTB is a major concern in Sicily , especially in the province of Messina . We found a large number of genetically different M . bovis strains causing the disease in livestock bred in a small geographical area . Numerous common genetic patterns were identified in farm animals from neighboring and even relatively distant districts . Factors involving the local environment , ecology and farming management practices may explain the high level of M . bovis heterogeneity we observed in both cattle and black pigs , and are likely to represent obstacles to bTB eradication . This information may contribute to bTB prevention efforts through increased controls at farms , especially during local livestock trade and exchange of stud animals , or increased vigilance on the part of veterinarians and farmers to those farming practices suspected to favor bTB transmission . Further studies will be needed to ascertain possible relationships between these farming practices and the spread of bTB in livestock , and to identify risk factors for infection , which would ultimately inform the implementation of targeted surveillance and control measures , especially in the ecosystem-wildlife-livestock-human interface areas .
|
Bovine tuberculosis is a widespread infectious disease affecting both domestic and wild animals , as well as humans . In addition to being of public health concern , the disease , caused mainly by Mycobacterium bovis , has a significant economic impact on the farming industry due to the costs of eradication efforts . In Sicily , the largest of the Italian islands , bovine tuberculosis in livestock is of great concern , and targeted control strategies are needed . Molecular epidemiology is an essential tool for determining the distribution of a disease , so as to control it and minimize its threat to the population . We typed M . bovis isolates isolated from cattle and pigs reared in a limited area of Sicily . An in-depth comparison of the genetic makeup of these isolates allowed us a better understanding of the genetic diversity and distribution of the pathogen in our population of animals . We found that the disease is widespread in the area and caused by a large variety of M . bovis strains , which are in several cases common to different species of livestock . The paper concludes with a discussion of the findings in light of the environmental and ecological setting , and of farming practices in the area . The results are expected to contribute to the improvement of surveillance and control programs of bovine tuberculosis in the region .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"biogeography",
"livestock",
"medicine",
"and",
"health",
"sciences",
"ecology",
"and",
"environmental",
"sciences",
"ruminants",
"population",
"genetics",
"vertebrates",
"animals",
"mammals",
"farms",
"population",
"biology",
"bacteria",
"veterinary",
"science",
"genetic",
"epidemiology",
"swine",
"veterinary",
"diseases",
"geography",
"epidemiology",
"actinobacteria",
"phylogeography",
"agriculture",
"eukaryota",
"earth",
"sciences",
"genetics",
"biology",
"and",
"life",
"sciences",
"cattle",
"evolutionary",
"biology",
"amniotes",
"bovines",
"organisms",
"mycobacterium",
"bovis"
] |
2019
|
Genotype diversity and distribution of Mycobacterium bovis from livestock in a small, high-risk area in northeastern Sicily, Italy
|
Rabies virus ( RABV ) is a neurotropic virus that depends on long distance axonal transport in order to reach the central nervous system ( CNS ) . The strategy RABV uses to hijack the cellular transport machinery is still not clear . It is thought that RABV interacts with membrane receptors in order to internalize and exploit the endosomal trafficking pathway , yet this has never been demonstrated directly . The p75 Nerve Growth Factor ( NGF ) receptor ( p75NTR ) binds RABV Glycoprotein ( RABV-G ) with high affinity . However , as p75NTR is not essential for RABV infection , the specific role of this interaction remains in question . Here we used live cell imaging to track RABV entry at nerve terminals and studied its retrograde transport along the axon with and without the p75NTR receptor . First , we found that NGF , an endogenous p75NTR ligand , and RABV , are localized in corresponding domains along nerve tips . RABV and NGF were internalized at similar time frames , suggesting comparable entry machineries . Next , we demonstrated that RABV could internalize together with p75NTR . Characterizing RABV retrograde movement along the axon , we showed the virus is transported in acidic compartments , mostly with p75NTR . Interestingly , RABV is transported faster than NGF , suggesting that RABV not only hijacks the transport machinery but can also manipulate it . Co-transport of RABV and NGF identified two modes of transport , slow and fast , that may represent a differential control of the trafficking machinery by RABV . Finally , we determined that p75NTR-dependent transport of RABV is faster and more directed than p75NTR-independent RABV transport . This fast route to the neuronal cell body is characterized by both an increase in instantaneous velocities and fewer , shorter stops en route . Hence , RABV may employ p75NTR-dependent transport as a fast mechanism to facilitate movement to the CNS .
Rabies virus ( RABV ) is a neurotropic negative-strand RNA virus of the Lyssavirus genus , belonging to the Rhabdoviridae family . It is transmitted mostly via bites of diseased animals and causes a fatal infection of the nervous system in both animals and humans . A key step in RABV pathogenesis is rapid transfer to the Central Nervous System ( CNS ) through the Peripheral Nervous System ( PNS ) [1] . Due to its extraordinary properties in directed axonal transport and trans-synaptic spread , RABV has also been used as a neuro-tracing agent to map neuronal circuitry [2]–[5] . Thus , understanding the mechanism of RABV transport is of high significance for both basic and applicative fields . RABV enters the peripheral nervous system and undergoes long-distance transport arriving at the cell soma and subsequently the CNS [6] . As peripheral neurons are highly polarized cells with long axons , active intracellular transport is vital to the maintenance of neuronal function and survival [7] , [8] . Axonal transport is the cellular process of trafficking proteins , organelles , vesicles , RNA and other cellular factors to and from the neuronal cell body . The molecular motor kinesin drives transport from the cell body anterogradely , supplying proteins , lipids and other essential materials to the cell periphery . Dynein/dynactin complexes drive retrograde transport , moving damaged proteins for degradation and critical signaling molecules such as neurotrophins to the cell body [9] , [10] . Although RABV phosphoprotein P , a component of the viral nucleocapsid of infecting virions , was shown to directly interact with a light chain of the dynein motor complex [11] , [12] , axonal RABV transport and CNS infection are independent of that interaction [13] and long distance transport of complete enveloped virions within internalized endosomes is more likely [14] . However , the cellular and molecular mechanisms involved in RABV's infection and retrograde trafficking are yet to be understood . Entry of RABV into the cell requires binding of the viral glycoprotein ( G ) and fusion of the virus envelope with the host cell membrane [15] . Following receptor binding and fusion , RABV may enter the host cell through the endosomal transport pathway . In neurons , infected cells may mistake RABV particles for cargo and thus recruit trafficking components , allowing viral particles to undergo long-range axonal transport to the neuronal cell body , as was found in the case of adenovirus and the CAR receptor [16] , [17] . Direct evidence for this notion is still lacking for RABV , as well as the identity and role of the molecular determinants of the axonal transport machinery RABV utilizes . Both the Neuronal Cell Adhesion Molecule ( NCAM ) and the p75 neurotrophin receptor ( p75NTR ) have been identified as RABV glycoprotein G binding receptors [18] , [19] . Other membrane-associated components have also been implicated in RABV binding [20] . By binding one of its receptors , RABV could enter the cell and activate downstream signaling which would allow it to hijack and manipulate axonal transport machineries . Although p75NTR is known to be involved in the retrograde transport of neurotrophic factors , little is known regarding its direct contribution to viral transport . It was recently shown , however , that lentiviral vectors pseudotyped with RABV-G are retrogradely transported in motor neurons and co-localize with both p75NTR and NCAM [21] . The p75NTR contains four cysteine-rich domains ( CRD ) in the N-terminal ectodomain and a type II death domain in its cytoplasmic C-terminal segment . Rabies virus glycoprotein specifically interacts with high affinity with the first Cysteine-Rich Domains ( CRDI ) of p75NTR [22] . Neurotrophins , on the other hand , bind to the second and third p75NTR cysteine-rich domains ( CRDII&III ) [23] . Hence , RABV and neurotrophins do not compete for each other's binding site . However , it was previously reported that treatment of cells with NGF and Neurotrophin-3 , ligands of p75NTR , modulates RABV infection of DRG-originated neurons [24] . Remarkably , although p75NTR binds RABV with high affinity , it is not essential for its infection [25] , further raising questions regarding the specific role of this interaction . Here we study the strategies used by RABV to exploit axonal transport mechanisms during CNS invasion . We tracked RABV entry at nerve terminals and studied its retrograde transport along the axon in comparison to the transport of NGF . We show that RABV and NGF are internalized in similar time frames at similar domains along nerve tips and that RABV enters the cell along with p75NTR , suggesting common entry machineries . Then , by tracking the transport of GFP labeled Rabies virions along the axon , we showed it moving in acidic compartments , mostly with neurotrophic factor receptors , yet faster than NGF . Finally , we determined that p75NTR-dependent transport of RABV is faster and more directed than p75NTR-independent transport . Our model suggests that RABV may enter the cell by receptor-mediated endocytosis following its binding to p75NTR , after which it enhances the efficiency of the retrograde co-transport of RABV – p75NTR complexes . The interaction with p75NTR modulates the cellular transport machinery and serves as a mechanism to facilitate movement of RABV to the CNS .
In order to study the mechanism of RABV long distance transport , we used an optimized compartmentalized microfluidic culture chamber . In this system , murine E12 . 5–13 . 5 DRG explants were plated in one side of the chamber , referred to here as the proximal channel ( Fig . 1A , B ) . Explants are encouraged to extend axons to the distal axon channel through microgrooves by introduction of a gradient of NGF known to promote DRG axonal growth ( Fig . 1D ) . By maintaining a difference in media volume we induce directional flow across grooves or channels . Thus , fluorescent dyes like Sulforhodamine B , introduced into the distal channel that contains less medium , are prevented from reaching the grooves or proximal channel where cell bodies are located ( Fig . 1C ) . Hence , EGFP-RABV added to the axon terminus , binds exclusively to the distal axon , enabling retrograde tracking of the virus along axons in the groove . Following serum and trophic factor starvation , EGFP-RABV virions were introduced into the distal channel and 1–2 hours later were observed to move retrogradely towards the cell body , as seen by time lapse imaging ( Fig . 2A , C and Movie S1 ) . X-Y coordinates of particles moving over time were manually registered and compiled into tracks . 87 . 29% of RABV particles ( n = 244 ) , were visible over at least 10 consecutive frames and had an average instantaneous velocity >0 . 1 µm/sec . These were considered directed particles and characterized further ( Fig . 2E–J ) . Since RABV-G is known to bind the p75NTR neurotrophin receptor , we asked whether RABV exploits NGF's endogenous transport machinery in order to facilitate its own transport to the cell body , and eventually to the CNS . To address this question , we applied Quantum-Dot conjugated NGF to axon tips in the distal side of the chamber after starvation , and tracked its retrograde transport along grooves ( Fig . 2B , D and Movie S2 ) . Detailed transport analysis of RABV and NGF puncta along axons , demonstrates that RABV moves significantly faster . The average speed of RABV was roughly 40% higher than that of NGF ( 0 . 93±0 . 03 versus 0 . 66±0 . 02 µm/sec , respectively ) ( Fig . 2E ) . While both RABV and NGF particles presented a “stop and go” motion ( Fig . 2C , D ) RABV particles demonstrated a more processive movement with fewer stops ( Fig . 2F ) . Though spot duration did not differ significantly ( Fig . 2F ) . RABV particles spent a larger percentage of their traffic time in direct movement ( Fig . 2H ) . Moreover , RABV's instantaneous velocity distribution profile was shifted towards the higher velocities ( Fig . 2I ) , averaging at 0 . 785±0 . 008 µm/sec , while the average for NGF was 0 . 546±0 . 007 µm/sec ( n = 7318 and 5452 , respectively ) . Hence , RABV moves faster and travels longer distances ( Fig . 2J ) . These findings suggest that RABV not only exploits the axonal mechanism for neurotrophin transport , but might also increase transport efficiency . Explants grown in microfluidic chambers tend to vary in the axonal meshwork formed at the distal channel , and consequently in the number of viral particles found in each groove . We therefore checked whether the number of tracked particles per groove had an effect on measured values . We found no correlation between number of EGFP-RABV puncta tracked per axon and that of percentage of directed puncta , their speed , displacement or run length ( Fig . S1 ) . In order to illustrate the differences leading to faster transport of RABV compared to NGF , we proceeded to inquire whether internalization of these ligands occurs over similar time frames . To this end , we performed a series of live imaging experiments using TIRF microscopy , and tracked fluorescent RABV or NGF particles at the axonal growth cone . Distinct features of the TIRF evanescent wave allow us to limit our view to the basal surface , an ideal set up for viewing internalization processes occurring at axon tips . Fluorescently labeled RABV or NGF was applied to DRG explant cultures and their dynamics at the axon tips were tracked ( Fig . 3 ) . Both RABV and NGF particles demonstrated similar internalization profiles . They arrived at the axon extremity , anchored to the cell membrane and then travelled for a few seconds before eventually internalizing into the cell . Internalization was manifested as a gradual decrease in particle intensity until complete disappearance ( see methods for more details ) , ( Fig . 3A , B and Movies S3 , S4 ) . Quantification of internalization durations revealed similar kinetics for RABV and NGF ( Fig . 3C , D ) ( 9 . 66±1 . 97 seconds and 13 . 25±1 . 58 seconds , respectively , p = 0 . 148 ) . RABV-G was shown to bind the p75NTR with high affinity [22] , yet there is no direct evidence demonstrating that p75NTR can act as a receptor to mediate RABV internalization . We therefore conducted a second series of live TIRF imaging assays where EGFP-RABV was applied to DRG explant cultures along with a fluorescent antibody against the extracellular domain of the p75 receptor . Dual color TIRF live imaging revealed that RABV and p75NTR were dynamically co-localized at the cell membrane , and were internalized together at the axon tips ( Fig . 4A , C and Movie S5 ) . Interestingly , tracking dual-color particles indicated that these followed a directed path towards the center of the axonal growth cone prior to their internalization ( Fig . 4B ) . Thus , RABV binds and is internalized at the axon tip together with p75NTR in a manner similar to that of NGF . In order to further validate the intimate proximity of p75 and RABV on the cell surface , we used single particle localization algorithms , to determine the relative position of co-localized RABV and p75 spots from live TIRF images at subpixel resolution using Gaussian and radial symmetry fitting ( Fig . 4D–F ) . Distances between p75 and RABV were measured according to the center positions of the radial symmetry fits , using Parthasarathy's Radial Center algorithm [26] , and averaged on 85 . 5±20 . 82 nm ( n = 7 ) . This measurement may reflect the actual distance between the rabies virion , roughly 100×200 nm in size , and the p75NTR it binds . Other factors , such as the size of the labeling antibody and imperfect optical alignment may have also contributed to the measured distance . Co-localization was further confirmed by stimulated emission depletion ( STED ) microscopy , on DRG explants were treated with either a combination of RABV-EGFP and anti-p75-550 , or RABV-mCherry and anti-p75-488 ( Fig . S2 ) . Our findings redirect attention to the question of p75NTR's role in RABV infection , in light of previous studies which have shown that p75NTR is not essential for RABV infection but afects clinical manifestation [27] . To address this question , we measured RABV infection rates in p75NTR-knocked down DRG cultures . Mixed infection with 4 shRNA constructs against p75NTR decreases p75NTR levels in DRGs ( Fig . S3 ) . Short inoculation times of 30 and 120 minutes were chosen , to study p75NTRs role in promotion of infection . Lower infection rates were observed in cultures infected with shRNA as opposed to GFP infected controls ( Fig . 4G ) , at both time points . Under these conditions , p75NTR expression enhances RABV infection of embryonic sensory neurons . Similar observations were made upon p75NTR knockdown in the NSC34 motor neuron cell line ( data not shown ) . As both RABV and NGF traffic retrogradely to the cell body , and use a similar internalization process , we asked whether RABV hijacks the NGF endosomal retrograde transport machinery . To address this issue we performed dual color live imaging of RABV and NGF retrograde axonal transport in microfluidic chambers . Tracking RABV and NGF , we analyzed ∼50 events in which RABV and NGF were retrogradely transported together in the same compartment along the axon ( Fig . 5A–F and Movie S6 ) . Examining the distribution of average track speeds we found that the co-transported RABV/NGF particles could be divided into two populations ( Fig . 5G ) . Separate characterization of these two populations shows that faster tracks were less prone to pausing mid-way ( 0 . 3±0 . 2 versus 1 . 7±0 . 2 pauses per 100 seconds , respectively ) . As only 2 pauses were recorded in the fast group , no significant difference was found between either group's stop durations ( Fig . 5H , I ) . Overall , the fast group , thus spent less time paused ( Fig . 5J ) . The presence of two distinct populations of RABV/NGF co-transport may suggest to a switch in “drivers” , where NGF leads the slower group while the faster is led by RABV . We proceeded to further characterize the transport mechanism of RABV particles; seeking to first determine the cellular compartment in which RABV particles are transported , post infection , within DRG axons . Using microfluidic chambers ( Fig . 1 ) , we infected axons in the distal compartments , while simultaneously treating cells with fluorescent markers of cellular compartments ( Fig . 6 ) . In order to check that RABV is transported in acidic compartments [28] , possibly late-endosomes , lysosomes or autophagosomes , and to quantify this co-localization , we treated cells with the PH-indicator dye , Lysotracker Red . In order to observe whether RABV is transported with mitochondria , we treated cells with the Mitotracker Deep Red marker . Using both markers together with EGFP-RABV , we acquired three channel time- lapse image series of RABV transport ( Fig . 6A–D ) . Detection and analysis of co-localized fluorescent spots along the axon determined that most of the RABV particles ( 75 . 9±4 . 09% , n = 3 separate experiments ) were located in acidic compartments ( Fig . 6E ) . Examination of merged kymographs of RABV and Lysotracker red amplified the outcome of co-localization analysis , as most transient RABV tracks were matched with a Lysotracker red track ( Fig . 6G–I ) . Thus , RABV particles are located in acidic axonal endosomes that retrogradely move towards the cell body . In contrast , we could not detect any significant co-localization of RABV and mitochondrial marker ( Fig . 6F ) . We therefore conclude that while RABV is mostly transported in an acidic compartment , it is not transported along with mitochondria in DRG axons . Having determined that RABV is internalized with p75NTR and can undergo transport with NGF , we then tested whether it is also transported with more selective neurotrophin receptors . In order to study this , we infected DRG axons grown in microfluidic chambers with EGFP-RABV , while labeling these cells with fluorescent antibodies against neurotrophin receptors . Specifically , we used fluorescent-tagged antibodies against the general neurotrophin receptor p75NTR and the specific NGF receptor TrkA and acquired three-channel time-lapse movies of RABV and these receptors ( Fig . 7A–D ) . Co-localization analysis determined that the over 60% of RABV particles co-localized with p75NTR , yet less than 40% with TrkA ( Fig . 7E , F , H and Movie S7 ) . Interestingly , when co-localized with neurotrophic factor receptors , mainly p75NTR , RABV tracks showed a greater processivity than that of RABV-only tracks ( Fig . 7G , H ) , suggesting that mutual transport of RABV with NTF receptors induces a more progressive transport . Although p75NTR may serve as a receptor for RABV ( Fig . 4 and Movie S5 ) and the viral G-protein binds with high affinity to the p75NTR , the receptor is not an absolute requirement for RABV infection , as RABV can also infect p75NTR deficient cells [25] . Furthermore , RABV could be transported retrogradely along the axon without p75NTR ( Fig . 7 ) . To assess the precise contribution of p75NTR to RABV transport , we tracked the transport of RABV particles along the axon , with and without p75NTR ( Fig . 8 A–C and Movie S8 ) . Plotted kymographs of RABV and p75NTR demonstrate that motile RABV tracks tend to co-localize with those of p75NTR ( Fig . 8D–F ) . Separate characterization of each group's transport ( Fig . 8G–O ) demonstrated that when traveling with p75NTR , RABV particles traveled faster compared to particles negative for the receptor , with respective speeds of 0 . 86±0 . 04 versus 0 . 63±0 . 04 µm/sec ( Fig . 8G ) . We attribute this alteration in speed to the fact that RABV-p75 puncta are less prone to pausing on their route to the cell body , with an average of 0 . 9±0 . 1 vs . 2 . 3±0 . 2 pauses per 100 seconds for the p75NTR positive and negative groups , respectively . Moreover , p75 positive particles paused for shorter times ( Fig . 8I ) and overall spent less time paused during their travel ( Fig . 8J ) . Another factor contributing to their higher respective speed was their instantaneous velocities . The distribution of instantaneous velocities of RABV particles positive for p75NTR is shifted towards the higher velocities when compared to RABV negative for p75NTR . RABV ( + ) p75NTR average velocity was higher than that of RABV ( − ) p75NTR , 0 . 71±0 . 008 versus 0 . 49±0 . 007 µm/sec ( n = 5494 and 3919 , respectively ) . Another interesting finding was that less than 10% of the recorded velocity events in the p75NTR positive group were anterograde , while over 17% of events in the p75NTR negative groups were anterograde , i . e . moving “backwards” towards their entrance area , the axon tip . This implies that p75NTR not only has a role in assisting the transport of NTF and viruses , but might also regulate the transport machinery , contributing to vesicle directionality towards the cell body . Measuring both the area and average intensities of the RABV particles in each group , we found that the p75NTR positive RABV particles were larger in size , ( average area of 1 . 34±0 . 09 µm2 vs . 0 . 81±0 . 07 µm2 , p<0 . 0005 ) , and had stronger intensity of GFP signal , when normalized to the average intensity of RABV particles in each experiment ( 1 . 24±0 . 11 vs . 0 . 61±0 . 1 , p<0 . 001 ) ( Fig . 8K–L ) . These larger and more prominent particles possibly represent larger endosomes , containing several RABV particles and receptors . These particles , positive for p75NTR , cover a greater net distance per time unit and are more directed towards the cell body , as shown when we compared their trajectories and mean squared displacements ( Fig . 8N and O , respectively ) . Taken together , the major differences that were observed between the two RABV groups suggest distinguishable transport mechanisms . It therefore seems that RABV binding to p75NTR allows the virus to exploit a rapid transport mechanism to facilitate its trafficking to the cell body . We continued to examine the role of p75 in RABV transport by tracking RABV in axons of a DRG explant after p75 knock-down . DRG explants were grown in microfluidic chambers and infected with LV-shRNA-p75-EGFP . mCherry-RABV was added to the distal channel as described before , and after 1 hour of incubation imaged for 1–2 hours . Although many axons crossed the grooves to the distal channel , only few were found to express GFP , hence were infected with sh-p75 . Unlike in non-infected axons , where RABV was easily identified when trafficked towards the cell body , RABV transport in sh-p75 axons was less frequent ( Fig . S4A–C ) . The number of transported RABV particles was reduced in sh-p75 axons when compared to adjacent , non-infected axons or to LV-EGFP infected controls ( Fig . S4D ) or to LV-EGFP infected controls ( not shown ) . The few RABV particles in sh-p75 axons were less directed than those transported in non-infected cells , as seen by their respective trajectories ( Fig . S4E ) .
A crucial initiating event for the mechanism outlined above is the binding of RABV to p75NTR . Here we provide direct evidence that p75NTR may serve as a receptor for internalization at axons tips , as well as mediate incorporation into the endosomal neurotrophic transport pathway . However , RABV does not strictly rely on p75NTR for internalization and may enter the cell in a p75NTR independent pathway , while is also known to bind other receptors [20] , [30] . Hence there are likely to be additional ways for RABV to merge into the p75NTR-RABV endosome . Indeed , we observed events where RABV particles merge with p75NTR-positive endosomes en route ( Movie S9 ) . The p75NTR neurotrophin receptor accelerates RABV transport to the cell body , yet there are instances of fast , processive transport of RABV particles without the receptor . We assume other identified RABV receptors such as NCAM [30] or other , un-identified ones , may facilitate RABV's retrograde axonal transport within endosomes in a similar fashion . Our experiment with DRG cultures where p75 was knocked down , show reduced RABV infection and transport ( Fig . 4 and Fig . S4 ) , and further support the role for p75NTR proposed here . Some viruses such as Herpes Simplex Virus can travel along the axon independently of a membrane compartment , as capsids [31] and control its long distance transport process directly [32]–[35] . Interestingly , the RABV phosphoprotein P directly interacts with a dynein light chain [11] , [12] , suggesting a mechanism whereby this interaction is key to RABV's retrograde trafficking . However , studies on the retrograde transport of RABV enveloped virions [14] and infection of the CNS from the periphery with dynein light chain binding defective virus mutants [13] already showed that such an interaction is not essential for retrograde axonal transport of the virus . Our data support this finding , as we demonstrated that RABV is transported in acidic compartments ( Fig . 6 ) , and mostly in p75NTR-positive endosomes ( Fig . 7 ) . A different role for dynein binding should thus be considered . Dynein is well characterized as a retrograde motor , yet can also act to tether and stabilize dynamic microtubules [36] . Possibly , RABV binding to dynein tethers projecting microtubules ( MT ) in the cell cortex thereby facilitating its retrograde trafficking from the cell periphery . Following this tethering , RABV particles can merge into the RABV-p75NTR endosomes and travel to the neuron cell body . Nonetheless , this RABV-MT interaction could be mediated by binding of RABV to NCAM , which was demonstrated to tether MT's at the synapse [37] . Suggestions relating to the function of various RABV populations that can either internalize with receptors to endosomes , or act as RNP capsids to manipulate and stabilize the cytoskeleton , require further testing . We have shown that the RABV and the acidic marker LysoTracker are co-localized and move together along the axon . This shows that the RABV is transported in membranal compartments similar to neurotrophic factors and is unlikely to be free in the cytoplasm . Interestingly , low pH induces conformational changes in the viral G protein , suggesting control of membrane fusion events [38] , [39] that may regulate RABV transport in acidic vesicles . Moreover , it could be that the alteration in RABV G-protein function , as pH changes , provides a signal for RABV-p75NTR complex , leading to its transport acceleration . The effects of RABV binding to p75NTR on axonal transport processivity and speed , suggest that down-stream signaling is activated and exerts an influence on the retrograde transport process . Axonal transport can be regulated at four main different levels: 1 . Microtubule tracks . 2 . Motor proteins . 3 . Motor-cargo adaptors . 4 . ATP supply . As we have shown that RABV-p75NTR complexes move both instantaneously faster and with fewer pauses , we speculate that more than one regulatory level may be involved . p75NTR activates different signaling cascades , as a result of binding to several distinct ligands or interactions with various co-receptors . Recently it was described that structural determinants underlie the signaling specificity of p75NTR to the JNK , RhoA and NF-kB pathways [40] . Interestingly , c-Jun N-terminal kinases ( JNKs ) can also regulate the axonal transport process in several different ways . JNK-interacting proteins ( JIPs ) are scaffolding proteins for JNK and serve as linkers between motor proteins and their membrane-associated cargos . JIP1 serves as a linker between kinesin-1 and dynein to vesicles , and JNK signaling can modulate its transport by regulating the two opposing motors [41] . JNK , by functioning as a kinesin-cargo dissociation factor , regulates axonal transport [42] . Additionally , JNK3 phosphorylates kinesin-1 and inhibits its microtubule-binding activity [43] , [44] . JNK3 and its scaffolding protein Sunday driver ( syd ) are activated after axonal injury and bind to p150 , the regulatory sub-unit of the dynein-dynactin retrograde complex [45] . Moreover , a recent study has shown that JIP1 phosphorylation serves as a molecular switch to regulate the direction of vesicle transport in neurons , by coordinating kinesin and dynein motors [46] . These studies show that scaffolding proteins such as the JIPs and JNK play an important role in the regulation of motor proteins and the axonal transport process , and thus might take part in RABV manipulation of the axonal transport machinery . Another interesting speculation to explain how RABV binding to p75NTR accelerates its transport is the potential involvement of axonal activated NF-kB . NF-kB can be activated downstream to both p75NTR and RABV in the axon before entering the nucleous , and may affect dynein activity [47]–[49] . Interestingly , NF-kB activation after NGF binding to p75NTR enhances neuronal survival [50] , suggesting that p75NTR may regulate NGF retrograde signaling . In the future , it will be interesting to examine the involvement of axonal p75NTR dependent downstream signaling activated by RABV and its effect on the neuronal cytoskeleton as well as axonal transport . It is also tempting to consider that local protein synthesis as the result of RABV binding to p75NTR can facilitate this processive transport . Indeed recently it was demonstrated that efficient retrograde transport of pseudorabies virus , require axonal protein synthesis [51] . These data reveal an unexpected role for p75NTR that is not necessarily related to its functions as a neurotrophin receptor , but rather to acceleration of RABV-axonal transport . Whether p75NTR might also provide a fast delivery of other axonal cargos , such as neurotrophic factors , pathogenic prions and tetanus toxin [1] is a question for future research . This study has addressed the question of how RABV is transported over long distances . Previous cell biology work in the field was performed mostly on RABV-infected cell lines or neuronal cell bodies , leading to a focus on mechanisms of infection and not on long distance transport [52] . Although neurotropic viruses need to progress over long distances to reach the CNS [1] , [35] , how RABV performs this task was not clear . Here , we suggest that RABV hijacks a specific mechanism that enables the neuron to transport cargos over long distances . Interestingly , p75NTR is internalized by clathrin dependent endocytosis and is sorted into distally transported endosomes after stimulation with NGF [53] , [54] . Furthermore , RABV internalization was characterized as a dominantly clathrin mediated process [52] , [55] . Here we show that RABV binds to and is internalized together with p75NTR , forming endocytic compartments which undergo processive long distance transport . This suggests similar mechanism by which RABV mimics neurotrophins for activation of p75NTR ligand-mediated internalization and transport . As p75NTR can bind many ligands and various co-receptors , it is possible that binding to some will trigger a signaling effect that will facilitate axonal transport of other cargos and not only RABV . As we demonstrate here for RABV , p75NTR interaction can modulate the cellular transport machinery and may serve as a novel route to increase transport efficiency and facilitate arrival of cargos to the CNS from the periphery .
ICR mice were bred and maintained at the Tel Aviv University animal care facility until the time of sacrifice . Spinal cords were dissected from E12 . 5–13 . 5 mice , followed by separation of Dorsal Root Ganglia ( DRG ) from meninges and additional spinal cord . The Institutional Animal Care Committee at the Tel Aviv University approved all the animal protocols in this work . Microfluidic chambers were fabricated using methods previously described in detail [56] . All microfluidic chambers were replica molded using PDMS ( #41201841 Dow Corning ) from masters that were patterned using the photosensitive epoxy SU-8 ( Microchem ) . All masters consisted of two permanent SU-8 layers on a 3″ silicon wafer and were made in the clean room facility in Tel-Aviv University . The first layer of SU-8 ( 3 µm depth ) contained the microgrooves , which were patterned by photolithography using a high-resolution chromium mask ( 5 µm minimum feature size; Advance Reproduction Corp . ) . The second layer of SU-8 ( 100 µm depth ) contained the compartments , which were patterned by photolithography using a 20 , 000 dpi printed transparency mask ( CAD/Art Services , Inc . ) . Chamber dimensions: channels: length 8 . 25 mm , width 1 . 5 mm; grooves: length 400 µm , width 15 µm , height 5 µm ( Fig . 1B ) . A single 7 mm well was punctured into the “proximal” or explant channel , into which a “cave” was carved using a scalpel , to prevent explants from floating , 2 additional 1 . 2 mm wells were punctured into the channel on either side of the “explant” well to allow flow . Two 7 mm wells were punctured into both ends of the “distal” or axons channel to allow control over the distal channel . Microfluidic devices were cleaned of surface particles using adhesive tape and sterilized in 70% high-grade ethanol for 1 h . Devices were allowed to completely air dry under sterile conditions , attached to sterile 50 mm glass bottom dishes ( FD5040-100 , WPI ) using gentle pressure and heated to 70°C for 20′ to improve adhesion to glass . Chambers were coated using 150 µl of 1 . 5 ng/ml Polyornithine ( P-8638 , Sigma ) in PBS for 24 hours , which was replaced with 150 µl Laminin ( L-2020 , Sigma ) 1∶333 in DDW for 24 hours . Laminin was replaced with culture medium until plating ( 1–3 days ) . At the day of plating , media were removed from all wells , and a single DRG was inserted to each explant “cave” using a 20 µl tip . Following 1 hour of incubation at 37°C , 150 µl of medium were added to each well . Basic culture medium consisted of Neurobasal medium ( Life Technologies ) supplemented with 2% B-27 ( Life Technologies ) , 1% penicillin-streptomycin ( Biological Industries , Israel ) and 1% Glutamax ( Life Technologies ) . A gradient of murine NGF was created in order to encourage axons to cross the grooves to the distal axon compartment , by adding 100 ng/ml and 62 . 5 ng/ml murine NGF ( Alomone labs ) to the distal and proximal wells , respectively . Cultures were maintained at 37°C and 5% CO2 , and media were refreshed every 2 days . Transport assays were initiated after axons had crossed the grooves and established an axonal network in the distal compartment , 3–5 days from plating ( Fig . 1D ) . For TIRF imaging assays , 4–5 DRG were placed on 35 mm glass bottom dishes ( FD35-100 , WPI ) with 20 µl each of DRG culture medium supplemented with 62 . 5 ng/ml NGF . DRG were incubated in 37°C for 4–5 hours to allow adhesion to glass , then supplemented with 2 ml DRG culture medium supplemented with 62 . 5 ng/ml NGF . Cultures were maintained at 37°c and 5% CO2 for 48 hours before imaging . rRABV EGFP-P ΔG/rRABV mCherry-P ΔG are recombinant rabies viruses in which EGFP or mCherry fluorescent reporters were fused to the phosphoprotein P and in which the glycoprotein G gene was deleted . Preparation and amplification was performed as previously described [14] , [57] . In brief , MG-136 cells , expressing the RABV matrix and glycoproteins after induction with doxycycline , were infected with rRABV EGFP-P ΔG . Cells were split , after which media were twice replaced with fresh medium supplemented with 1 mg/ml Doxycycline , to be collected 48 hours later . rRABV EGFP-P ΔG was concentrated from cell culture supernatants using the PEG virus precipitation kit ( ab102538 , Abcam ) . HEK293t cells were transfected using calcium-phosphate precipitation with the viral vector pLL-EGFP and the helper plasmids pVSVG and pGag-PolGpt ( kind gift from Eran Bacharach ) or mix of 4 shRNA plasmids against murine p75 , produced from Mouse GIPZ lentiviral target gene shRNAmir glycerol set ( GE healthcare RMM4532-EG18053 ) . Viral particles were collected 48 and 72 hours after transfection , filtered and concentrated using the PEG virus precipitation kit ( ab102538 , Abcam ) . DRG were collected from E12 . 5–13 . 5 mice , trypsinized with Trypsin-EDTA solution B ( Biological industries 03-052-1B ) for 5 minutes in 37°C , and washed with complete F12 medium . Cells were dissociated by pipetting , counted and plated on PLO/Laminin coated 96 well plates , in a density of 15K cells per well . Cells were maintained with DRG culture medium supplemented with 62 . 5 ng/ml NGF . On day of plating , cells were infected with either LV-EGFP or a mix of 4 LV-shRNA-p75-EGFP ( MOI≈5 and MOI≈10 , respectively , due to high infectivity of LV-EGFP ) . 3–4 days from plating , cells were infected with ≈120K mCherry RABV particles for 0′ , 30′ and 120′ in duplicates after which they were washed ×3 with DRG culture medium . 48 hours later , neurons showing obvious mCherry foci were counted in 6–9 fields per well , and divided by total number of counted neurons . Although cultures varied in axon network , non-neuronal cell populations and LV-infection levels , infection rates were lower in sh-RNA-p75-EGFP groups when compared to GFP control in all experiments . Due to differences in absolute infection rates , these were normalized to the rate obtained at the LV-EGFP control group . Infection of DRG explants in microfluidic chambers were performed on the day of plating , with ≈1×106 LV-EGFP particles , or a mix of 4×≈1–2×106 LV-sh-RNA-p75-EGFP particles . To estimate knockdown of p75 , a DRG culture plated on a glass bottom dish was infected with sh-RNA-p75-EGFP as described above . At 4DIV , culture was treated with 1∶100 anti-p75-550 for 15′ and washed ×3 prior to imaging using spinning disc confocal . Qdot-NGF was prepared by mixing biotinylated murine-NGF 10 µg/ml ( #N-240-B , Alomone Labs ) with Quantum-Dot 605 streptavidin conjugate 1 µM ( Q10101 , Molecular Probes ) in a molar ratio of 3∶1 , respectively . Acidic compartments were delineated using Lysotracker Red DND-99 ( Life Technologies ) in a final concentration of 50 nM . Mitotracker Deep Red FM ( Life Technologies ) was used to denote mitochondria , in a final concentration of 100 nM . In order to track the co-transport of RABV with its receptors , the fluorescent extracellular antibodies Anti-p75NTR-ATTO-550 , Anti-p75-ATTO-488 and Anti-TrkA-ATTO-633 ( ANT-007-AO , ANT-007-AG and ANT-0180-FR , respectively , Alomone Labs ) were used in a 1∶100 dilution . Live imaging was performed on DRG explants grown in compartmental chambers , 3–5 DIV , upon forming an axonal network at the distal axon compartment . Prior to imaging , media from each well were replaced with 150 µl Neurobasal medium supplemented with 1% penicillin-streptomycin and 1% Glutamax ( poor medium ) . Following two hours of starvation , 30 µl of poor medium were added to the proximal well , to induce compartmental separation . 2 µl of concentrated EGFP-RABV ( ≈120K particles ) , Qdot-NGF or both were added to one distal well , with additional 5 µl medium , to encourage flow through the distal axon channel ( Fig . 1B , C ) . Chambers were incubated for 1–2 hours in 37°C and 5% CO2 prior their placement on the microscope stage . Time lapse images of axons in the groove area ( Fig . 1B , C ) , were acquired at 37°C and CO2 controlled environment , using Nikon Eclipse Ti microscope equipped with Yokogawa CSU X-1 spinning disc confocal , controlled via iQ software ( Andor ) . Multi-channel time lapses were captured using 60× lens , NA = 1 . 4 with 2000 msec intervals , with an approximate lag of 100–400 ms between channels . Digital images were taken with Andor iXon DU-897 EM-CCD camera . Time-lapse image analysis was carried out using Fiji , following subtraction of average intensity z-projection to exclude completely stationary fluorescent artifacts . XY coordinates of tracks were registered from 9 or more grooves ( which contains typically 2–5 axons ) in each experiment , using Manual Tracking plugin , while distances , velocities and MSD's were computed using MATLAB implementation for Fiji Trackmate plugin [58] . Tracks with run lengths <10 µm and/or average speed <0 . 2 µm/sec were filtered out of the analysis . A single puncta traveling with instantaneous velocity <0 . 1 µm/sec for three or more consecutive frames ( i . e 6 seconds ) , was considered paused . Instantaneous velocities were calculated as the distance a puncta traveled between two consecutive frames , divided by frame interval ( 2 sec ) . Speed distribution was presented for 0 . 2 µm/sec bins , and data were interpolated to include 103 points . Track speed was calculated as total run length divided by track duration , while displacement considered the net-distance a puncta traveled in retrograde direction . Kymographs were drawn using the Kymotoolbox plugin [59] In order to quantify co-localization of RABV with receptors and cellular components , RABV-EGFP images threshold was manually set , then RABV puncta detected using Fiji particle analysis feature , and saved as ROI's . ROI's were measured using Fiji measure function to illustrate both puncta diameter ( area ) and GFP intensity . Intensities were normalized to the average intensity of RABV puncta in each experiment . Images were subsequently matched with corresponding images from requested channel , and occurrences of co-localized puncta inside ROI's were manually counted . For random co-localization , RABV ROI's were matched with a random image from the same channel and time lapse . DRG explants were imaged with TILL Photonics iMIC TIRF microscope ( FEI , Munich , Germany ) , using Olympus 100× NA 1 . 49 TIRF objective . Images were acquired with an Andor iXon 897 EMCCD camera ( Belfast , UK ) and imaging protocol was controlled by TILL Photonics LiveAcquisition software . Throughout the experiment , explants were imaged inside an environmental chamber maintaining 37°C and 5% CO2 . TIRF angle was set to give minimal penetration of the evanescent wave still giving measurable signal from the RABV-EGFP/NGF-QD particles . We used the 360° TIRF feature which azimuthally spins the laser beam on the circumference of the objective back focal plane , creating homogenous TIRF illumination across the field . Broad neuronal tips were selected for imaging after RABV/NGF addition , selected fields were imaged for 5–10 minutes each , for up to 2 hours . Exposure times were 50 msec , and camera gain was set to 300 . Final frame rate was 1 and 1 . 12 seconds per frame in RABV/NGF alone and RABV+p75NTR experiments , respectively . A frame is comprised of RABV/NGF/p75 fluorescent channels over a bright-field image to localize fluorescent particles in context . For RABV-p75 imaging , explant cultures were incubated with fluorescent anti-p75NTR ( ANT-007-AO , Alomone Labs ) for 10 minutes and washed 3 times in poor neurobasal medium prior to imaging . Time-lapse image analysis was carried out using Fiji . RABV/NGF/p75 particles , which were manually identified in the area of the neuronal tip according to bright-field images , were marked by circular ROI in every time point . Average intensity was measured and plotted over time . A three-point moving average was applied to smooth out noisy spiking . As movement of RABV/NGF particle farther from the evanescent wave was measured as reduction in signal intensity , we used this property to detect and illustrate internalization kinetics . More specifically , internalization was defined as gradual reduction of RABV/NGF signal to undetectable levels for more than 4 consecutive frames ( 4–4 . 5 seconds ) . Internalization time was calculated as the time of the gradual reduction from high level intensity to baseline level . In RABV-p75 co-labeling experiments , p75 intensity was measured inside the identified RABV ROI .
|
Rabies virus ( RABV ) is a neurotropic virus that depends on long distance axonal transport in order to reach the central nervous system ( CNS ) . The strategy RABV uses to hijack the cellular transport machinery is unknown . Here we use live cell imaging to track RABV entry at nerve terminals and study its retrograde transport along the axon . First , we demonstrate that RABV interacts with the p75 neurotrophin receptor ( p75NTR ) at peripheral neuron tips to enter the axon . Then , characterizing RABV retrograde transport along the axon , we showed that the virus moves in acidic compartments , mostly with p75NTR . Interestingly , RABV is transported faster than NGF , an endogenous p75NTR ligand . Finally , we determine that p75NTR-dependent transport of RABV is faster and more directed than p75NTR-independent RABV transport . Hence , RABV not only exploits the neurotrophin transport machinery , but also has a positive influence on transport kinetics , thus facilitating its own arrival at the CNS .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"molecular",
"biology"
] |
2014
|
Rabies Virus Hijacks and Accelerates the p75NTR Retrograde Axonal Transport Machinery
|
Quantification of gene expression levels at the single cell level has revealed that gene expression can vary substantially even across a population of homogeneous cells . However , it is currently unclear what genomic features control variation in gene expression levels , and whether common genetic variants may impact gene expression variation . Here , we take a genome-wide approach to identify expression variance quantitative trait loci ( vQTLs ) . To this end , we generated single cell RNA-seq ( scRNA-seq ) data from induced pluripotent stem cells ( iPSCs ) derived from 53 Yoruba individuals . We collected data for a median of 95 cells per individual and a total of 5 , 447 single cells , and identified 235 mean expression QTLs ( eQTLs ) at 10% FDR , of which 79% replicate in bulk RNA-seq data from the same individuals . We further identified 5 vQTLs at 10% FDR , but demonstrate that these can also be explained as effects on mean expression . Our study suggests that dispersion QTLs ( dQTLs ) which could alter the variance of expression independently of the mean can have larger fold changes , but explain less phenotypic variance than eQTLs . We estimate 4 , 015 individuals as a lower bound to achieve 80% power to detect the strongest dQTLs in iPSCs . These results will guide the design of future studies on understanding the genetic control of gene expression variance .
Robustness , or the ability to maintain a stable phenotype despite genetic mutations and environmental perturbations , is an important property of many key biological processes , such as those underlying embryogenesis and development [1 , 2] . Conversely , evolvability , or the ability to generate heritable phenotypic variation , is a fundamental requirement of evolutionary processes [3] . A long-standing question in genetics , therefore , is how the balance between these two seemingly opposite processes has been fine-tuned [4] . To make progress in understanding the balance between robustness and evolvability , we need to characterize the mechanisms that underlie robustness . Robustness can arise through a number of different mechanisms: for example , redundancy of regulatory elements or feedback loops in regulatory circuits . In these different scenarios , we hypothesize evolvability could be maintained through different selective pressures . If we are able to characterize gene-specific robustness to expression variability , we can begin to ask about the balance between natural selection of gene function and the ability to maintain evolvability . In model organisms , robustness and evolvability can be studied using experimental evolution approaches . These approaches typically quantify robustness as the change in trait variation after applying an experimental perturbation [5 , 6] . However , in such experiments the phenotypic outcomes , rather than the underlying mechanisms of robustness , are measured . Moreover , experimental evolution studies have almost always considered population-average measurements of phenotypes using entire organisms , tissues , or cell cultures , with few exceptions [7 , 8] . To truly understand how robustness and evolvability are established and encoded in the genome , we need to consider phenotypic variation across individual cells [9] , and connect it to genetic variation , an approach termed “noise genetics” [10] . Using the yeast Saccharomyces cerevisiae as a model system , studies have shown that heterogeneity in the expression of certain genes across cells is highly heritable and placed under complex genetic control , suggesting that the level of noise in gene regulation may also differ between individuals of multicellular organisms depending on their genetic background [11] . Follow-up studies further demonstrated that gene expression noise mediated by promoter variants can provide a fitness benefit at times of environmental stress in yeast , highlighting the direct role of genetically controlled stochastic cell-cell variation in evolutionary robustness [12] . However , the genetic and molecular circuits that lead to robustness remain largely uncharacterized in mammals . Here , we take an unbiased , genome-wide approach to identify quantitative trait loci associated with gene expression variance across cells ( vQTLs ) . We study human induced pluripotent stem cells ( iPSCs ) , which offer a homogeneous population of cells allowing a relatively simple statistical model . Investigating iPSCs also provides the possibility to study gene expression variance across cells during differentiation in follow-up studies . To directly measure the mean and variance of gene expression within cell populations as phenotypes , we generated single cell RNA-seq ( scRNA-seq ) data from cells derived from multiple individuals .
Using the Fluidigm C1 platform , we isolated and collected scRNA-seq from 7 , 585 single cells from iPSC lines of 54 Yoruba in Ibadan , Nigeria ( YRI ) individuals . We used unique molecular identifiers ( UMIs ) to tag RNA molecules and account for amplification bias in the single cell data [13] . To estimate technical confounding effects without requiring separate technical replicates , we used a mixed-individual plate study design ( Fig 1A ) . The key idea of this approach is that having observations from the same individual under different confounding effects and observations from different individuals under the same confounding effect allows us to distinguish the two sources of variation [14] . We excluded data from one individual ( NA18498 ) with evidence of contamination , then filtered poor quality samples as previously described [14] . After quality control , we analyzed the expression of 9 , 957 protein-coding genes in a median of 95 cells per individual in 53 individuals ( total of 5 , 597 cells; S1 Fig ) . To ensure that our measurements are comparable across samples , we first sought to assess the impact of observed technical variation on the data and to identify unobserved technical confounders . To this end , we performed principal components analysis ( PCA ) on the matrix of log counts per million ( log CPM ) . We found that across samples , the loading on the top principal component ( PC ) was correlated with gene detection rate ( the proportion of genes with at least one molecule detected ) , but not with the biological variable of interest ( individual ) or the expected technical confounders ( batch and C1 chip; Fig 1B ) . Indeed , as previously reported [15] , the entire distribution of observed log CPM ( over all genes ) varies across samples , and appears to be associated with the gene detection rate ( Fig 1C ) . After accounting for gene detection rate ( Methods ) , the top PCs were correlated with individual , batch , and C1 chip , as expected ( Fig 1D ) . We developed a method to estimate the mean and variance of gene expression across cells for each gene in each individual ( Fig 1A; Methods , S1 Text ) . Briefly , for each individual and each gene , our method uses maximum likelihood to fit a zero-inflated negative binomial distribution ( ZINB ) to the observed UMI counts across cells , and derives the mean and variance of gene expression from the estimated model parameters . When fitting the ZINB model the method controls for technical confounders ( e . g . C1 chip ) and library size , and when deriving the mean and variance it accounts for Poisson measurement noise in the UMI counts [16 , 17] . These desirable , and arguably crucial features would not be achieved by directly computing the sample mean and variance of either the UMI counts or log CPM . To evaluate the accuracy of the method , we first simulated data from the model and compared the estimated parameters , as well as the derived mean and variance , to the true values used to generate the data . We fixed the number of cells and number of molecules detected per cell to the median of those values in our observed data , and varied the ZINB parameters . Assuming that mean expression is high enough , we found the method produces accurate estimates of the underlying negative binomial parameters , but not the zero inflation parameter ( S2 Fig ) . Despite not accurately estimating the zero inflation parameter , the method still produces accurate estimates of the derived mean and variance for genes that are expressed at intermediate to high levels . Next , we tested for goodness of fit on each simulated data set ( Methods ) . The key idea underlying the test is that if the data are truly distributed according to some cumulative distribution function F , then the values of F evaluated at the data should be uniformly distributed between 0 and 1 . Applying the test to the simulated data , we rejected the null that the model fit the data for zero of 2 , 451 simulation trials after Bonferroni correction ( p < 2 × 10−5; S3 Fig ) . The results suggest the method is successfully able to fit the observed data , and also suggest that inaccuracy in the estimated parameters is likely explained by noise due to small sample sizes . We then applied our method to the observed data , correcting for batch and C1 chip . Importantly , we did not correct for gene detection rate , reasoning that the dependence on gene detection rate is only an artifact introduced by analyzing log CPM . We tested the goodness of fit for each individual and each gene , and rejected the null that the model fit the data for only 60 of 537 , 658 individual-gene combinations ( 0 . 01% ) after Bonferroni correction ( p < 9 × 10−8; S4 Fig ) . Our results emphasize that careful experimental design as well as careful statistical modeling are required to robustly map effects on gene expression variance across cells . Previous studies have shown a clear relationship between the mean and variance of gene expression [18 , 19]; therefore , apparent genetic effects on the variance could potentially be explained by effects on the mean . In our model , the mean-variance relationship is controlled by a single dispersion parameter per gene per individual . We sought to directly map QTLs which could alter the variance independently of altering the mean by using the estimated dispersion parameter as a quantitative phenotype . However , we found zero dispersion QTLs ( dQTLs ) using this approach ( FDR 10% ) . Further , we found the QQ plot of association p-values did not show deviation from the null ( S5 Fig ) . Alternative approaches to decouple the mean-variance relationship include using the coefficient of variance ( CV; ratio of standard deviation to mean ) or Fano factor ( ratio of variance to mean ) as quantitative phenotypes . However , prior work shows these quantities have predictable relationships with the mean , and therefore effects could still be explained away [14 , 19] . Therefore , we proceeded to map eQTLs , variance QTLs ( vQTLs ) , CV-QTLs , and Fano-QTLs , and then asked whether we could discover variance effects which could not be explained as effects on mean expression . We found 235 eQTLs , 5 vQTLs , 0 CV-QTLs , and 0 Fano-QTLs ( FDR 10%; S5 Fig ) . To validate the eQTLs , we estimated the replication rate against eQTLs discovered in bulk RNA-seq from the same iPSC lines [20] . We found that 79% of the single cell eQTLs replicate in the matched bulk data ( Fig 2A ) , and 80% of bulk eQTLs replicate in the single cell data . However , we found 1 , 390 eQTLs ( FDR 10% ) using all of the individuals in the bulk RNA-seq study ( n = 58 ) , and still recovered 1 , 136 eQTLs ( FDR 10% ) after subsampling to n = 53 . Our results therefore suggest that eQTL discovery in scRNA-seq ( as opposed to replication of previously discovered eQTLs ) loses power compared to equal-sized studies in bulk RNA-seq , likely due to increased experimental noise . We found 85% of the eQTLs were also discovered as vQTLs ( when restricting to testing only at the eQTL SNP ) , and 100% of vQTLs were discovered as eQTLs ( Fig 2B ) . We then sought to directly explain away vQTLs as eQTLs by regressing out the mean from the variance . Treating the residuals from the regression as the phenotype , we recovered zero vQTLs . These results suggest the significant variance effects detected in this study are all likely to be explained as effects on mean expression . Our goal in this study was to find QTLs which alter the variance of gene expression independently of altering the mean expression . Under our model , these QTLs should explain variation in the dispersion parameter across individuals; however , we failed to find dQTLs . Further , all of the vQTLs we were able to identify could be explained by mean effects . In contrast , we were able to discover eQTLs , but fewer than expected based on bulk RNA-seq in matched samples . To understand why we failed to discover dQTLs , and why we discovered fewer eQTLs than expected , we first derived the power function in terms of effect size ( log fold change ) , sample size , noise ratio ( ratio of measurement error variance to phenotypic residual variance ) , and significance level ( Methods ) . We then sought to estimate the distribution of QTL effect sizes and the typical noise ratio , for both mean expression and dispersion . To estimate the distribution of QTL effect sizes , we fit a flexible unimodal distribution for the true effect sizes which maximizes the likelihood of the observed effect sizes and standard errors [21] . Surprisingly , we found that dQTL effects could be larger than eQTL effects ( S6 Fig ) . For example , we estimate that the 99th percentile eQTL effect size is 0 . 022 , but is 0 . 090 for dQTLs . Given this result and the power function we derived , there are two possible explanations for why we still failed to find dQTLs: ( 1 ) the noise ratio of dispersion is large ( measurement error reduced power ) , or ( 2 ) the residual variance of dispersion is large ( genetic variation explains little phenotypic variance ) . To estimate the typical noise ratio , we developed a two-step procedure to estimate the measurement error variance and residual variance per gene ( Methods ) . Briefly , in our approach we have one measurement error variance per individual , per gene , which equals the sampling variance of our ZINB model . To estimate each error variance , we used non-parametric bootstrapping . To estimate the measurement error variance for each gene , we took the median of the estimated measurement error variances across individuals . To estimate the residual variance for each gene , we fit a flexible unimodal distribution for the true phenotypes which maximizes the likelihood of the observed phenotypes and measurement errors , and estimated the variance of the posterior mean true phenotypes . Using our approach , we estimated that the typical noise ratio of the dispersion is 2 . 99 , compared to 4 . 18 for the mean ( S7 Fig ) . This result suggests that we did not fail to find dQTLs only due to measurement error , because the noise ratio was lower for dQTLs than for eQTLs . As a reference point , a noise ratio equal to 1 has the same impact on power to detect a QTL as cutting the sample size in half , explaining why our study lost power to detect eQTLs . We found that the typical phenotypic standard deviation of dispersion is 7 . 2 fold larger than that of the mean expression , suggesting we failed to find dQTLs because the effect sizes of dQTLs ( relative to phenotypic standard deviation ) are smaller than the effect sizes of eQTLs . We finally asked how much power our current study had to detect the 99th percentile dQTL effect size , assuming the typical noise ratio estimated above . We found that our study had only 0 . 001% power to detect that effect size at Bonferroni–corrected level α = 5 × 10−6 ( Fig 3 ) . Fixing the typical noise ratio ( a function of the number of cells per individual and sequencing depth ) , we estimate 16 , 015 individuals would be required to achieve 80% power . As a lower bound ( setting the noise ratio to zero ) , we estimate 4 , 015 individuals would be required regardless of the number of cells per individual . Overall , our results suggest a much larger study , both in terms of number of individuals and number of cells per individual , would be required to detect the strongest dQTLs in iPSCs .
Individual cells must tolerate both external and internal perturbations arising from the environment or mutations . It has long been argued that this outcome of robustness is an inherent property of biological systems [22] , and arises from natural selection [23 , 24] . Robustness is especially critical in the context of cell fate transitions during differentiation [25] . Other dynamic physiological processes must also be robust , and as a result , loss of robustness is associated with clinically relevant phenotypes and complex genetic disease [26 , 27] . Cells maintain their identity and other phenotypes despite perturbations because of the robust regulation of key sets of genes [28] . We hypothesized that QTLs could disrupt the mechanisms underlying robust regulation , and therefore reveal new insights into the genetic regulation of differentiation and disease . To investigate this hypothesis , we directly observed gene expression variance across multiple individuals using scRNA-seq , and sought to identify QTLs which could alter the variance of gene expression across cells within a single individual , independently of altering the mean expression . However , we failed to discover such QTLs , and demonstrated that QTLs which are associated with the variance of gene expression can be explained by effects on mean expression . We found that relative to the phenotypic standard deviation , effects on the dispersion are smaller than effects on the mean , partially explaining why this study failed to find them . Our results do not rule out genetic effects on variance independent of mean effects , due to limitations of our analysis . First , our estimated distributions of effect sizes are based on an empirical Bayes estimate of the underlying effect sizes , given the observed effect sizes . Our results in simulation and observed data suggest the observed effect sizes may be not be accurately estimated given the size of the current study . Therefore , the empirical Bayes estimate may not accurately reflect the true distribution of effect sizes . However , we chose to bias the estimation procedure towards putting prior mass on zero , so our estimates of effect sizes are conservative . Additionally , our estimates may not generalize beyond iPSCs , because the distribution of dispersion effect sizes could vary across cell types and conditions . Second , we made a strong assumption that latent gene expression is point-Gamma distributed . In this study , we directly assessed whether or not this was true using a simple statistical diagnostic , and did not find any gross violations of this assumption in the data . However , it is likely that this assumption will be violated in heterogenous populations of cells . One possible extension of our method to this case would be to assume there are K homogeneous subpopulations of cells , each described by a ( possibly different ) point-Gamma distribution . This mixture of ZINB model suggests an expectation-maximization approach where each cell is assigned to a subpopulation , and then the distributions of the subpopulations are re-estimated . Finally , we took a modular approach to map QTLs in this study: ( 1 ) we estimated parameters for each individual using only the scRNA-seq data , and then ( 2 ) we mapped QTLs using phenotypes derived from the estimated parameters . An alternative approach would be to include genotype in the count model for the data , and jointly learn the mean , dispersion , proportion of excess zeros , and genetic effect sizes for mean and dispersion . Such an approach could borrow information across cells with common genotypes to improve power , holding the experiment size fixed . However , further development will be needed to efficiently fit the models at QTL mapping scale . We stress that our power calculation is only a rough guideline for designing QTL mapping studies using scRNA-Seq . Intuitively , some minimum number of cells per individual is required to adequately estimate means and variances . However , having achieved that lower bound , the most important quantity to maximize is the number of individuals . In support of this argument , we estimate thousands of individuals would be required to detect a dQTL no matter how many cells were collected per individual . We based our power calculations on typical values of the noise ratio for the mean expression and dispersion , and chose a conservative significance level . However , we found considerable variation in the noise ratio across genes , suggesting that our results may not generalize even across genes . Overall , our results suggest that the technical noise introduced by scRNA-seq greatly reduces the power to discover eQTLs . Our results also suggest that , for iPSC lines , dramatically larger studies will be required to map both eQTLs and dQTLs from scRNA-seq .
The cell lines used in this study were obtained from the NHGRI Sample Repository for Human Genetic Research at the Coriell Institute for Medical Research . All samples were collected by the Coriell Institute for Medical with written informed consent and with IRB approval . We cultured YRI iPSCs [20] in feeder-free conditions for at least ten passages in E8 medium ( Life Technologies ) [29] . We collected cells using the C1 Single-Cell Auto Prep IFC microfluidic chip ( Fluidigm ) . We used a balanced block-incomplete design to randomize individuals across chips . For each chip , we freshly prepared a mixture of cell suspensions from four individuals . We measured live cell number via trypan blue staining ( ThermoFisher ) , to ensure equal cell numbers across individuals per mixture . We performed single cell capture and library preparation as previously described using 6 bp Unique Molecular Identifiers [14] . We pooled the 96 samples on each C1 chip and sequenced them on an Illumina HiSeq 2500 using the TruSeq SBS Kit v3-HS ( FC-401-3002 ) . We mapped the reads to human genome GRCh37 ( including the ERCC spike-ins ) with Subjunc [30] , deduplicated the UMIs with UMI-tools [31] , and counted molecules per protein-coding gene ( Ensembl 75 ) with featureCounts [32] . We then matched single cells back to YRI individuals using verifyBamID [33] . We filtered samples on the following criteria , derived as previously described [14]: We filtered genes for QTL mapping on the following criteria: We applied principal component analysis ( PCA ) to the matrix X of log counts per million ( log CPM ) , using the pseudocount proposed in edgeR [34] . We corrected for gene detection rate by simultaneously regressing out quantiles of gene expression , correcting for sample-specific and gene-specific means , and performing PCA . Let X = ( x1 , … , xn ) be observed p-vectors , and let ( z1 , … , zn ) be latent k-vectors where k ≪ p . Then , PCA corresponds to maximum likelihood estimation in the following latent variable model [35]: x i ∼ N ( · ; W z i + μ , σ 2 I ) ( 1 ) In this parameterization , μ denotes a per-coordinate mean ( in our application , per-gene ) . However , as previously reported [15] , we additionally have to account for the per-sample mean . Our approach is based on the latent variable model: x i j ∼ N ( W j z i + q i ′ β j + u i + v j , σ 2 I ) ( 2 ) where u is an n-vector of per-sample means , v is a p-vector of per-gene means , and Q = ( q1 , … , qn ) is a n × k matrix of expression quantiles . We fit the model as follows: We estimated the squared correlation between each PC and categorical covariates ( batch , C1 chip , individual , well ) by recoding each category as a binary indicator , fitting a multiple linear regression of the PC loadings against the binary indicators , and then estimating the coefficient of determination of the model . We assume the count data are generated by a zero-inflated negative binomial ( ZINB ) distribution ( S1 Text ) . Let: Then , we assume: r i j k∼ Poisson ( · ; R i j exp ( x i j ′ β k ) λ i j k ) ( 6 ) λ i j k∼ π i k δ 0 ( · ) + ( 1 - π i k ) Gamma ( · ; μ i k , ϕ i k ) ( 7 ) Under this model , the mean and variance of gene expression are: E [ λ i j k ]= ( 1 - π i k ) μ i k ( 8 ) V [ λ i j k ]= ( 1 - π i k ) μ i k 2 ϕ i k + π i k ( 1 - π i k ) μ i k 2 ( 9 ) Considering just the non-zero component , marginalizing out λ yields the negative binomial ( NB ) log likelihood , weighted by 1 − πik: l ( · ) = ln ( 1 - π i k ) + r i j k ln ( R i j exp ( x i j ′ β k ) μ i k R i j exp ( x i j ′ β k ) μ i k + ϕ i k - 1 ) + ϕ i k - 1 ln ( ϕ i k - 1 R i j exp ( x i j ′ β k ) μ i k + ϕ i k - 1 ) + ln Γ ( r i j k + ϕ i k - 1 ) - ln Γ ( r i j k + 1 ) - ln Γ ( ϕ i k - 1 ) ( 10 ) Then , marginalizing over the mixture yields the ZINB log likelihood: ln p ( r i j k ∣ · ) = ln ( π i k + exp ( l ( · ) ) ) if r i j k = 0 ( 11 ) ln p ( r i j k ∣ · ) = l ( · ) otherwise ( 12 ) To estimate the model parameters , we maximized the ZINB log likelihood . The parameters must satisfy the constraints μik > 0 , ϕik > 0 , 0 ≤ πik ≤ 1 . To make the problem easier , we re-parameterized in terms of ln μik , ln ϕik , logit ( πik ) and performed unconstrained optimization . The ZINB log likelihood is nonconvex; therefore , we used a two stage optimization procedure . In the first stage , we optimized the NB log likelihood with respect to ln μik , ln ϕik , initializing from zero . In the second stage , we used the NB solution and logit ( πik ) = −8 ( corresponding to a suitably small value of πik ) as the initialization and optimized the ZINB log likelihood . In both stages , we used batch gradient descent for 30 , 000 iterations with fixed learning rate 10−3 , accelerated by RMSProp [36] . We implemented the method in Tensorflow [37] . We defined the size factor of each cell as the total number of molecules detected ( before excluding genes in QC ) . To correct for technical confounders , we included C1 chip as an observed confounder , recoded as binary indicator variables and centered . This approach is sufficient to also correct for batch , because in our experimental design , batch is a linear combination of C1 chip . Intuitively , if there were a batch effect independent of C1 chip , then we could add the batch effect to each chip effect and set the batch effect to 0 . To assess the goodness of fit of the method , we used a diagnostic test based on the following simple fact: if the data x1 , … , xn are continuous random variables generated from a continuous CDF F , then F ( xi ) ∼ Uniform ( 0 , 1 ) . Then , to test for goodness of fit of an estimated F ^ to the data x1 , … , xn , we apply the Kolmogorov-Smirnov ( KS ) test to test whether the values F ^ ( x 1 ) , … , F ^ ( x n ) are uniformly distributed . ( This test is slightly conservative because it uses the data to estimate F ^ ) . Here , we have to modify this simple procedure to account for the fact that our data are discrete counts , so F is not continuous . To address this issue , we used randomized quantiles [38]: we sample one random value per observation u i ∣ x i ∼ Uniform ( F ^ ( x i - 1 ) , F ^ ( x i ) ) . These have the property that if xi ∼ F then ui ∼ Uniform ( 0 , 1 ) . In our model , each observed UMI count xijk comes from a different distribution Fijk , because it depends on the library size which is cell-specific . We therefore draw u i j k ∣ x i j k ∼ Uniform ( F ^ i j k ( x i j k - 1 ) , F ^ i j k ( x i j k ) ) . Then , for each individual i and gene k , we apply the KS test to whether the randomized quantiles uijk across cells j are uniformly distributed . We imputed dosages for 120 Yoruba individuals from the HapMap project ( Phase 3 , hg19 ) as previously described [39] . We restricted our analysis to 8 , 472 , 478 variants with minor allele frequency at least 0 . 05 . For each single cell expression phenotype tested , we standardized and quantile-normalized the phenotype matrix to a standard normal as previously described [40] . We called QTLs within 100 kilobases of the transcription start site of each gene and controlled the gene-level false discovery rate using QTLtools [41] . We included principal components ( PCs ) of the normalized expression matrix as covariates for QTL mapping , and selected the number of PCs for each phenotype by greedily searching for the number of PCs which maximized the number of QTLs discovered on even chromosomes only at FDR 10% . We did not include genotype PCs as covariates . We additionally recalled eQTLs in the matched bulk RNA-seq data [20] using the re-processed dosage matrix . We performed replication testing by taking each SNP-gene pair from the discovery cohort , and testing that pair in the replication cohort . We defined a hit as replicating if it passed the Benjamini–Hochberg procedure at level 10% ( restricted to the set of SNP-gene pairs tested ) and had the same effect size direction . For individual i and gene k , we assume the generative model: y i k= x i b + e i k ( 13 ) y ˜ i k= y i k + e ˜ i k ( 14 ) where y ˜ i k is the observed phenotype , yik is the true phenotype , xi is the genotype at the SNP of interest , e ˜ i k ∼ N ( 0 , σ m 2 ) , and e i k ∼ N ( 0 , σ r 2 ) . To perform QTL mapping , we fit a working model which ignores measurement error: y ˜ i k = x i β + ϵ i k ( 15 ) where ϵik ∼ N ( 0 , σ2 ) . From this model , we estimate β ^ . Assuming V [ x ] = 1 , we have σ 2 = σ r 2 + σ m 2 and: β ^ ∼ N ( b , σ r 2 + σ m 2 n ) ( 16 ) where n is the number of individuals . Under the working model , the power function is: Pow ( · ) = Φ ( Φ - 1 ( α 2 ) + b SE ( β ^ ) ) ( 17 ) where α denotes the significance level , SE ( ⋅ ) denotes standard error , and Φ ( ⋅ ) denotes the standard Gaussian CDF . Under the assumed generative model , the power function equals: Pow ( λ , n , δ , α ) = Φ ( Φ - 1 ( α 2 ) + λ n 1 + δ ) ( 18 ) where λ = b/σr , and δ = σ m 2 / σ r 2 . Parameterized in terms of δ , the power function implies useful reference points; for example , δ = 1 is equivalent to cutting the sample size in half . To determine the effect size b , we estimate the distribution of true effect sizes b given observed effect sizes β ^ j and associated standard errors s ^ j . We assume the hierarchical model: β ^ j ∣ b j , s ^ j∼ N ( b j , s ^ j 2 ) ( 19 ) b j ∣ s ^ j∼ g ( · ) ( 20 ) where g is a unimodal mixture of Gaussians . We estimate g using adaptive shrinkage ( ash ) [21] . We took b to be the 99th percentile of the fitted distribution . Although we assumed a single measurement error variance σ m 2 , we actually have measurement errors for each individual and gene σ m i k 2 . To estimate σ m i k 2 , we used non-parametric bootstrapping . For each individual and gene , we resampled the counts ( matched with the library size and technical confounders ) with replacement , and refit the ZINB model . To reduce computational burden , we restricted our analysis to 200 randomly chosen genes , warm-started the optimization from the optimal parameters for the original data , and ran gradient descent for 30 , 000 iterations . To estimate the typical noise ratio δ , we estimate a measurement error variance per gene σ m k 2 and a residual variance per gene σ r k 2 . We take σ ^ m k 2 = median ( σ m i k 2 ) . To estimate σ r k 2 , we solve a deconvolution problem [42]: y ˜ i k ∣ y i k , σ ^ m i k 2∼ N ( y i k , σ ^ m i k 2 ) ( 21 ) y i k ∣ σ ^ m i k 2∼ g ( · ) ( 22 ) where g is a unimodal mixture of uniforms , estimated using ash . To fit the model , we centered the y ˜ i k for each gene k , concatenated them across genes , and assumed a common prior . Then , the required estimates are: σ ^ r k 2= V ^ [ E [ y i k ∣ · ] ] ( 23 ) δ ^= median ( σ ^ m k 2 σ ^ r k 2 ) ( 24 ) λ= b median ( σ ^ r k 2 ) ( 25 ) where V ^ denotes sample variance .
|
Common genetic variation can alter the level of average gene expression in human tissues , and through changes in gene expression have downstream consequences on cell function , human development , and human disease . However , human tissues are composed of many cells , each with its own level of gene expression . With advances in single cell sequencing technologies , we can now go beyond simply measuring the average level of gene expression in a tissue sample and directly measure cell-to-cell variance in gene expression . We hypothesized that genetic variation could also alter gene expression variance , potentially revealing new insights into human development and disease . To test this hypothesis , we used single cell RNA sequencing to directly measure gene expression variance in multiple individuals , and then associated the gene expression variance with genetic variation in those same individuals . Our results suggest that effects on gene expression variance are smaller than effects on mean expression , relative to how much the phenotypes vary between individuals , and will require much larger studies than previously thought to detect .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[] |
2019
|
Discovery and characterization of variance QTLs in human induced pluripotent stem cells
|
Stimulus-specific adaptation ( SSA ) in single neurons of the auditory cortex was suggested to be a potential neural correlate of the mismatch negativity ( MMN ) , a widely studied component of the auditory event-related potentials ( ERP ) that is elicited by changes in the auditory environment . However , several aspects on this SSA/MMN relation remain unresolved . SSA occurs in the primary auditory cortex ( A1 ) , but detailed studies on SSA beyond A1 are lacking . To study the topographic organization of SSA , we mapped the whole rat auditory cortex with multiunit activity recordings , using an oddball paradigm . We demonstrate that SSA occurs outside A1 and differs between primary and nonprimary cortical fields . In particular , SSA is much stronger and develops faster in the nonprimary than in the primary fields , paralleling the organization of subcortical SSA . Importantly , strong SSA is present in the nonprimary auditory cortex within the latency range of the MMN in the rat and correlates with an MMN-like difference wave in the simultaneously recorded local field potentials ( LFP ) . We present new and strong evidence linking SSA at the cellular level to the MMN , a central tool in cognitive and clinical neuroscience .
A critical function of the brain is to identify uncommon and potentially important stimuli while ignoring irrelevant ambient backgrounds [1–3] . In humans , this ability is reflected by an electrophysiological brain response called mismatch negativity ( MMN ) , a mid-late ( 150–200 ms ) deflection of the auditory event-related potentials ( ERP ) that is elicited by uncommon , but not by repetitive , sounds [4–7] and serves to automatically redirect attention toward potentially relevant stimuli [8] . Importantly , the MMN signal is altered in patients with schizophrenia and other psychiatric disorders and can be used as an index of cognitive decline in normal and pathological neurodegenerative processes [9 , 10] . The MMN has been extensively studied using the “oddball” paradigm , in which infrequently occurring sounds , i . e . , “deviant” tones , are randomly interspersed among frequent monotonous sounds , i . e . , “standard” tones . MMN studies have advanced our knowledge on many aspects of change and novelty detection , but scalp recordings limit our ability to pinpoint its regions of generation . Recent studies over the past decade have taken advantage of the oddball paradigm to study adaptation in single auditory neurons . Stimulus-specific adaptation ( SSA ) may be a counterpart phenomenon to MMN that is studied in single neurons using this paradigm [11] . As in MMN , neurons showing SSA adapt to frequently occurring stimuli ( standards ) yet respond strongly to rare stimuli ( deviants ) . Within the auditory system , SSA was originally reported in the primary auditory cortex ( A1 ) [12] as a higher level of adaptation to a specific stimulus , different from firing rate adaptation resulting from changes in the intrinsic properties of the neuron . SSA shares many properties with the MMN , and it is important because it may be a neural correlate of the MMN , or at least one of its early generators [11 , 13] . The basic properties of SSA have been studied in great detail not only in A1 but also in the subcortical inferior colliculus ( IC ) [14–16] and medial geniculate body ( MGB ) [17 , 18] . One important difference between SSA in the auditory cortex and subcortical stations is their anatomical location . SSA is strong and widespread only in the nonlemniscal regions of the IC and MGB [16] , while SSA has been described as strong and widespread in lemniscal A1 [12 , 19] . However , detailed studies on SSA within the different cortical fields beyond A1 are lacking . Since SSA is stronger in the nonlemniscal regions of the IC and MGB , it is reasonable to hypothesize that SSA in the nonprimary regions of the auditory cortex would also be stronger than in A1 . Indeed , previous studies on the general response properties of the auditory cortex reported that nonprimary neurons in the cat [20 , 21] and rat [22–24] auditory cortex adapt more strongly than in A1 . Even studies in human subjects have shown differential adaptation between primary and nonprimary cortical areas [25–27] . Moreover , two recent studies that mapped auditory ERPs in the rat showed robust MMN-like responses in nonprimary auditory cortical fields [28 , 29] . The main goal of the present study was to generate a complete and fine-grained map of SSA across all known cortical fields in the rat . Despite interspecies differences , the rat auditory cortex shares many common anatomical and physiological features with other species [23 , 30 , 31] , including primary and nonprimary regions . Primary regions of the auditory cortex are characterized by a thick , dense , granular layer and receive major layer IIIb/IV thalamocortical projection from the first-order ( or lemniscal ) auditory thalamus . The nonprimary auditory cortex is formed by surrounding regions that subsequently process input from primary regions and receive major layer IIIb/IV projection from the higher-order ( or nonlemniscal ) auditory thalamus [31] . Detailed electrophysiological mapping studies [23 , 24 , 32] have identified at least five tonotopically organized fields in the rat auditory cortex . The A1 , the anterior auditory field ( AAF ) , and the ventral auditory field ( VAF ) are all considered primary fields [23 , 33] . Additionally , two distinct nonprimary regions have been identified: the posterior auditory field ( PAF ) , located in the dorsocaudal border of A1; and the suprarhinal auditory field ( SRAF ) , in the ventral margin of the auditory cortex [23 , 31 , 34 , 35] . Unfortunately , there are no specific stains or molecular markers that cause one cortical region in the rat to stand out unambiguously from another , but they show a robust organization of multiple response properties that follow a particular spatial organization [23 , 36] . Our results demonstrate that , although SSA is indeed present in A1 and the other two primary fields , it is markedly stronger in the nonprimary fields PAF and SRAF , consistent with the SSA observed in nonlemniscal parts of the IC and MGB . Another important finding in our data is that SSA observed in auditory cortex is robust up to 200 ms after stimulus onset , well within the latency range of the MMN-like potentials in the rat [37] . These data suggest the existence of a hierarchically organized system for SSA processing [13] and reinforce the notion that nonprimary SSA is a more direct neural correlate of the MMN than the SSA observed in A1 .
The main aim of this study was to quantify and compare SSA levels between the five cortical fields . Thus , we computed the stimulus-specific adaptation index ( SI ) for each stimulus , SI ( f1 ) and SI ( f2 ) , and the common SSA index ( CSI ) for every recording site , using baseline-corrected spike counts during stimulus presentation ( 5 to 80 ms from stimulus onset; see Materials and Methods ) . Fig 3A shows a series of scatterplots illustrating the joint distribution of SI ( f1 ) and SI ( f2 ) , for the whole population and for each field separately , and Fig 3B illustrates corresponding histograms of CSI distributions ( total number of recording sites included in this analysis , as detailed in Materials and Methods , are also indicated ) . In all cases , points are symmetrically clustered around the main diagonal , with no significant differences between the median SI ( f1 ) and SI ( f2 ) for any field ( paired Wilcoxon signed rank test , p > 0 . 1 in all fields ) , indicating that adaptation was equal on average for f1 and f2 . The drift of the population medians toward the upper-right corner ( Fig 3A ) reveals a gradual shift of the cloud of points , from A1 to PAF fields , toward higher levels of SSA . The global population shows a CSI distribution that is slightly skewed to the right ( Fig 3B , top panel ) . The origin of this skewness emerges once we split these distributions into the five cortical fields: the CSI distributions for the primary fields , especially A1 and AAF , are more symmetrical , centered on medium CSI values , and span the full range of possible values ( Fig 3B ) . The same distributions for the nonprimary fields , SRAF and PAF , on the other hand , are clearly asymmetric , sharply skewed to the right toward the extreme positive CSI values , with a virtual absence of low CSI values . Moreover , the center of the distribution progressively moves to the right ( i . e . , toward higher CSI values ) from A1 to PAF , ( CSI , [Q1 , median , Q3]: A1 , [0 . 22 , 0 . 38 , 0 . 61]; AAF , [0 . 32 , 0 . 50 , 0 . 68]; VAF , [0 . 39 , 0 . 56 , 0 . 72]; SRAF , [0 . 57 , 0 . 76 , 0 . 90]; PAF , [0 . 56 , 0 . 76 , 0 . 89] ) , with the median CSI in every primary field being significantly smaller than in every nonprimary field ( Kruskall-Wallis test , χ2 ( 4 ) = 121 . 43 , p < 5×10−24 ) . Correcting for baseline activity was required to measure the actual evoked response , given the high spontaneous rates seen in many recordings , particularly from the nonprimary fields ( spontaneous firing rate , mean ± SEM: A1 , 8 . 2 ± 0 . 7 spk/s; AAF , 7 . 3 ± 0 . 6 spk/s; VAF , 10 . 7 ± 0 . 7 spk/s; SRAF , 9 . 2 ± 0 . 6 spk/s; PAF , 13 . 0 ± 1 . 0 spk/s ) . This correction may have a major impact when using a contrast index such as the CSI [38] , so that higher CSI values in nonprimary fields could result in part from this procedure . Therefore , we repeated the CSI calculation using the absolute spike counts for the same time window . As expected , all CSI values were overall reduced , but the same trend was observed between fields , since median CSI in all fields were lower than in SRAF; only CSI levels in PAF were differentially affected , so that they were no longer higher than in primary fields ( CSI without baseline correction , [Q1 , median , Q3]: A1 , [0 . 14 , 0 . 24 , 0 . 40]; AAF , [0 . 18 , 0 . 30 , 0 . 45]; VAF , [0 . 21 , 0 . 32 , 0 . 42]; SRAF , [0 . 26 , 0 . 39 , 0 . 52]; PAF , [0 . 17 , 0 . 25 , 0 . 42] ) . However , given the higher spontaneous rate relative to evoked activity seen in PAF , uncorrected CSI does not faithfully represent the strong SSA ( i . e . , contrast ) clearly observed in responses from this field ( Fig 2B ) . Therefore , we kept using these corrected measures for the rest of the analyses . Consistent with previous studies [23] , nonprimary fields showed longer response onset latencies than primary fields for both deviant ( mean ± SEM: A1 , 11 . 6 ± 1 . 2 ms; AAF , 11 . 1 ± 1 . 1 ms; VAF , 17 . 3 ± 1 . 5 ms; SRAF , 27 . 0 ± 2 . 0 ms; PAF , 23 . 9 ± 2 . 2 ms; Kruskal-Wallis test , χ2 ( 4 ) = 152 . 78 , p < 10−31 ) and standard tones ( A1 , 16 . 7 ± 1 . 7 ms; AAF , 22 . 5 ± 3 . 3 ms; VAF , 29 . 8 ± 3 . 5 ms; SRAF , 45 . 8 ± 4 . 7 ms; PAF , 50 . 0 ± 8 . 0 ms; χ2 ( 4 ) = 77 . 59 , p < 10−15 ) . From these figures , it is apparent that onset latency was significantly delayed for standards as compared to deviants in all five fields ( onset latency difference , standard–deviant , mean ± SEM: A1 , 7 . 6 ± 1 . 5 ms; AAF , 13 . 6 ± 3 . 1 ms; VAF , 18 . 0 ± 3 . 1 ms; SRAF , 23 . 2 ± 3 . 8 ms; PAF , 31 . 8 ± 7 . 2 ms; all significantly greater than zero , Wilcoxon signed rank test , p < 0 . 01 in all cases ) . Thus , in addition to an overall reduction in spike counts , SSA also produced a delay in onset latency to the standard tones . Furthermore , this delay was significantly longer in nonprimary fields than in primary fields A1 and AAF ( Kruskal-Wallis test , χ2 ( 4 ) = 34 . 13 , p < 10−6 ) . The sharp differences in SSA levels observed between primary and nonprimary fields derive from a distinct topographic organization of adaptation throughout the whole auditory cortex ( Fig 4 ) . The absolute position of the map with respect to bregma differed between animals by up to 0 . 6 mm , but the relative position and orientation of the five cortical fields were highly conserved from one animal to the next . Thus , we constructed a synthetic map of CSI from all available data . Using the CF gradient as the main reference landmark , an appropriate shift was applied to each map to maximize the degree of CF coincidence between them ( Fig 4A; cf . Fig 1 in [23] and Fig 1 in [36] ) . We quantified the quality of the alignment as the local coincidence of CF values . The resulting correlation of CF between neighboring sites was next to maximal ( Topological product , PT = 0 . 9686 , permutation test , p < 0 . 001 ) [39] . Fig 4B shows the CSI map , while Fig 4C and 4D show the corresponding maps of the response to deviant and standard stimuli ( within the stimulus-fitted window ) , from which the CSI was computed . The CSI follows a statistically significant topographic distribution ( Topological product , PT = 0 . 2342 , permutation test , p < 0 . 001 ) , meaning that neighboring sites are likely to have more similar CSI values than more distant ones . To better determine the nature of this topography , we traced a boundary following the median iso-CSI contour ( Fig 4B; median population CSI = 0 . 60 ) whenever this line enclosed a region of area greater than 0 . 5 mm2 . This procedure revealed an emergent organization of SSA , showing a large region of low-to-medium CSI values that covers the central and rostral portions of the auditory cortex and two separate and distinct high-CSI regions confined to the posterodorsal and ventral margins of the map , respectively ( Fig 4B ) . Remarkably , the CSI-based boundary that defines the posterodorsal high-CSI region matches almost perfectly the boundary between A1 and PAF previously traced from the CF gradient reversal ( Fig 4A ) . Similarly , the iso-CSI contour that separates the ventral high-CSI region matches very well the caudal SRAF/VAF and rostral SRAF/AAF boundaries . Finally , these high-SSA regions revealed in Fig 4B can be seen also as regions of extremely low spike count to the standard stimuli in Fig 4D . Indeed , the “CSI” and “Standard” maps are almost complementary , such that regions of extreme CSI values correspond to those with virtually no response to standard stimuli , while regions of low-medium CSI match those with significant response to standards . This observation reveals a strong CSI dependence on the standard response being low , rather than on the deviant response being high . In fact , CSI was negatively correlated with both deviant ( DEV ) and standard ( STD ) response strength , yet much more strongly to the standard ( Spearman correlation coefficient , ρ[CSI , DEV] = −0 . 19 , p < 10−6; ρ[CSI , STD] = −0 . 81 , p < 10−152 ) . This also indicates that CSI values tend to be higher for neurons with an overall lower firing rate , as confirmed by a subsequent analysis ( v . i . ) . SSA was suggested as a potential neural correlate for the MMN , but previous studies neglected an analysis of the responses to deviant and standard tones at different temporal courses during stimulus presentation and beyond . Since we observed responses of long durations to deviant tones in many recordings ( deviant response offset , mean ± SEM: A1 , 162 . 6 ± 5 . 7 ms; AAF , 149 . 8 ± 6 . 9 ms; VAF , 194 . 2 ± 4 . 6 ms; SRAF , 196 . 4 ± 4 . 4 ms; PAF , 167 . 9 ± 7 . 4 ms ) , we wanted to further investigate the variation of the CSI across different components of the neural response . Hence , we computed baseline-corrected spike counts for different time intervals after stimulus onset ( Fig 5A ) : onset ( 5–30 ms ) , sustained ( 30–80 ms ) , offset ( 80–105 ms ) , and late ( 105–200 ms ) . Corresponding CSI distributions and their topography for these different time windows are shown in Fig 5B and 5C , respectively . First , we compared median CSI between fields for every time window separately . For the onset , sustained , and offset components , we found the same trend already observed for the stimulus-fitted response window: the median CSI in every primary field was significantly lower than in every nonprimary field , and lowest of all in A1 ( Fig 5B; Kruskall-Wallis test , onset: χ2 ( 4 ) = 73 . 95 , p < 10−14 , sustained: χ2 ( 4 ) = 109 . 81 , p < 10−22; offset: χ2 ( 4 ) = 60 . 95 , p < 10−11 ) . The CSI for the late component of the response , however , behaved differently . At this time window , there were no significant differences in SSA between fields ( Fig 5B; Kruskall-Wallis test , χ2 ( 4 ) = 7 . 78 , p > 0 . 1 ) . Then , we compared CSI levels within each field for the four time windows to analyze the trend of SSA throughout the different response components . Within nonprimary fields , we found no significant differences between median CSIs measured at the four different time windows ( Fig 5B; Friedman test , SRAF: χ2 ( 3 ) = 5 . 03 , p > 0 . 1; PAF: χ2 ( 3 ) = 4 . 72 , p > 0 . 1 ) . By contrast , a highly significant window effect was found for the three primary fields ( Friedman test , A1: χ2 ( 3 ) = 109 . 58 , p < 10−22; AAF: χ2 ( 3 ) = 18 . 18 , p < 0 . 001; VAF: χ2 ( 3 ) = 55 . 3 , p < 10−11 ) . Post-hoc comparisons revealed that this effect was due to a specific increase of CSI at the late component ( Fig 5B ) , with no significant differences between median CSI measured at the onset , sustained , or offset components of the response , except for a slightly significant increase from the sustained to the offset component in A1 , consistent with the overall trend . Therefore , SSA in the nonprimary fields is maintained high throughout the entire response ( Fig 5B and 5C ) . By contrast , SSA in the primary fields is moderate during stimulus presentation , followed by a specific enhancement in late components ( Fig 5B and 5C ) , in which SSA reaches the same levels found in nonprimary fields . Upon visual inspection , regions with lowest SSA in the CSI landscape seemed to coincide with low-CF regions of the auditory cortex , particularly within A1 ( Fig 4A and 4B ) . Since a strong dependence of SSA on frequency and intensity of pure-tone stimulation has been shown in the IC [15] , we wanted to test whether a similar dependence was present in the auditory cortex . Fig 6A shows a spotlight-average map of the SI across all frequency/intensity combinations tested in the whole set of recordings . High SSA is sharply skewed toward the high frequencies and low intensities of stimulation . When we analyzed primary and nonprimary fields separately ( Fig 6B and 6C ) , we observed that this dependence of the SI on frequency and intensity was more evident within primary ( Fig 6B ) than nonprimary fields ( Fig 6C ) . Additionally , average firing rate had a topographical distribution in the dataset and was different between cortical areas ( Fig 4C and 4D ) . Since firing rate may also have a strong impact on the amount of adaptation [17] , the topography of SSA could result in part from a topography of firing rates . Finally , the observed effect of stimulus intensity on the SI ( Fig 6 ) might be an indirect consequence of the effect of firing rate , with higher intensities of stimulation producing higher firing rates and , therefore , lower SSA . To address these observations quantitatively , we fit a multivariate linear regression model for the SI , following a stepwise strategy ( “fitlm” function in Matlab , with robust fitting options; sample data used to fit this model can be found in S5 Data ) . First , we used average spike count ( SPK , as the sum of average response to deviant and standard stimuli ) and frequency of stimulation ( OCT , in octaves with respect to 1 kHz ) as predictors . The resulting model was: SI= 0 . 51−0 . 046·SPK+0 . 057·OCT ( F2 , 1215 = 166 , p < 5×10−64 ) . This model accounted for 21 . 3% of the variability of the SI , but , more importantly , it provided a specific quantification of each effect: on average , SI decreases 0 . 046 points per spike of the response , while it increases 0 . 057 points per octave of the stimulus . Then , we added intensity of stimulation ( SPL , in dB SPL ) to the model , obtaining: SI=0 . 72−0 . 051·SPK+0 . 050·OCT−0 . 003·SPL ( F3 , 1214=122 , p<5×10−69 ) . Thus , SI is also negatively correlated to intensity of stimulation . This model , however , explained 23% of the variability of the SI , only 1 . 7% more than the previous one . Therefore , most of the dependence of the SI on SPL is already explained by its dependence on SPK , confirming the fact that higher intensities produce lower SSA because of a higher firing rate . Therefore , we removed SPL from the model and replaced it with FIELD as a categorical factor . Now , the explanatory power of the model increased to 30 . 6% , mainly due to overall higher SI in the nonprimary fields: SI=0 . 41+0 . 12·VAF+0 . 24·SRAF+0 . 20·PAF−0 . 04·SPK+0 . 05·OCT ( F6 , 1211=90 . 6 , p<10−94 ) . According to this model , mean SI is 0 . 41 in A1 and AAF ( not significantly different from each other ) , 0 . 53 ( 0 . 41 + 0 . 12 ) in VAF ( p < 5×10−9 ) , 0 . 65 in SRAF ( p < 5×10−28 ) , and 0 . 61 in PAF ( p < 5×10−16 ) , and this difference cannot be explained by differences in firing rate within fields , since the FIELD factor explains an extra 9 . 3% of the SI variability . Note also that these are mean values and , therefore , lower than the median values shown in Fig 3 , given the rightward skewness of the distributions . As a final step , we tested this model for interactions between FIELD and the other three predictors separately , and we found significant interactions only between FIELD and OCT: SI=0 . 19−0 . 24·VAF+0 . 36·SRAF+0 . 36·PAF+0 . 078·OCT−0 . 042·VAF·OCT−0 . 031·SRAF·OCT−0 . 034·PAF·OCT ( F9 , 1208=43 . 8 , p<5×10−68 ) , indicating that the effect of frequency was weaker in VAF ( p < 0 . 005 ) , SRAF ( p < 0 . 05 ) , and PAF ( p < 0 . 05 ) than in A1 and AAF . Therefore , the dependence of SSA on firing rate ( and , indirectly , on intensity of stimulation ) is comparable among the five fields , but the observed dependence of SSA on frequency of stimulation is mainly due to the fact that A1 and AAF show lower SSA for low frequencies of stimulation , as illustrated in Figs 4A , 4B and 6B . Incidentally , A1 and AAF are the cortical fields that show the most clear tonotopic gradient , each the mirror reversal of the other ( Fig 4A ) [23] . Since frequency and intensity of oddball stimulation were selected according to the frequency tuning and threshold of each recording site , and since there is a tendency for tuning bandwidth in auditory cortex to decrease as a function of CF [40 , 41] , differences in SSA between fields could simply reflect differences in tuning bandwidth or CF threshold in the auditory cortex . To check this possibility , we analyzed the correlation between CSI and frequency tuning characteristics in our sample . Distributions of tuning bandwidth and threshold in our sample were consistent with previous mapping work in the rat [23] . Particularly , PAF and AAF featured the broadest tuning bandwidth and highest response thresholds ( bandwidth 30 dB above threshold , in octaves , mean ± SEM: A1 , 1 . 89 ± 0 . 06; AAF , 2 . 30 ± 0 . 1; VAF , 1 . 75 ± 0 . 06; SRAF , 1 . 98 ± 0 . 08; PAF , 2 . 95 ± 0 . 16; CF threshold in dB SPL , mean ± SEM: A1 , 23 . 7 ± 0 . 9; AAF , 29 . 3 ± 1 . 3; VAF , 14 . 8 ± 0 . 9; SRAF , 22 . 5 ± 1 . 1; PAF , 28 . 3 ± 1 . 3 ) . Both bandwidth and threshold in AAF and PAF were different from the other fields , but not from each other ( Kruskal-Wallis test , bandwidth: χ2 ( 4 ) = 55 . 60 , p < 5×10−11; threshold: χ2 ( 4 ) = 96 . 03 , p < 10−20 ) . By contrast , CSI was 50% higher in PAF than in AAF , as already shown ( Fig 3B ) . Similarly , CF threshold in VAF was significantly lower than in A1 or AAF , but the median CSI was not different between these primary fields ( Fig 3B ) . Indeed , correlation between CSI and either tuning bandwidth or threshold was extremely weak in our sample ( Spearman correlation coefficient: ρ[CSI , BW30] = 0 . 083 , p = 0 . 04; ρ[CSI , THR] = −0 . 09 , p = 0 . 02 ) . These considerations demonstrate that the distinct topography of SSA that we have found is genuine and not an artifactual effect of differences in other response properties between cortical fields . In order to study the dynamics of adaptation to the repetitive stimuli over time , we averaged responses to standard and deviant stimuli across recordings for every trial number within the sequence and plotted them in relation to the time elapsed since the beginning of the sequence , separately for each field ( Fig 7A ) . Then , we fitted these responses to different simple models . None of the models tested could explain any amount of the variance of the deviant responses , indicating that deviant responses did not show dependence on trial number within any field . In sharp contrast , a power law model of three parameters , y ( t ) = a · tb + c , yielded very good quality fits for the responses to standards in all fields , explaining about 80% of their variability ( adjusted r2: A1 , 0 . 80; AAF , 0 . 74; VAF , 0 . 84; SRAF , 0 . 83; PAF , 0 . 69 ) and indicating that SSA in all fields matches stimulus statistics at many timescales [42] . The most obvious difference between fields was that nonprimary fields reached a much lower plateau at their final steady-state responses ( gray dashed line in Fig 7B; c parameter ( spk/trial ) : A1 , 0 . 84; AAF , 0 . 50; VAF , 0 . 60; SRAF , 0 . 22; PAF , 0 . 17; all significantly different from each other as derived from the 95% confidence intervals reported by the “fit” function in Matlab ) . Also , according to this model , adaptation was fastest in PAF , slowest in VAF , and not significantly different between the other three fields ( b parameter: A1 , –0 . 78; AAF , –0 . 93; VAF , –0 . 68; SRAF , –0 . 73; PAF , –1 . 32 ) . This result indicates a distinct high sensitivity of PAF to repetitive stimuli , needing only a few presentations to reach its fully adapted state . This phenomenon can be readily appreciated when analyzing the responses to the first 10 standard trials of the sequence ( Fig 7B ) . Responses to standards in the nonprimary fields adapt below half their initial strength with three ( PAF ) or four ( SRAF ) presentations of a stimulus ( black arrows in Fig 7B ) , whereas in the primary fields it takes up to six ( A1 ) presentations to produce this same relative reduction . Therefore , adaptation occurs faster and is stronger in nonprimary than in primary fields . Whereas SSA in spike responses is a local measure at the neuron level , the MMN is a large-scale brain potential . One reasonable way to bridge this gap is to probe the correlation between adaptation of neural responses and LFP , which represent average synaptic activity in local cortical circuits [43] . Thus , we recorded LFP simultaneously with MUA in four out of the 12 animals , with a total yield of 268 recording sites ( A1 , 49; AAF , 48; VAF , 55; SRAF , 54; PAF , 42; Unlocalized , 20 ) . We averaged the recorded LFP waveforms evoked by standard and deviant tones for each field separately and computed the difference wave ( DW ) at every time point after stimulus onset ( Fig 8A ) . In all five cortical fields , these potentials showed the typical morphology in response to pure tones [44 , 45] , with a fast negative deflection ( Nd ) followed by a slower positive deflection ( Pd ) . These two components were present in responses to both standard and deviant tones , but their amplitudes were , in all cases , smaller for the standards , giving rise to a DW of similar shape but varying amplitudes ( Fig 8A ) . For each recording , the peak amplitude and peak latency of the DW was measured for the Nd and Pd components , within a time window in which the DW reached statistical significance at the whole population level ( 16–37 . 6 ms for Nd and 41 . 5–86 . 7 ms for Pd , respectively , paired t test , Bonferroni correction for 268 comparisons , p < 0 . 05 ) . Peak amplitude of the DW at the Nd component showed a clear trend to be larger in primary than in nonprimary fields , being significantly smaller in PAF than in the three primary fields and smaller in SRAF than in AAF ( Fig 8B; one-way ANOVA , F4 , 243 = 8 . 24 , p < 5×10−6 ) . This trend was still present , albeit much less clear , for the Pd component of the DW , being significantly smaller in PAF than in A1 and AAF but not different between the other fields ( Fig 8B; one-way ANOVA , F4 , 243 = 3 . 74 , p < 0 . 01 ) . Thus , the fast Nd component of the DW showed a topographical distribution within the auditory cortex , whereas the slower Pd component of the DW showed a more homogenous distribution across cortical fields . A similar pattern was apparent for the peak latencies of each of these components ( Fig 8B ) . The Nd component of the DW peaked earlier in the primary than in the nonprimary fields , significantly so between A1 or AAF and SRAF or PAF ( mean ± SEM: A1: 24 . 6 ± 0 . 9 ms; AAF: 24 . 8 ± 0 . 8 ms; VAF: 28 . 3 ± 0 . 6 ms; SRAF: 31 . 1 ± 0 . 8 ms; PAF: 32 . 0 ± 1 . 7 ms; one-way ANOVA , F4 , 243 = 11 . 78 , p < 5×10−8 ) . Peak latencies for the Pd component , on the other hand , were not statistically different between fields ( mean ± SEM: A1: 61 . 7 ± 2 . 0 ms; AAF: 57 . 4 ± 2 . 2 ms; , VAF: 59 . 5 ± 2 . 0 ms; , SRAF: 59 . 8 ± 1 . 7 ms; PAF: 61 . 4 ± 2 . 1 ms; one-way ANOVA , F4 , 243 = 0 . 70 , p = 0 . 59 ) . The steady progression of the Nd peak latency is consistent with a bottom-up propagation of the signal from primary to nonprimary fields , whereas the homogeneity of the Pd peak latency suggests a stronger contribution of intracortical processing and reciprocal interaction between fields . To facilitate a more direct comparison between SSA for the MUA and for the LFP components , we also computed CSI values for the Nd and Pd peaks of the LFP ( S8 Data ) . Overall , SSA at both components of the LFP was appreciably lower than for the MUA ( paired signed rank test for the whole set of recordings with LFP; CSI-Nd versus CSI-onset , z-score = 6 . 98 , p < 5×10−12; CSI-Pd versus CSI-sustained , z-score = 10 . 12 , p < 5×10−24 ) , but it followed the same trend to be lower in the primary than in nonprimary fields ( Median CSI-Nd: A1 , 0 . 32; AAF , 0 . 31; VAF , 0 . 45; SRAF , 0 . 50; PAF , 0 . 47; Kruskall-Wallis test , χ2 ( 4 ) = 21 . 12 , p < 5×10−4 . Median CSI-Pd: A1 , 0 . 25; AAF , 0 . 24; VAF , 0 . 33; SRAF , 0 . 37; PAF , 0 . 40; Kruskall-Wallis test , χ2 ( 4 ) = 13 . 09 , p < 0 . 05 ) . Furthermore , CSI-Nd and CSI-Pd were strongly correlated with their corresponding CSI values at comparable time windows ( Spearman correlation coefficient: ρ[CSI-Nd , CSI-onset] = 0 . 66 , p < 10−40; ρ[CSI-Pd , CSI-sustained] = 0 . 43 , p < 5×10−12; ρ[CSI-Pd , CSI-offset] = 0 . 21 , p < 0 . 005 ) .
In this account , we compared the level of SSA in primary and higher-order auditory cortex to validate SSA as a candidate neural correlate of the MMN . To study the topographic organization of SSA , we mapped the whole rat auditory cortex with MUA recordings from middle layers IIIb/IV using an oddball paradigm . We demonstrate that SSA occurs beyond A1 , and its properties differ between primary and nonprimary fields . Our major findings are: ( 1 ) Highest SSA is sharply segregated to nonprimary fields , creating a distinct topographic gradient of SSA within the auditory cortex . ( 2 ) High SSA is present in nonprimary fields up to 200 ms after stimulus onset , and it remains stronger than in primary fields during the first 100 ms of the neuronal responses . ( 3 ) In all cortical fields , SSA is correlated in time and strength with the difference wave seen in both the fast ( Nd ) and slower ( Pd ) deflections of the LFP . As additional novel findings , we show that ( 4 ) SSA produces a delay in the responses to standard tones , as compared to deviants , and this delay is longer in nonprimary fields . ( 5 ) SSA is significantly higher for high frequencies of stimulation , and this dependence is more pronounced in primary fields . ( 6 ) SSA occurs faster and reaches a much lower plateau in the nonprimary fields . One key aspect of our data is the high coincidence in the relative position of the fields across animals and in comparison with previous mapping studies [23 , 24 , 32 , 36] . Our analysis revealed a systematic meta-organization of SSA in the auditory cortex of the rat [23 , 36] , such that the CSI gradient shows a steep increase at the boundaries between primary and nonprimary fields ( Fig 4B ) . In particular , the sharp CSI enhancement between A1 and PAF ( Fig 4D ) bears striking resemblance with the same border found previously for bandwidth and latency [24] . Our results conform with previous studies that showed SSA in A1 [12 , 19 , 44–50] and extend their findings , as we present new SSA properties hitherto unknown . Importantly , the distribution of SSA indices in our A1 sample is largely equivalent to those shown in previous studies of SSA in the rat or mouse A1 that used similar paradigm parameters [19 , 47 , 50] , making further comparisons more reliable . To the best of our knowledge , there were no previous studies of SSA outside A1 , although higher SSA levels were expected to be found in nonprimary fields , since neurons in nonprimary cortical areas are known to show fast adaptation [20 , 21] . In particular , many studies independently reported that PAF neurons in the rat adapt strongly even to slow repetition rates [22–24] , and novel sounds produced greater cellular activity than familiar sounds in auditory association cortex in area Te3 [51] , where the SRAF is located [35] . There is also strong evidence of enhanced adaptation in nonprimary areas of the auditory cortex from large-scale brain responses ( ERP , magnetoencephalography [MEG] , fMRI ) in both animals [28 , 29 , 52 , 53] and humans [25–27 , 54] . Our findings also parallel the topography of subcortical SSA ( Fig 9 ) . Previous studies consistently found stronger SSA in the nonprimary ( or nonlemniscal ) subdivisions of the IC [14–16] and MGB [17 , 55] . Importantly , an identical dependence of SSA on frequency of stimulation as well as a delay in onset latency of responses to standards have already been shown in the IC [15] . Our data sharply contrast with previous studies showing that the SSA level in A1 neurons is independent of their CF and in which less than 4% of neurons showed a latency effect [56] . However , the presence of strong SSA in spiking responses at 50–100 ms and beyond represents the major difference with previous SSA studies . Only very recently , two studies in mouse auditory cortex [49 , 50] and one in rat somatosensory cortex [57] found SSA in either subthreshold Vm fluctuations of layer II/III pyramidal neurons [49] or spiking responses of inhibitory interneurons [49 , 50] and layer IV pyramidal neurons [57] occurring more than 50–100 ms after stimulus onset . Importantly , we recorded mainly form layer IIIb/IV neurons , receiving direct thalamocortical inputs , which are more likely to show long-latency spiking responses [58] . Finally , previous studies reported SSA for LFP in A1 , but they failed to show any correlation between MMN-like components of the LFP and SSA . Some did not find significant spiking activity for latencies beyond 50 ms [44 , 45] or observed SSA only for the fast Nd [46]; others did not measure MUA [59] , or their analysis was restricted to the fast Nd only [19] . Such a correlation has only been described in the somatosensory cortex [57] . The mechanisms and location of the neural generators of SSA and their relation to MMN are still subjects of debate [11 , 13 , 60 , 61] . In the lemniscal pathway ( Fig 9 ) , SSA undergoes a first enhancement at the thalamocortical synapses from the ventral division of the MGB to A1 [12 , 17] . Here , we show a further enhancement of SSA in nonprimary cortical fields , which integrate the thalamocortical projection from nonlemniscal MGB [31] and the corticocortical projection from primary fields [62] and redirect their output to prefrontal and limbic brain regions involved in spatial attention and emotional memory [34 , 35] . Thus , our study confirms that SSA is a prevalent property of the nonlemniscal auditory pathway , even at the cortical level ( Fig 9 ) . This organization may underlie its functional significance as a higher-order stage of sensory processing beyond the faithful representation of the auditory stimuli that predominates in the lemniscal pathway [63] . Cumulating evidence indicates the existence of a hierarchy of processing stages for regularity encoding in the auditory brain , with later response components showing sensitivity for changes in more complex aspects of the acoustic scene [13 , 60 , 64] . Repetition positivity ( RP ) has been proposed as the electrophysiological correlate of the memory trace formation required for subsequent change detection and , in turn , rapid SSA in auditory cortex is likely to contribute to its generation [65 , 66] . Here , we show very strong SSA in nonprimary auditory cortex , supposed to contain the main generators of the MMN in humans [25 , 27 , 54 , 67 , 68] , cats [53] , and rats [29] , that resembles MMN in several ways . First , SSA results in stronger responses to deviants than to standards in the oddball paradigm , to the extent that responses to standards can get totally suppressed in some recordings from nonprimary fields . Critically , we show strong SSA in these areas between 50 and 100 ms , correlated with a consistent difference wave at the slow Pd component of the LFP ( Fig 8A ) . The latency of this Pd deflection ( 60–80 ms ) is considerably shorter than the human MMN ( 150–200 ms ) but matches perfectly the range of MMN-like potentials in the rat [28 , 29 , 69–73] , which tend to occur , on average , 50–100 ms after stimulus onset , probably due to the smaller size of the rat brain [37] . Interestingly , this SSA resembles RP in the first standard presentations ( Fig 7B ) and matches stimulus statistics at multiple time scales [56 , 74] . We also show stronger SSA for high- than for low-frequency tones , paralleling a commonly observed effect of frequency in both animal [71–73] and human [75 , 76] MMN recordings . Therefore , we present strong evidence linking animal SSA to the human MMN , a result thus far missing in animal research . Importantly , we show that an MMN-like difference signal can readily result from SSA to standard tones that leaves responses to deviants unaffected ( Fig 7A ) . Additionally , our LFP recordings show that the same components were present in responses to both standard and deviant tones ( Fig 8A ) , consistent with the view that the MMN is a differentially adapted obligatory component of the ERPs . If so , our results would suggest a purely SSA explanation for the MMN [6 , 7 , 26] . Before we conclude , we should draw attention to three major caveats of our study . First , anesthesia reduces neuronal responsiveness to auditory stimuli as well as spontaneous firing , and may change some receptive field properties [77–79]; thus , an increased sensitivity to anesthetics in higher-order fields may lead to an overestimation of the SSA seen in those areas . However , we observed high spontaneous rates as well as strong , sustained responses to deviants in nonprimary fields ( Fig 5A; baseline-corrected spike counts within 0 and 200 ms , mean ± SEM: A1 , 3 . 2 ± 0 . 1; AAF , 2 . 8 ± 0 . 1; VAF , 4 . 7 ± 0 . 2; SRAF , 3 . 9 ± 0 . 2; PAF , 2 . 6 ± 0 . 2 ) . We used urethane as anesthetic because it preserves balanced neural activity better than other agents [80] , retains the higher-order processing capabilities of the auditory cortex [81] , and shows no significant effects on SSA levels , at least in the IC [82] . Most importantly , MMN-like responses have been successfully recorded from anesthetized [29 , 69–71] and awake [28 , 72 , 73] animals alike ( for review , see [10] ) . Second , the MMN is a negative-going component , in contrast to the positive late potential ( Pd ) examined here . Depending on the location of recording and anesthetic state , epidural MMN recordings in rats can be positive in polarity [72 , 73] , an effect commonly observed in urethane-anesthetized preparations [69 , 71] . Moreover , an inversion of the LFP has been extensively described using laminar probes in A1 [45 , 59] , such that positivities in layers IIIb/IV may appear as negativities in superficial layers . Third , there are some discrepancies between the SSA seen in MUA and in LFP data . Namely , whereas the MUA shows prominent activity between 100 and 200 ms ( i . e . , beyond the rat-MMN range ) , the LFP is relatively flat within this time window . Similar late-spiking activity has been observed in parvalbumin-positive inhibitory interneurons [49] and interpreted as delayed reverberating network activity specifically triggered by deviant stimuli , but we cannot rule out that MUA includes activity from thalamocortical afferents in layers IIIb/IV , which would not produce a prominent LFP component . Alternatively , the late enhancement of SSA ( 100–200 ms ) seen in the primary fields ( Fig 5B and 5C ) might result from processing in the nonprimary fields , subsequently transmitted downwards through the massive feedback corticocortical connections ( Fig 9 ) [34 , 35 , 57 , 83] . A more relevant discrepancy is that the difference-wave amplitude for the later Pd component of the LFP is comparable between primary and nonprimary auditory cortex and even significantly smaller in PAF than in A1 or AAF ( Fig 8B ) , not supporting the notion of enhanced SSA in nonprimary fields . However , previous ERP studies [28 , 29] failed to find differences in the MMN amplitude between primary and nonprimary fields . One simple reason for this could be that ERPs and LFPs are large-scale potentials , reflecting overall synaptic activity within a wide volume of tissue [43] , most probably spanning the boundaries between fields . Therefore , local measures at the cellular level , such as MUA , are much better indicators of specific differences between fields . Furthermore , it is consistent to find higher SSA at the MUA than at the LFP level ( i . e . , output versus input , respectively ) within any particular area , as also shown at the single-neuron level [48] . Additionally , the amplitude of the difference wave is an absolute measure , whereas SSA is commonly expressed as a contrast , such as the CSI . When computed this way , SSA for the Pd amplitude is already higher in nonprimary than in primary fields , yet this difference is much sharper for the MUA , reflecting the operations carried out by nonprimary fields to their already-adapted inputs . At this juncture , it is important to note that the slower Pd component of the difference wave peaked with the same latency throughout the entire auditory cortex ( Fig 8B ) , and so did its epidural counterpart in the rat [29] . By contrast , the fast Nd deflection of the LFP occurs earlier in primary than in nonprimary fields ( Fig 8B ) , suggesting a lemniscal origin and bottom-up propagation . Therefore , the higher degree of reciprocal interaction between fields is likely involved in the generation of the Pd , consistent with the idea that intracortical processing contributes to SSA at longer latencies [12 , 50 , 57 , 59 , 84] . Thus , MMN-like potentials may be readily recorded from both primary and nonprimary auditory cortex , but nonprimary fields seem to contribute critically to their generation at the microcircuit level [27 , 85] . In conclusion , we demonstrate that strong SSA occurs in nonprimary auditory cortex at the latency range of the MMN in the rat . This finding overcomes the two main discrepancies hitherto alleged against the suggestion that SSA in the auditory cortex may underlie the generation of the MMN [7 , 86] , namely , its anatomical location and its temporal development . We provide empirical evidence of the missing link between SSA in single neurons and scalp-recorded potentials , thus bridging the gap between animal physiology studies and the human MMN . Given the wide use of the MMN as a tool in clinical and cognitive neuroscience [9 , 10 , 87 , 88] , such a connection is potentially of high relevance for future research in these fields .
The experimental protocols were approved by , and used methods conforming to the standards of , the University of Salamanca Animal Care Committee and the European Union ( Directive 2010/63/EU ) for the use of animals in neuroscience research . Experiments were performed on 12 adult female Long-Evans rats with body weights within 200 and 250 g . Surgical anesthesia was induced and maintained with urethane ( 1 . 5 g/kg , i . p . ) , with supplementary doses ( 0 . 5 g/kg , i . p . ) given as needed . Dexamethasone ( 0 . 25 mg/kg ) and atropine sulfate ( 0 . 1 mg/kg ) were administered at the beginning of the surgery and every 10 h thereafter to reduce brain edema and the viscosity of bronchial secretions , respectively . Prior to surgery and recording sessions , we recorded auditory brainstem responses ( ABR ) with subcutaneous electrodes to ensure the animal had normal hearing . ABRs were collected using Tucker-Davis Technologies ( TDT ) software ( BioSig ) and hardware ( RX6 Multifunction Processor ) following standard procedures ( 0 . 1 ms clicks presented at a 21/s rate , delivered in 10 dB ascending steps from 10 to 90 dB SPL ) . The animal was then placed in a stereotaxic frame in which the ear bars were replaced by hollow specula that accommodated a sound delivery system . After the animal reached a surgical plane of anesthesia , the trachea was cannulated for artificial ventilation and a cisternal drain was introduced to prevent brain hernia . Corneal and hind-paw withdrawal reflexes were monitored to ensure that a moderately deep anesthetic plane was maintained as uniformly as possible throughout the recording procedure . Isotonic glucosaline solution was administered periodically ( 5–10 ml every 6–8 h , s . c . ) throughout the experiment to prevent dehydration . Body temperature was monitored with a rectal probe and maintained between 37 and 38°C with a homoeothermic blanket system ( Cibertec ) . The skin and temporal muscles over the left side of the skull were reflected , and a 6 × 5 mm craniotomy was made in the left temporal bone to expose the entire auditory cortex . The dura was removed and the exposed cortex and surrounding area were covered with a thin , transparent layer of agar to prevent desiccation and to stabilize the recordings . At the end of the surgery , a magnified picture ( 25× ) of the exposed cortex was taken with a digital SLR camera ( D5100 , Nikon ) coupled to the surgical microscope ( Zeiss ) through a lens adapter ( TTI Medical ) . The picture included a pair of reference points previously marked on the dorsal ridge of the temporal bone , indicating the absolute scale and position of the image with respect to the bregma . This picture was displayed on a computer screen , and a micrometric grid was overlapped to guide and mark the placement of the electrode for every recording made . Recording sites ( 150–250 μm spacing; Fig 1A ) were evenly distributed across the cortical region of interest while avoiding blood vessels . The vascular pattern was used as a local reference to mark the position of every recording site in the picture , but otherwise differed largely between animals . To confirm the actual depth and cortical layer of the recorded neurons , at the end of the experiment we made electrolytic lesions at one to four of the recording sites at the same depth that recordings were made . Experiments were performed inside a sound-insulated and electrically shielded chamber . Sounds were generated using an RX6 Multifunction Processor ( TDT ) and delivered monaurally ( to the right ear ) in a closed system through a Beyer DT-770 earphone ( 0 . 1–45 kHz ) fitted with a custom-made cone and coupled to a small tube ( 12-gauge hypodermic ) sealed in the ear . The sound system response was flattened with a finite impulse response ( FIR ) filter , and the output of the system was calibrated in situ using a ¼-in condenser microphone ( model 4136 , Brüel & Kjær ) , a conditioning amplifier ( Nexus , Brüel & Kjær ) , and a dynamic signal analyzer ( Photon+ , Brüel & Kjær ) . The output of the system had a flat spectrum at 76 dB SPL ( ±3 dB ) between 500 Hz and 45 kHz , and the second and third harmonic components in the signal were ≤ 40 dB below the level of the fundamental at the highest output level ( 90 dB SPL ) [14] . MUA was recorded with self-manufactured glass-coated tungsten electrodes ( 1–5 MΩ impedance at 1 kHz ) [89 , 90] . A single electrode was positioned orthogonal to the pial surface ( forming a 30° angle with the horizontal plane ) and advanced 350–550 μm into the thalamorecipient layers IIIb–IV using a piezoelectric micromanipulator ( Sensapex ) until we observed a strong spiking activity synchronized with the train of searching stimuli . The signal was amplified ( 1000× ) and band-pass filtered ( 1 Hz to 3 kHz ) with a differential amplifier ( DAM-80 , WPI ) . This analog signal was digitized at a 12K sampling rate and further amplified and band-pass filtered for action potentials ( between 500 Hz and 3 kHz ) . Spike waveforms and relative times in respect to the start of the recording were displayed and stored in a PC running Windows XP ( Microsoft ) . A bilateral threshold for automatic action potential detection was set at about two to three standard deviations of the background noise . In a subset of the experiments , the digital signal was further filtered for LFP ( between 3 and 50 Hz ) , decimated to a 508 Hz sampling rate and stored in continuous form for posterior analysis . Stimulus generation and neuronal response visualization were controlled online with custom software created with the OpenEx suite ( TDT ) and Matlab ( Mathworks ) . Sounds used for stimulation were white noise bursts or pure tones with 5 ms rise-fall ramps . Sounds used for searching for neuronal activity were trains of noise bursts or pure tones ( 1–8 stimulus per second ) . We used short stimulus duration for searching ( 30 ms ) to prevent strong adaptation . In addition , type ( noise , pure tone ) and parameters ( frequency , intensity , presentation rate ) of the search stimuli were varied manually when necessary to facilitate release from adaptation and , thus , prevent overlooking responses with high SSA . Once a suitable recording site was reached , the FRA was determined using 75 ms pure tones at varying frequencies and intensities ( Fig 2A; 0 . 5–44 kHz logarithmically spaced at 0 . 25 octave steps , 0–70 dB SPL at 10 dB steps , 375 ms onset-to-onset interval , one to three randomized repetitions of each stimulus ) . The FRA was displayed on a computer screen using custom software , and the frequency-tuning curve was automatically outlined as the minimum sound intensity that elicited a firing rate over 20%–40% of the maximum firing for each frequency . Thus , the minimum response threshold and CF were computed for each site ( excluding isolated “islands” of spontaneous activity ) , and two frequencies ( f1 , f2 ) were selected to use in the oddball paradigm [12] at 20–30 dB above threshold . The two stimuli were selected so as to evoke strong responses of similar magnitude at that recording site . In some cases , one or more extra pairs of stimuli were selected to ensure at least one recording met this requirement . Two oddball sequences with fixed parameters ( 250 trials each , 75 ms stimulus duration , 0 . 5 octaves frequency separation , 10% deviant probability , 300 ms onset-to-onset interval , minimum of three standard tones before a deviant ) were presented for every pair of stimuli thus selected . In one of the sequences , the low frequency ( f1 ) was the “standard” and the high frequency ( f2 ) was the “deviant , ” and in the other sequence their roles were swapped . The order of presentation of these two sequences was randomized across sites . Peristimulus time histograms ( PSTH ) were generated for every stimulus and condition tested . Only the last standard tones preceding each deviant were used for the analyses , except for the time course analysis , where all standard trials were analyzed . Every PSTH was analyzed to test for significant auditory responses and to extract several different metrics of response strength and latency . For these analyses , the original PSTH was smoothed with a 6 ms gaussian kernel ( “ksdensity” function in Matlab ) in 1 ms steps to estimate the spike-density function ( SDF ) over time , and the baseline spontaneous firing rate ( SFR ) was determined as the average firing rate during the 75 ms preceding stimulus onset . For any given time window , the excitatory response was measured as the area below the SDF and above the baseline SFR . This measure will be referred to as “baseline-corrected spike count” ( BCSC ) . To test for statistical significance of the BCSC we used a Monte Carlo approach . First , 1000 simulated PSTHs were generated using a Poisson model with a constant firing rate equal to the SFR . Then , a “null distribution” of BCSC was generated from this collection of PSTHs , following these same steps . Finally , the p-value of the original BCSC was empirically computed as p = ( g + 1 ) / ( N + 1 ) , where g is the count of “null” measures greater than or equal to BCSC and N = 1000 is the size of the “null” sample . Note that using this approach , the minimum p-value that can be obtained is 1/1001 ≈ 0 . 001 . When a significant evoked activity was detected , onset and offset latencies of the whole excitatory response were computed as follows . First , a “noise” threshold was computed , as the firing rate below which the pure-spontaneous simulated SDFs remained 97 . 5% of the time . Every SDF , including the simulated ones , was scanned for stretches of “signal” above this threshold , and the amount of “signal” for each stretch was measured as the area below the SDF and above the SFR during that particular interval . Using the distribution of all the signal stretches thus found within the 1 , 000 pure-spontaneous SDFs , a Monte Carlo test was used to compute empirical p-values for every stretch of signal found in the target SDF under study . For each significant signal stretch ( p < 0 . 05 ) , the start/end times ( Ton , Toff ) were determined as the time points when the SDF trace cuts the noise threshold , and onset/offset latencies of the whole excitatory response ( ONSET , OFFSET ) were defined as the Ton/Toff of the first/last significant excitatory component of the response , respectively . Peak firing rate amplitude was defined as the maximum firing rate reached by the SDF within the analysis window , minus the SFR baseline , and peak latency as the time point respect stimulus onset that this peak takes place . Finally , the duration of the whole significant response interval was defined as OFFSET–ONSET , and the duration of the strong peak of the response , or “half-peak response duration , ” was measured as the total length of time that the SDF remains above 50% of the peak amplitude . In order to quantify and compare SSA levels between the five fields , we computed the frequency-specific SSA index for each stimulus , SI ( f1 ) and SI ( f2 ) , and the common SSA index ( CSI ) for every recording site in the usual way [12]: SI ( fi ) =DEV ( fi ) −STD ( fi ) DEV ( fi ) +STD ( fi ) ;i=1 , 2 CSI=∑DEV ( fi ) −∑STD ( fi ) ∑DEV ( fi ) +∑STD ( fi ) ;i=1 , 2 Where DEV ( fi ) , STD ( fi ) are baseline-corrected spike counts in response to frequency fi when it was a deviant and standard , respectively . The CSI was calculated only for recordings with significant auditory responses to at least one frequency in the oddball paradigm ( either as deviant or as standard ) . In cases in which more than one stimulus pair was tested at the same recording site , we selected only one to compute SSA for that site , according to the following criteria: ( 1 ) Recordings with significant responses to both frequencies ( either as deviant or as standard ) were always preferred to recordings with significant response to only one of them . ( 2 ) We selected the recording with most similar responses to f1 and f2 ( as deviants ) ; the similarity between responses was measured as their ratio , f1/f2 or f2/f1 , whichever was less than 1 . ( 3 ) If there were two or more recordings with similar deviant-to-deviant ratios ( difference of ratios < 0 . 1 ) , we selected the one with the lowest sound level ( SPL ) used for stimulation . For the analysis of the LFP signal , we aligned the recorded wave to the onset of the stimulus for every trial and computed the mean LFP for every recording site and stimulus condition ( deviant , standard ) as well as the difference wave ( DW = deviant−standard ) . Then , grand-averages were computed for deviant , standard , and DW across the whole auditory cortex and for every field separately . The p-value of the grand-averaged DW was determined for every time point with a two-tailed t test , Bonferroni-corrected for 204 comparisons ( overall significance level of 0 . 05 ) , and the time intervals in which a significant DW was observed were computed . For each individual ( mean ) LFP wave , the peak amplitude and latency were computed within two time windows: [10–40 ms] and [50–90 ms] , corresponding to the first Nd and second Pd seen in the grand-averages within all fields . When comparing response features between fields , such as onset latency or CSI , we used nonparametric Kruskall-Wallis or Friedman tests , given the non-normal nature of these measures . Each of these tests was followed by a post-hoc multiple comparison test , using the Dunn-Sidak method at a 5% significance level , to detect specific differences between fields . For the sake of readability , p-values for all tests are reported using an upper bound equal to the minimum power of ten or half a power of ten that is greater than the actual p-value ( e . g . , p < 5·10−6 ) . For the time course analysis , we first computed the average standard and deviant response at each absolute position within the sequence for all neurons tested within each cortical field separately . A single-trial spike count for any given PSTH was computed as the number of spikes between the previously determined ONSET and OFFSET times , minus the baseline SFR . Then , we fitted these time series to different models ( linear , exponential , double exponential , polynomial inverse , and power law with two or three coefficients ) using the “fit” function in Matlab , which also computes the coefficient of determination ( adjusted-r2 ) of the whole fit and confidence intervals for the fitted parameters . To quantify the topographical organization of a feature map and test for statistical significance thereof , we used the “MapTools” library in Matlab , applying the topographic product statistic [39] . This metric was used instead of other alternatives ( Pearson and Spearman linear correlation , Zrehen measure , etc . ) due to the highly non-normal nature of the data under study ( i . e . , CSI ) and assuming a local , linear nature of the topography of the CSI . To generate averaged maps for CF , CSI , and other response features , we followed a spotlight-average approach: starting with the set of sample points in which actual recordings were made and the associated values of the feature , we computed the averaged feature value for any other point in the map from its nearest neighbors . Specifically , we placed a bivariate Gaussian kernel of 100-μm radius , ker ( x , y ) =12πr⋅exp{x2+y22r2} , centered on every sample point and multiplied it by its associated feature value . Then we summed all these functions over the entire map and divided the result by the sum of all kernels at every point , to compute a weighted average throughout the whole surface . Thus , the feature value V at every point of the map was calculated as: V ( x , y ) =Σi=0nvi⋅ker ( x−xi , y−yi ) Σi=0n ker ( x−xi , y−yi ) , where x , y are the coordinates of a generic point in the map , and xi , yi ( i = 1 , … , n ) are the sample points used to generate the map . To impose a limit on the influence span for every point , this weighted average was computed only for points where the sum of all kernels ( denominator in the last formula ) was greater than 0 . 05 . Further , to avoid single-point averages , we computed V ( x , y ) only when at least two neighboring sample points had been used for averaging . To combine data from different animals , we followed an iterative process to improve the quality of the alignment in successive stages . We first generated the CF map for the case with the greatest number of recordings ( shown in Fig 1 ) . Then , we applied a manual shift to each of the remaining maps in turn so as to put them into register with the former . We used the CF gradient , the “unresponsive spot” at the wedge between A1 , AAF , and VAF , and the low-frequency centers in A1 , AAF , and SRAF as main references to determine , for each animal , the absolute position of the map with respect to the bregma [23] . Finally , we computed the topographic product statistic for the whole set of aligned recordings . This alignment was refined and the test statistic was recalculated until no improvement was detected in the correlation . We repeated this process for every animal until the alignment was completed .
|
Sensory systems automatically detect salient events in a monotonous ambient background . In humans , this change detection process is indexed by the mismatch negativity ( MMN ) , a mid-late component of the auditory-evoked potentials that has become a central tool in cognitive and clinical neuroscience over the last 40 years . However , the neuronal correlate of MMN remains controversial . Stimulus-specific adaptation ( SSA ) is a special type of adaptation recorded at the neuronal level in the auditory pathway . Attenuating the response only to repetitive , background stimuli is a very efficient mechanism to enhance the saliency of any upcoming deviant or novel stimulus . Thus , SSA was originally proposed as a neural correlate of the MMN , but previous studies in the auditory cortex reported SSA only at very early latencies ( circa 20–30 ms ) and only within the primary auditory cortex ( A1 ) , whereas MMN analogs in the rat occur later , between 50 and 100 ms after change onset , and are generated mainly within nonprimary fields . Here , we report very strong SSA in nonprimary fields within the latency range of the MMN in the rat , providing empirical evidence of the missing link between single neuron response studies in animal models and the human MMN .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"auditory",
"cortex",
"medicine",
"and",
"health",
"sciences",
"action",
"potentials",
"engineering",
"and",
"technology",
"signal",
"processing",
"topographic",
"maps",
"membrane",
"potential",
"brain",
"brain",
"electrophysiology",
"electrophysiology",
"neuroscience",
"signal",
"filtering",
"brain",
"mapping",
"computational",
"neuroscience",
"bioassays",
"and",
"physiological",
"analysis",
"event-related",
"potentials",
"bandwidth",
"(signal",
"processing)",
"electroencephalography",
"neuroimaging",
"research",
"and",
"analysis",
"methods",
"sensory",
"physiology",
"geography",
"imaging",
"techniques",
"animal",
"cells",
"clinical",
"neurophysiology",
"cartography",
"electrophysiological",
"techniques",
"auditory",
"system",
"diagnostic",
"medicine",
"anatomy",
"cell",
"biology",
"physiology",
"neurons",
"earth",
"sciences",
"single",
"neuron",
"function",
"biology",
"and",
"life",
"sciences",
"sensory",
"systems",
"cellular",
"types",
"computational",
"biology",
"neurophysiology"
] |
2016
|
Topographic Distribution of Stimulus-Specific Adaptation across Auditory Cortical Fields in the Anesthetized Rat
|
Collective motion phenomena in large groups of social organisms have long fascinated the observer , especially in cases , such as bird flocks or fish schools , where large-scale highly coordinated actions emerge in the absence of obvious leaders . However , the mechanisms involved in this self-organized behavior are still poorly understood , because the individual-level interactions underlying them remain elusive . Here , we demonstrate the power of a bottom-up methodology to build models for animal group motion from data gathered at the individual scale . Using video tracks of fish shoal in a tank , we show how a careful , incremental analysis at the local scale allows for the determination of the stimulus/response function governing an individual's moving decisions . We find in particular that both positional and orientational effects are present , act upon the fish turning speed , and depend on the swimming speed , yielding a novel schooling model whose parameters are all estimated from data . Our approach also leads to identify a density-dependent effect that results in a behavioral change for the largest groups considered . This suggests that , in confined environment , the behavioral state of fish and their reaction patterns change with group size . We debate the applicability , beyond the particular case studied here , of this novel framework for deciphering interactions in moving animal groups .
Collective motion occurs across a variety of scales in nature , offering a wealth of fascinating phenomena which have attracted a lot of attention [1]–[5] . The self-organized motion of social animals is particularly intriguing because the behavioral rules the individuals actually follow and from which these remarkable collective phenomena emerge often remain largely unknown due to the tremendous difficulties to collect quality field data and/or perform controlled experiments in the laboratory . This situation does not prevent a thriving modeling activity , thanks to the relative ease by which numerical simulations can be conducted . However , most models of moving animal groups are built from general considerations , educated guesses following qualitative observations , or ideas developed along purely theoretical lines of thought [6]–[9] . Even when authors strive to build a model from data , as in the recent paper by Lukeman et al . [10] , this model building amounts to writing down a fairly complicated structure a priori , involving many implicit assumptions , and to fit collective data to determine effective parameters , yielding a best-fit model . On the other hand , recent studies within the physics community of simple , minimal models for collective motion have revealed an emerging picture of universality classes [11]–[15]: Take , for instance , the Vicsek model , arguably one of the simplest models exhibiting collective motion . In this model , point particles move at constant speed and choose , at discrete time-steps , their new heading to be the average of that of their neighbors located within unit distance . Many of these behavioral restrictions can be relaxed without changing the emerging collective properties . Fluctuations of speed can be allowed , some short-range repulsion ( conferring a finite size to the particles ) can be added , even explicit alignment can be replaced by inelastic collisions , etc . , all these changes will still produce the remarkable nonlinear high-density high-order bands emerging near onset of collective motion , and , deeper in the ordered moving phase , the anomalously strong number fluctuations which have become a landmark of the collective motion of polarly aligning self-propelled particles [16]–[20] . The Vicsek model , in this context , is one of the simplest members of a large universality class defined by all models sharing the same large-scale properties . This universality class can be embodied in the continuous field equations that physicists are now able to derive . With such a viewpoint , different models in this class merely differ in the numerical values of their parameters [21]–[23] , very much like different fluids are commonly described by the Navier-Stokes equations and differ only in their viscosity and other constitutive parameters . Significant features nevertheless may be altered when a qualitatively important feature is changed , such as the symmetry of the aligning interaction , or added , as when local attraction/repulsion between individuals is also considered [8] , [24] In this latter case , for instance , no strong clustering and high density band appears when attraction is sufficiently strong , and finite groups may keep cohesion in open space as most natural groups do . These models yield a more complex phase diagram where collectively moving groups may assume gas-like , liquid-like or even moving crystal states as the two parameters controlling alignment and cohesion are varied . So , it remains important to know how individuals make behavioral choices when interacting with others , not only from a social ethology and cognitive viewpoint , but also because i ) different behavioral rules may make a difference in small enough groups and ii ) the analysis of local-scale data that this requires may lead to discover features eventually found to give rise to different qualitative collective properties . A recent instance can be found in the results on the structure of starling flocks gathered by Ballerini et al . [25]: They have ignited an ongoing debate about the possibility that individuals might interact mostly with neighbors determined by topological rules and not by metric criteria as assumed in most models . While this message has intrinsic value for the study of decision-making processes in animal groups , it was also shown recently that such metric-free , topological interactions are relevant , in the sense that they give rise to collective properties that are qualitatively different from those of metric models [26] . Thus , in this case , an individual-level ingredient suggested by data , which had been only partially and theoretically considered before [6] , [7] , [27] , defines new classes of collective properties . Given that animals are likely to possess more sophisticated behavior than , say , sub-cellular filaments displaced by molecular motors , one can expect more hidden features to play an important role at the collective level . This is a central finding of the recent work by Katz et al . where a careful analysis of groups of two and three fish revealed that the mechanisms at play are , at least in the golden shiners studied there , much more subtly intertwined that in existing fish models [28] . Indeed they concluded that alignment emerges from attraction and repulsion as opposed to being an explicit tendency among fish . Whether fish display some mechanisms of active alignment or only attraction/repulsion is likely to lead to different patterns as interactions accumulate over time . In short , extracting interaction rules from individual scale data is crucial not only for animal behavior studies , but also because heretofore overlooked features can be found decisive in governing the emergent collective properties of moving animal groups . Here , we assess the power of a bottom-up methodology to build models for animal group motion from data gathered at the individual scale in groups of increasing sizes . We use data obtained by recording the motion of barred flagtails ( Kuhlia mugil ) in a tank . In natural conditions , the barred flagtail form schools with a few thousands individuals along the reef margin of rocky shorelines , from just below the breaking surf to a depth of a few meters . However the size of these schools is much smaller than in species like the sardine or the Atlantic herring . Our analysis is incremental: in a previous work we characterized the spontaneous behavior of a single fish , including wall-avoidance behavior [29] . Here , using pairs of fish , we first characterize the response function of one fish depending on the position and orientation of the other fish . Then we calibrate multiple fish interactions , using data in larger groups . At each step , the already-determined factors and parameters are kept unchanged and the new terms introduced in the stimulus-response function and the corresponding new parameters are determined from data with nonlinear regression routines ( see Statistical Analysis in Materials and Methods ) . The resulting model is validated by comparing extensive simulations to the original data . Often , different functional forms are tested and we determine which one is most faithful to the data . When no significant difference is found , the simplest version is retained , following a principle of parsimony .
Experiments with 1 to 30 fish were performed in shallow circular swimming pools that let the fish form quasi 2-dimensional schools ( see Fig . 1A and Video S1 , S2 , S3 , S4 ) . At the collective level , we observe a transition from schooling to shoaling behavior when the density of fish increases in the tank: the group polarization , which measures the degree of alignment , is high in groups of two and five fish , even if sometimes we do observe some breaks in the synchronization , while in larger groups , when , it remains low ( Fig . 1B ) . Within each group size , we notice some variability , the most striking effect being an increase of the synchronization level with the individuals velocity in groups of two fish . For every group size , fish move continuously and quickly synchronize their speed to a well defined , but replicate-dependent value ( Fig . S1 ) . The fish trajectories are smooth , differentiable and the instantaneous speed has a well-defined mean and root mean square fluctuations of about 10–20% which are found to be uncorrelated to , the angular velocity of the fish orientation ( Fig . S2 ) . On this basis , fish can be modeled as self-propelled particles moving in 2D space at constant speed and the only dynamical variable retained is . Moreover , since the recorded trajectories , be they extracted from a single fish or from small groups in the tank , are always irregular/stochastic , our model takes the form of coupled stochastic differential equations for the angular velocities of each fish . Note that if noise acts on rather than the fish position or heading , trajectories are smooth and differentiable , as observed . We have shown elsewhere that single fish trajectories in barred flagtails are very well described by an Ornstein-Uhlenbeck process acting on the instantaneous curvature , or , equivalently , on [29] . When the fish is away from the tank wall , the distribution of is nearly Gaussian with zero mean and variance , where is the characteristic time of the ( exponentially decaying ) autocorrelation function of . To avoid collisions with the tank walls , we found that a single fish adjusts its current turning speed towards a ( time-dependent ) target value where is a parameter , is the distance to the point of impact on the wall should the fish continue moving straight ahead , and is the angle between the current heading of the fish and the normal to the point of impact ( see Fig . 2A ) . In short , obeys the stochastic differential equation: ( 1 ) where is a Wiener process of variance reflecting the stochasticity of the behavioral response . Non-linear regression analysis of the above model against our experimental data yielded excellent agreement and accurate estimations of and . Note that in the present work we adopted a slightly different form for the wall avoidance term with regards to the exponentially decreasing one of Ref . [29] , since it actually prevents fish from crossing the tank boundary , while both ansatz are similar as fish moves away from tank walls ( Fig . S8A ) . The stimulus/response function of a single fish in the tank is directly expressed by how varies with the relative position of the fish and the wall . We now assume that this framework holds when two fish and are present in the tank by defining how , for fish , its turning speed is modulated by the combined stimuli due to the wall and to fish . Almost all existing fish behavior models , on the basis of common sense , intuition , and sometimes experimental evidence [30]–[37] , offer a combination of three basic ingredients: short distance repulsion ( to avoid collisions ) , alignment for intermediate distances , and attraction up to some maximal range . Here , we dispose of repulsion not only because we want to allow for the rare experimentally observed over- and under-passings events , but mostly because we do not need to incorporate it explicitly to avoid collisions ( see below and Video S1 , S2 , S3 , S4 ) . In contrast with most existing “zonal” models , and because there is little cognitive/physiological evidence for a sudden switch between alignment and attraction , we want to allow for continuous , distance-dependent weighting between alignment and attraction in agreement with the recent findings of Katz et al . [28] . These two factors a priori depend on the geometrical quantities defining the location of fish from the viewpoint of fish : their distance is involved , but also , the angular position of fish with respect to , the current heading of fish , as well as their relative heading difference ( Fig . 2A ) . The main angular variable for explicit alignment is , as usual , , whereas for attraction it is ; both may also depend on . The stimulus/response function of fish thus combines a priori wall avoidance , alignment and attraction in some unknown function with parameters and ( reaction to the wall ) , , and : . Next , in the spirit of an expansion around the no-interaction case , we write the expression for above as the sum of three terms: ( 2 ) where the “main” variables have been placed first for each term . The wall avoidance term depends explicitly on to reflect a possible screening of the wall by the other fish . We have tested the influence of this by introducing a dependence in the wall avoidance term determined for the single-fish behavior . Essentially , was made smaller for . But this brought no significant improvement , so we keep as found previously . On general grounds , one expects that the relative importance of the positional interaction ( attraction ) to the velocity interaction ( alignment ) increases with . Given that the fish are constrained in a rather small tank , a limited range of inter-distances is effectively explored . In the spirit , again , of a small-distance expansion , a satisfactory choice is given by a linear dependence of on , while is independent of . Of course , such a functional choice cannot be correct at large distances since then would take large unrealistic values , meaning that the fish would spend enormous amounts of energy turning toward a distant “neighbor” ( see the Discussion for more comments on this point ) . The attraction interaction must depend on , the relative angle with the other fish position: it is reasonable to assume that a fish is not attracted much towards a neighbor located behind , and of course this term must be zero when the other fish is right ahead , yielding . A simple , compatible , trigonometric function representing the leading term of a Fourier expansion is the sine function . We thus write where is a parameter controlling the weight of the positional information . Finally , we neglect the possible dependence on : the way a fish would turn toward the position of a neighbor does not depend on the orientation of that fish . This is especially natural when this interaction dominates , i . e . when the neighbor is far away . Moreover knowing the other fish orientation is a cognitively expensive and/or time consuming process at larger distances . The alignment interaction is mostly characterized by its functional dependence on . The main constraint here is that ( the two fish are then already aligned ) . Here again , the simplest choice is as in most models [8]–[10] . Including higher harmonics ( e . g . ) would allow to account for the few observed nematic alignment events where a fish remains anti-aligned with its neighbors . However , incorporating this term did not improve the faithfulness of the model to our dataset , so we keep only the leading sine function . In principle , the strength of alignment can also depend on : less attention may be paid to “back neighbors” . We have tested simple and reasonable choices for the dependence of on , e . g . , but this did not lead to significant improvement so we kept no angular position dependence in the alignment interaction . We thus write , finally: where is a parameter controlling the weight of the orientational information . To summarize the case of two fish and , the stimulus/response function in the general evolution equation ( 1 ) is thus finally written: ( 3 ) Using nonlinear regression analysis , the faithfulness to our data of the model consisting of Eqs . ( 1 ) and ( 3 ) was found very good for each of our two-fish recordings and the 5 parameters , , , and were estimated for each fish . We find clear dependences of the estimated parameters on , the average speed of each fish ( see Fig . 3A ) . In particular , , , and are found proportional to , whereas and no significant -dependence appears for . Results regarding this last parameter are the least convincing , with a large dispersion of individual values . This is mostly due to the confinement of fish in the tank: the positional interaction never dominates alignment , preventing its accurate estimation . Nevertheless it is crucial to note here that without these positional interactions the model fails to match the data . Furthermore , we have tested a posteriori our ansatz by testing each contribution ( either wall avoidance , neighbor position or neighbor orientation ) after the other twos have been subtracted from the fish response according to Eq . ( 3 ) . Results show an excellent agreement between our ansatz and the mean fish response ( for more details see Fig . S8 B–D ) . Note that these results mean also that the wall avoidance is actually governed by , the time it would take the fish to hit the wall , rather than the distance . Conversely , , the relaxation time of the angular velocity , is better expressed as the ratio between a characteristic length and the speed . These -dependences were then incorporated explicitly in the model: ( 4 ) with ( 5 ) where , and are now constants over all fish . Running again our nonlinear regressions using this form , and using data for all replicate , allows for a more accurate estimation of the parameters , , , and now the same for all fish . We find , , , and . To validate this experimental finding , these parameter values were used in simulations of the model which were compared directly to the data . Good agreement is found not only for statistical quantifiers of the emergent synchronization between the two fish ( see Fig . 3C ) , but in fact also for the dynamics: see for instance Video S1 , S2 , S5 , S6 and the time series of polarization which show the same intermittent behavior ( Fig . 3B ) . We emphasize that the model captures the experimental observation that the orientational order is lower when the swimming speed is lower , and is better in faster groups ( Fig . 3B , C ) . Can multiple-fish interactions be factorized into pairs ? This is often taken for granted , following a typical physics approach where this assumption is routinely made . However , recent work has suggested that this is not valid when describing pedestrian interactions in a crowd [38] . Even more recently , Katz et al . argued that this is also the case for groups of three golden shiners [28] ( but see [39] for the case of birds ) . Here , our data set is too small to allow for an in-depth analysis of group behavior at the level of detail that was accomplished above for two fish , mostly because many more variables are involved , but the quality of the pair approximation can be evaluated a posteriori . Assuming that multiple fish interactions are indeed essentially made of the sum of the pair interactions involved , Eq . ( 5 ) is extended to ( 6 ) where is the ( current ) neighborhood of fish which contains individuals . In our observations with fish , individuals mostly stayed together , suggesting that individuals remains aware of all others . Using all-to-all , equal-weight coupling , we found good agreement between data and simulations of Eqs . ( 4 ) and ( 6 ) ( see Fig . S3 ) . This justifies a posteriori the factorization in pairs and the use of two-fish parameters for groups , but also the overall normalization factor in Eq . ( 6 ) , which indicates that , in the stimulus response of a fish , wall avoidance and the averaged influence of neighbors keep , on average , the same relative importance irrespective of the group size . The raw , “force-like” un-normalized superposition would yield too strong a coupling . For the larger group sizes , all-to-all equal-weight coupling quickly becomes unrealistic , and one must determine the set of neighbors a fish interacts with . In principle , abundant data recorded in larger tanks would allow to discriminate between alternative choices , but our experimental recordings are too short for this . Nevertheless , many choices can be eliminated: the usual one , which consists in cutting off interactions at fixed distances ( zonal models ) , is inconsistent with our continuous weighting of alignment and attraction with fish inter-distance . Based on an analysis of starling flocks , Ballerini et al . have argued that these birds actually pay attention to their 6–8 closest neighbors , irrespective of the density of the flock [25] . Coming back to our observations , this non-metric choice of neighbors can , however , lead to unrealistic situations when , for instance , a fish is leading a small group , since then this fish will only pay attention to those behind , even if individuals are located at intermediate distances ahead ( but see Fig . S7 ) . A simple , reasonable , non-metric solution is that of neighbors determined by the Voronoi tessellation around each individual: this allows for continuous weighting between alignment and attraction and avoids the caveat mentioned above in the case of a fixed number of closest neighbors . Moreover , given the rather small inter-distances observed , individuals beyond the first shell of Voronoi neighbors are largely screened out , so that our final choice was that of the first shell of Voronoi neighbors ( see Fig . 2B ) . Using this , the validation of the model simulated with fish using the parameters is again quite satisfactory ( see Fig . S3 ) . This is however not true anymore for larger groups which display too high a polarization when using the parameters ( whereas distance predictions remains satisfactory , see Fig . S3 ) . Our approach actually allows to further investigate this discrepancy . We estimate the parameters at the individual scale for each fish with our nonlinear least-square procedure using the Ito-integrated version of the Ornstein-Uhlenbeck process of Eqs ( 4 ) and ( 6 ) for each fish time series ( see Statistical Analysis ) . Thanks to this parametric inversion strategy , we have been able to extract the parameter values for each replicate separately ( Fig . 4A ) . The model predictions with these replicate-based parameters yield a near-perfect match with the data ( Fig . 4B ) . The results confirm that , within the limits of statistical accuracy , the parameters and their v-dependence remain about the same up to N = 10 , in agreement with the above findings ; but in larger groups there is a decreased tendency of fish to react to their neighbors , which both concerns the alignment and positional interactions ( Fig . 4A ) .
Characterizing and modeling the interactions between individuals and their behavioral consequences is a crucial step to understand the emergence of complex collective animal behaviors . With the recent progress in tracking technologies , high precision datasets on moving animal groups are now available , thus opening the way to a fine-scale analysis of individual behavior [37] , [40]–[42] . Here we adopted a bottom-up modeling strategy for deciphering interactions in fish shoaling together . This strategy is based on a step-by-step quantification of the spontaneous motion of a single fish and of the combined effects of local interactions with neighbors and obstacles on individuals motion . At each step , one model ingredient is considered and checked against experimental data . The required parameters are determined using a dedicated inversion procedure and the numerical values of these parameters are kept unchanged in the following steps , yielding , in the end , a model without any free parameter . Such an incremental procedure fosters the explicit enunciation of the rationale behind each functional choice , and differs from searching the best set of free parameters to fit large-group data [10] , [43] . Proceeding step by step also puts stronger constraints on matching , since the incorporation of additional behavioral features at each step assumes the stability of the previously explored behaviors and of the corresponding model parameters . Using pairs of fish , we were able to show how positional and directional stimuli combine , and the crucial role of the swimming speed in the alignment interaction . At intermediate sizes , multiple fish interactions could be faithfully factorized into pair interactions albeit in a normalized form . However we found that at even larger group sizes our incremental modeling approach fails to accurately reproduce the collective dynamics . We explored this point further , still considering the statistical behavior of each fish separately , but only using the data corresponding to the large-group experiments . We concluded that our model could still grasp the observed individual and collective features but with smaller positional and alignment coefficients . We believe that this decrease in reactivity to neighbors is a consequence of the high density already imposed by confinement effects . Indeed , our model predicts that large groups adopting the high neighbor reactivity found in smaller groups would remain polarized also in open space , keeping group cohesion with an average distance to neighbors of about two body lengths ( Fig . S6 ) . Since the largest groups we observed in the tank are already characterized by such a typical neighbors distance due to confinement effects , we argue that lower interaction strengths may simply indicate the fish vanishing need to actively react to neighbors position and heading in order to maintain a high density . This could be , for instance , a physiological consequence of the density per se: the physiological and behavioral consequences , for an individual , of living in dense groups , known as group effect , have been described in numerous species from insects to vertebrates [44] , [45] . Our results investigation suggests that this sensitivity may be represented in a quite straightforward manner , preserving the model shape of Eqs . ( 4 ) and ( 6 ) and only modifying the interaction parameters . This conjecture , of course , could only be validated by experiments on large groups conducted in open space or larger tanks . While we believe in a positive answer , namely that without too strong a confinement , individuals would react to the perceived neighbors the same way regardless of the overall group size , we leave this question for future investigations on group effect in fish schools . Our approach yielded a novel type of fish school model whose main features are its built-in balancing mechanism between positional and orientational information , a topological interaction neighborhood , and explicit dependencies on fish speed . Note that similar features were recently uncovered for another species thanks to a novel data analysis procedure [28] . The smooth transition from a dominant alignment reaction when a neighbor is close to attraction when it is far away is in line with a simple additive physiological integration of both information [46] . The linear dependence of the positional interaction strength on fish inter-distance obviously cannot hold for sparse groups , and will have to be modified by introducing a long-distance saturation when dealing with situations where confinement effects are weaker . Even if we claim that a Voronoi neighborhood was the best choice to account for our data thus extending the relevance of topological interactions , we also checked that our conclusions were robust against this choice , by testing a simple K-Nearest Neighbors network of interactions ( which remains topological [25] ) . We computed the model predictions with the parameters estimated for groups of N = 2 fish , but considering only the K nearest neighbors for increasing values of K ( K = 1 to 7 , and 10 ) . The results are reported in Fig . S7 ; the main impact of a lower level of connectivity is a decrease of polarization , but it does not lead to better predictions at the collective scale . Interestingly , the best predictions were found with a number of nearest neighbors that corresponds to the average number of neighbors belonging to the first shell in a Voronoi neighborhood ( , Fig S7–B ) . This number of influential nearest neighbors is remarkably similar to the one found in starlings [25] and in contrast with recent results found by Herbert-Read et al . in mosquito fish [47] . Further dedicated experiments will be required to discriminate between alternative choices of the relevant neighborhood . The speed dependence of the parameters , directly derived from our data , is in contrast with most previous fish school models . It leads to an increase of group polarization with swimming speed , a direct consequence of the predominance of alignment at high speed ( see Video S7 ) . In natural conditions , this mechanism could be involved in the transitions from shoaling at low speed often associated with feeding behavior to polarized schooling at high speed associated with searching for food . Such speed change could also be elicited by the detection of a threat and abrupt transitions can occur when fish suddenly increase their speed , for instance generating a flash expansion ( see Video S8 ) . The question of whether the propagation of such an excitation wave within large schools can generate an efficient collective evasion call for further experimental tests [48] . The reason why our approach was fruitful in spite of the limited amount of data available lies largely in the suitable properties of the behavior of the fish studied: the smooth fluctuations of tangential speed and their de-correlation from angular velocity variations were essential in limiting the number of variables at play but also allowed for a faithful account of single fish behavior by a simple Ornstein-Uhlenbeck process . Clearly it is likely that more complicated solutions will be needed for other species where tangential and angular accelerations are intimately coupled and/or the underlying stochastic process is not as transparent [28] . Nevertheless , we expect that , pending sufficient amounts of data , our approach could be successfully applied to more complex situations occurring in various biological systems at different scales of organization .
Our experiments were all carried out in full accordance with the ethical guidelines of our research institutions and comply with the European legislation for animal welfare . The welfare of fishes in the tanks was optimized with a continuous seawater flow , a suitable temperature , and oxygen content . The maximum density in the holding tank was lower than . During the experiments , low mortality occurred ( five individuals ) . At the end of the experiment , the fish were released at their capture site . The experiments were performed from April to June 2001 at the Sea Turtle Survey and Discovery Centre of Reunion Island . Barred flagtail Kuhlia mugil ( Forster ) were caught in March 2001 in the coastal area around Reunion Island . 80–100 fishes were conveyed to the marine station and housed in a holding tank of 4 m diameter and 1 . 2 m depth . Fishes were fed daily ad libitum with a mixture of aquaria flake-food and pieces of fish flesh . Fishes were considered acclimatized when all of them feed on the aquaria flake-food . This weaning period lasted 15 days . Experiments were performed in a circular tank similar to the holding tank . Opaque curtains were placed around and above the tank to obtain diffuse lighting and to reduce external disturbances from the environment . The tank was supplied with a continuous flow of seawater [49] . Since currents may influence fish behavior , the seawater inlet pipe was placed vertically and the water flow was stopped throughout the observation periods . A digital video camera ( Sony model CDR-TRV 900E ) was fixed at 5 meters above the tank and tilted at to observe the whole tank . The remotely operated video camera was fitted with a polarizing filter and a wide-angle lens . Groups of N = 1 to 30 fish were introduced in the experimental tank and acclimatized to their new environment for a period of 20 min . Their behavior was then recorded at 24 fps for 2 mins . Prior to each trial , the fish were deprived of food for 12 hours to standardize the hunger level and were transferred to the experimental tank . The relative shallowness of the water ensured quasi two-dimensional motion . Five replicates per group size using different individuals were performed . Eighty per cent of the trials were performed in the morning to avoid possible conditions of strong wind that may disturb the fish , and sunshine that may render light inside the tank unsuitable for video recording . A first data processing consisted in sampling 12 images per second out of the 24 images recorded by the video camera . A custom-made tracking software was then used to extract high-quality , smooth trajectories from the video recordings , with crossing ambiguities resolved by eyes ( see Video S3 , S4 ) . In order to get even higher precision data , the head position and the orientation of each fish in groups of N = 2 were acquired with a manual tracking software ( Video S1 , S2 ) . Model parameters were estimated from each fish time series separately ( typical series are shown on Fig . S4 ) . In order to perform the estimation of the parameters , , , and in the stochastic differential equation ( 1 ) , ( 3 ) and ( 5 ) , we considered its discrete-time version using Ito integration over , assuming is small enough so that is constant [50]: ( 7 ) where i = 1 , 2 and is given by Eq . ( 3 ) or ( 5 ) . Estimates for the parameters were obtained using a standard non-linear least squares procedure ( we employed the nls package of the statistical environment R [51] ) either separately for each fish using Eq . ( 3 ) or for all fish together using Eq . ( 5 ) . Residuals given by were checked to be Gaussian-distributed ( see Fig . S5 ) and their variance yielded . The model was simulated within a virtual tank , using the estimates of behavioral parameters extracted by statistical analysis from time-series in groups of fish . The fish heading ( direction of motion ) and position were updated by Euler integration , following: ( 8 ) where . For each value , numerical simulations were performed over 120 seconds ( a time corresponding to the duration of individual experiments with real fish ) with a time step . A transient time of was discarded before measuring statistical averages . We computed the mean value and the variance over time of the global polarization ( 9 ) and of the neighbor inter-distance ( 10 ) This yielded an estimation of the expected measures distribution under model hypothesis and over the typical observation time of experiments . We then computed the mean and confidence interval of such distributions , to obtain the expected mean and variance ( with their confidence intervals ) of alignment and of neighbor inter-distance . This provided the check of the model against experimental data . The above procedure was repeated varying the mean speed over the range covered by the experimental data , with the results plotted in Fig . 3C . The same procedure was adopted to make predictions for higher group sizes , using the stimulus/response function as determined by equation ( 5 ) with interacting neighbors defined by first neighbors in a Voronoi tessellation ( For a set of points , Voronoi tessellation divides the space in different cells , each the locus of space closer to its center than to any other points in : at each time step space is divided in Voronoi cells centered around the fish position , with Voronoi neighbors being the fish lying in neighboring cells ( Fig . 2B ) . For each experimental replicate , the same measures were repeated with the parameters extracted from the replicate , and the corresponding initial conditions ( Fig . 4B ) . By construction , our method does not “learn the parameters to make the model fit” , contrasting with a more usual procedure which consists in stating an a priori model and searching a best set of free parameters that optimizes its collective patterns towards the observed collective properties ( namely , make the model fit at the collective scale ) . In such cases , it is known that several models can adjust the data at the collective scale ( because the search for best match is unconstrained and can be performed for each model , so that the collective level underdetermines the individual level ) . In the present study , once the model has been formulated , that is , once we identified in the experiments with pairs of fish the nature of stimuli ( the orientation and relative position of neighboring fish , and how they combine to determine the response of a focal fish ) , we estimated the values of 5 parameters at the individual scale . So for each fish , we measured its behavioral response ( i . e . the change of its turning speed ) for each configuration of stimuli encountered in its path . Only then , we tested whether these parameters measured at the individual level can explain the observations at the collective scale with no free parameters . For each group independently , we thus checked that the model allows a quantitative matching concurrently at individual and collective scales . This confirmed that our model calibrated with the parameters estimated from the third derivative of the fish position ( i . e . the change in the turning speed ) was able to reproduce quantitatively the statistics resulting from the time integration of the coupling between fish ( polarization , inter-distance ) . Moreover the same procedure applied separately on each group size revealed , on the one hand , the dependences of the estimated parameters on the swimming speed ( using groups of N = 2 fish ) , and on the other hand , the modulation of interactions' strength with group size ( in the largest groups ) .
|
Swarms of insects , schools of fish and flocks of birds display an impressive variety of collective patterns that emerge from local interactions among group members . These puzzling phenomena raise a variety of questions about the behavioral rules that govern the coordination of individuals' motions and the emergence of large-scale patterns . While numerous models have been proposed , there is still a strong need for detailed experimental studies to foster the biological understanding of such collective motion . Here , we use data recorded on fish barred flagtails moving in groups of increasing sizes in a water tank to demonstrate the power of an incremental methodology for building a fish behavior model completely based on interactions with the physical environment and neighboring fish . In contrast to previous works , our model revealed an implicit balancing of neighbors position and orientation on the turning speed of fish , an unexpected transition between shoaling and schooling induced by a change in the swimming speed , and a group-size effect which results in a decrease of social interactions among fish as density increases . An important feature of this model lies in its ability to allow a large palette of adaptive patterns with a great economy of means .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"statistical",
"mechanics",
"neuroscience",
"animal",
"behavior",
"animal",
"management",
"phase",
"transformation",
"biology",
"agriculture",
"behavioral",
"ecology",
"physics",
"systems",
"biology",
"computer",
"science",
"condensed-matter",
"physics",
"computer",
"modeling",
"ecology",
"biophysic",
"al",
"simulations",
"interdisciplinary",
"physics",
"computational",
"biology",
"animal",
"cognition"
] |
2012
|
Deciphering Interactions in Moving Animal Groups
|
Enterotoxigenic Escherichia coli ( ETEC ) is an endemic health threat in underdeveloped nations . Despite the significant effort extended to vaccine trials using ETEC colonization factors , these approaches have generally not been especially effective in mediating cross-protective immunity . We used quantitative proteomics to identify 24 proteins that differed in abundance in membrane protein preparations derived from wild-type vs . a type II secretion system mutant of ETEC . We expressed and purified a subset of these proteins and identified nine antigens that generated significant immune responses in mice . Sera from mice immunized with either the MltA-interacting protein MipA , the periplasmic chaperone seventeen kilodalton protein , Skp , or a long-chain fatty acid outer membrane transporter , ETEC_2479 , reduced the adherence of multiple ETEC strains differing in colonization factor expression to human intestinal epithelial cells . In intranasal challenge assays of mice , immunization with ETEC_2479 protected 88% of mice from an otherwise lethal challenge with ETEC H10407 . Immunization with either Skp or MipA provided an intermediate degree of protection , 68 and 64% , respectively . Protection was significantly correlated with the induction of a secretory immunoglobulin A response . This study has identified several proteins that are conserved among heterologous ETEC strains and may thus potentially improve cross-protective efficacy if incorporated into future vaccine designs .
Enterotoxigenic Escherichia coli ( ETEC ) is a significant cause of human morbidity due to infectious diarrhea and resultant malnutrition [1] . A recent Global Enteric Multicenter study conducted over a 3-year period to identify the etiology of pediatric diarrheal diseases in sub-Saharan Africa and South Asia found that ETEC infection led to moderate to severe diarrhea in 60–70% of ETEC infected patients and found that ETEC was present at all study sites [2] . ETEC are a diverse group of pathogens that colonize the small intestine , where they attach to mucosal surfaces using surface antigens known as colonization factors [CFs; [3] . ETEC infections are associated with an acute watery diarrhea that can lead to rapid dehydration [1] . At least 25 unique CFs have been identified [4] . ETEC strains also express heat-labile ( LT ) and/or heat-stable ( ST ) enterotoxins [5] . The enzymatic activities of these enterotoxins cause diarrhea by ultimately inducing water and electrolyte loss from the intestine of infected subjects [5] . Several strategies have been used for ETEC vaccine development . Purified CFs have been used as oral immunogens to provide protection against later challenge with ETEC expressing homologous CFs [6] . A commonly used approach has involved using the cholera toxin B subunit ( CT-B ) with formalin-inactivated ETEC strains expressing the most prevalent CFs [7] . This approach showed that the vaccine elicited IgA responses against the different CFs that were used [6] . However , further trials based on this approach with vaccines expressing CFA/I , CS1-3 , CS5 , and a recombinant CT-B suggested the need to improve vaccine safety in infants and young children [8 , 9] . A new version of this oral vaccine with an increased level of CF expression is being tested [10] . An approach with a live attenuated oral ETEC vaccine was also taken where an ETEC variant ( E1392/75-2A ) that had lost the capacity to produce toxin but still expressed CFA/II was used for oral vaccination . The vaccination showed 75% protection against ETEC expressing CFA/II [11] . E1392/75-2A was further attenuated and found to be immunogenic and safe to administer to humans [11] . However , challenge studies have , to our knowledge , not been conducted to determine protective efficacy . A recent study combined six ETEC vaccine strains expressing different CFs with the LT B subunit [12] . This vaccine formulation ( ACE527 ) , which was used in a phase I trial [12] , was well tolerated and immunogenic [13 , 14] and may be the subject of future development . An attenuated S . flexneri 2a vaccine strain CVD 1204 , bearing deletions in the guanine nucleotide biosynthetic pathway ( ΔguaBA ) , was used to express different ETEC colonization factors along with mutated forms of LT ( LThK63 or LThR72 ) ; [15] ) . Following intranasal immunization of guinea pigs , both serum IgG and mucosal IgA responses against the CFs were detected . However , the immune response was not consistent against the mutant LT [16] . This approach was improved upon by using four different ETEC fimbrial antigens ( CFA/I , CS2 , CS3 , and CS4 ) and LThk63 expressed separately in CVD 1204 [17] . However , the invasive capacity of S . flexneri ΔguaBA was significantly reduced , which may limit its ability to stimulate robust immune responses [18] , and expression of ETEC CFs further reduces its invasiveness [19] . An invasive strain , S . flexneri 2a ( SC608 ) was also developed for heterologous antigen expression [20] . All these studies showed a significant immune response against ETEC CFs . However , none of these immunization studies were , to our knowledge , followed with ETEC challenge , due to lack of a proper animal model to assess directly protection against ETEC infection . Plasmid-based antigens such as EtpA and EatA have protective efficacy against ETEC challenge in a mouse model [4] . EtpA is secreted by a two-partner secretion system and functions as an ETEC adhesin that promotes colonization of mice [21] . Immunization with recombinant EtpA provided protection against ETEC H10407 challenge , based upon reduced ETEC colonization of the mouse intestine [21] . EatA is a serine protease that degrades EtpA [22] . EatA also degrades MUC2 , a major component of intestinal mucin [23] . The EatA passenger domain is immunogenic and protective against ETEC infection of mice [23] . The eatA gene was present in ~70% of ETEC strains collected from Chile and Guinea Bissau [23] . Other proteins under investigation for their potential to be used in ETEC vaccines are the YghJ metalloprotease and EaeH , a putative adhesin [4] . YghJ degrades MUC2 and antibodies raised against YghJ inhibit LT delivery [24] . EaeH has immunoglobulin-like domains which are similar to domains of other proteins involved in adhesion [4] . Further studies are needed to realize the full potential of these proteins as vaccine candidates . Genes involved in the biogenesis of Type I fimbriae in ETEC are upregulated in their expression upon contact with epithelial cells [25] . Type I fimbriae were used as a vaccine candidate in pigs and significantly reduced the intestinal load of ETEC [26] . It may be possible to use highly conserved subunits of Type I fimbriae as potential vaccine candidates [4] . A major challenge to ETEC vaccine development lies in the genomic diversity of the strains [27] . Ultimately , it may be desirable to identify antigens that are sufficiently conserved to provide a broad range of protection against the diverse population of ETEC strains that cause diarrheal disease [4] . ETEC protective antigens remain relatively poorly defined in comparison with studies of other diarrheal pathogens . Indeed , with the exception of antibody responses to LT and CFs , relatively little is known about the human immune response to ETEC infection . Genome sequence data are now available for several ETEC isolates that have previously been studied in human vaccine trials . These data suggest that the genomes of these organisms are variable [27] . The apparent sequence diversity among strains suggests that identifying broadly conserved antigens with a “reverse vaccinology” approach [28] , rather than focusing on a limited class of proteins ( i . e . CFs ) may be an effective strategy for vaccine development . Bacterial surface proteins are attractive as targets as vaccine components [29] . However , identifying these proteins can be difficult , and in the past has typically required using chromatographic separation methods [30] susceptible to contamination [31] . ETEC utilizes a type II secretion system ( T2SS ) to transport proteins from the cytoplasm to the outer membrane [32] . While the substrates of the T2SS are incompletely characterized , we hypothesized that at least a subset of these proteins may be effective vaccine components . In the current study , we identified several ETEC proteins that differed in their abundance in membrane protein preparations from wild-type ( WT ) vs . a type II secretion mutant of ETEC . After purifying these proteins , we generated antisera and characterized the ability of these antisera to protect cultured intestinal epithelial cells from adherence by a diverse panel of ETEC strains differing in CF type . We also identified three ETEC proteins that provide protective immunity in an intranasal mouse challenge model .
The Kansas State University Institutional Animal Care and Use Committee approved the animal procedures ( IACUC protocol #3196 ) in the context of the Kansas State University Animal Welfare Assurance Number A3609-01 , in compliance with the Public Health Service ( PHS ) Policy on Humane Care and Use of Laboratory Animals . The ETEC strains used are described in Table 1 . ETEC H10407 was originally isolated from a patient with a severe cholera-like diarrhea in Bangladesh [33] . The ETEC H10407 ΔgspE mutant was a gift from Dr . James M . Fleckenstein [21] . ETEC strains used for in vitro adherence assays were gifts from Dr . Stephen Savarino . Deletion mutants in the skp , mipA , and ETEC_2479 genes were generated using the λ-Red recombinase method [46] . Mutants were confirmed by both PCR screening and DNA sequencing . ETEC were plated on colonization factor antigen ( CFA ) agar plates ( 1% Casamino acids , 0 . 15% yeast extract , 2% agar , 0 . 4 mM magnesium sulfate , 0 . 04 mM manganese chloride ) and grown at 37°C . For adherence assays , individual colonies were inoculated in Luria-Bertani ( LB ) broth overnight followed by re-inoculation of overnight cultures in Eagle’s minimal essential medium ( EMEM ) to an OD600 of 0 . 6 . Cytoplasmic and membrane protein fractions were obtained by growing WT and ΔgspE mutant ETEC H10407 strains overnight in 50 ml CAYE media ( 2% Casamino acids , 0 . 6% yeast extract , 43 mM NaCl , 38 mM K2HPO4 , 0 . 25% glucose , 0 . 1% trace minerals ) . Bacterial cultures were centrifuged , washed in PBS , and then treated with lysozyme and recentrifuged . The pellet was resuspended in 10 mM Tris-HCl , pH 7 . 0 , sonicated , and centrifuged . The supernatant was subjected to ultracentrifugation ( 1 h , 50 , 000 g , 4°C ) , after which the supernatant was retained as the cytoplasmic fraction . The pellet was resuspended in 10 mM Tris-HCl , pH 7 . 0 and retained as the membrane fraction . Equal amounts of protein from each fraction ( ~50 μg ) were digested to peptides with trypsin , differentially labeled with light- and heavy-isotopologs of formaldehyde ( Cambridge Isotope Labs ) , combined , resolved by isoelectric focusing , and analyzed by nLC-MS/MS using an LTQ-Orbitrap mass spectrometer . The Mascot search engine was used to compare the measured fragment spectra against an in silico translation of the ETEC H10407 genome . PCR primers were designed so as to amplify genes without predicted signal-peptide coding sequences . PCR products were generated from chromosomal ETEC H10407 DNA and introduced into pET42a to generate recombinant proteins fused to a glutathione-S-transferase ( GST ) epitope . Plasmids were transformed into E . coli BL21 ( DE3 ) cells . Bacterial cultures were grown overnight at 37°C and subcultured 1:100 into fresh media . The subcultured cells were grown for 2 h at 37°C until reaching an OD600 of 0 . 2–0 . 5 . Isopropyl β-D-1-thiogalactopyranoside ( IPTG; 1 mM ) was added and the bacterial cultures were grown for an additional 2 h , after which the cells were centrifuged ( 10 min , 10 , 000 g , 4°C ) . The cells were lysed in 1/20 culture volume of Bugbuster Protein Extraction Reagent ( Novagen ) , rotated for 15 min at room temperature , and subjected to ultracentrifugation ( 1 h , 2 , 500 g , 4°C ) , after which the supernatant was retained . Supernatants were applied to GST-Bind Resin ( Novagen ) , incubated at 4°C overnight with rotation , centrifuged ( 5 min , 1 , 000 g , 4°C ) , washed 3 times with GST Wash Buffer ( Novagen ) , and eluted in GST elute buffer . The GST protein was also expressed separately and purified using similar conditions for use as an immunization control in subsequent experiments . The GST tags were not removed after protein purification . Potential contamination of antigen preparations with LPS was quantified using the Chromogenic Endotoxin Quantitation Kit ( Pierce ) . LPS was present at negligible quantities in antigen preparations used for immunization ( less than 3 pg/dose ) . Mice were immunized with 200 μg of each purified protein suspended in 50 mM Tris HCl pH 7 . 5 , 50 mM NaCl , mixed with Complete Freund’s Adjuvant , and administered into multiple sites subcutaneously . Booster injections were administered twice at 2-week intervals , using 200 μg of antigen mixed with Incomplete Freund’s Adjuvant . Two weeks after the final immunization , the mice were euthanized , exsanguinated , and the blood was processed into serum . Control serum was also obtained from mice treated with PBS or a GST-epitope control protein . For each antigen , mouse sera were individually tested using ELISAs , then pooled by group , and stored at -80°C . IgG and IgA concentrations in mouse sera and feces , respectively , were analyzed using ELISAs . We collected blood by intracardiac puncture after sacrificing the animals after challenge . To obtain fecal antibodies , we collected five fresh stool pellets from each animal . Stool pellets were added to 1 ml of fecal reconstitution buffer ( 50 mM ethylenediaminetetraacetic acid ( EDTA ) , 0 . 1 mg/ml soybean trypsin inhibitor , 1 . 39 μg/ml phenylmethylsulfonylfluoride ( PMSF ) , and homogenized . The samples were centrifuged ( 5 min , 5 , 000 g ) to remove insoluble material and the supernatants were stored at -80°C . ELISAs were performed in polystyrene 96-well , flat bottom plates ( Whatman ) coated with 0 . 5 μg/ml of each purified protein or BSA and incubated overnight at 4°C . Plates were washed 3X in PBS , 0 . 1% Tween-20 and blocked with 5% milk in PBS , 0 . 1% Tween-20 for 1 h at room temperature . For serum IgG analysis , 50 μl of each serum sample was added in duplicate to antigen-coated wells and incubated at 37°C for 1 h . Goat anti-mouse Ig ( H+L ) HRP detection antibody ( Southern Biotech ) diluted 1:4 , 000 in 0 . 1% PBS-Tween was added to the wells and incubated 37°C for 30 min . Plates were developed with 1-StepTM Ultra TMB-ELISA ( Thermo ) and quenched with 3 N H2SO4 . Absorbance was read at 450 nm . Antibody titers were transformed logarithmically and the Student’s t test was used to compare the mean serum antibody titer values of different groups of mice with those of non-immunized mice . Differences in P values of < 0 . 05 were considered significant . For fecal IgA analysis , 50 μl of fecal supernatants were used with a rabbit anti-mouse IgA HRP detection antibody ( Sigma ) and developed as described above . HCT-8 cells ( 5*104 cells/well ) were grown in 24-well plates and incubated for 16 h in EMEM . We added serial dilutions ( 5 μg/ml to 500 pg/ml ) of antisera prepared against ETEC antigens or control sera from non-immunized mice , followed by 5*105 colony-forming units ( CFUs ) of ETEC strains that are diverse in CF-type ( Table 1 ) . HCT-8 cells were incubated with ETEC for 1 h at 37°C . Following incubation , cells were washed three times with sterile PBS to remove unbound ETEC and then lysed in 1% saponin . Cell lysates were serially diluted and plated for enumeration of adherent bacteria . We calculated the number of bacteria that remained cell-associated , relative to the bacterial inocula and computed the % reduction in adherence resulting from serum-containing polyclonal antibodies raised against ETEC proteins vs . serum from non-immunized mice . We compared the data statistically using paired Student’s t tests and considered p-values < 0 . 05 significant . Female BALB/c mice ( 15/group ) were obtained from the Jackson Laboratory ( Bar Harbor , Maine ) . Mice were housed in microisolator cages and provided with food and water ad libitum . Antigens were administered intranasally at 20 μg/dose by mixing ETEC proteins with 2 . 5 μg of cholera toxin ( Sigma-Aldrich ) in 25 μl PBS to the external nares of mice lightly anesthetized with isoflurane . Booster doses were administered 2- and 4-weeks after the initial vaccination . Intranasal challenge studies were conducted as previously described [47 , 48] , with minor modifications . Two weeks after the final immunization , the mice were lightly anesthetized with isoflurane in a VetEquip RC2 isoflurane anesthesia machine and challenged intranasally with 5*108 CFUs of ETEC H10407 . To quantify changes to mouse clinical signs of illness as a function of ETEC challenge and immunization , we observed mice every 4 h after challenge and recorded the clinical signs of illness ( lack of responsiveness to stimulation , hunched posture , ruffled hair coat , dehydration ) as a function of time . Data were analyzed statistically using log-rank tests . If mice displayed clinical signs of illness , or at the end of the study ( 7 d ) , they were euthanized , necropsied , and their lungs were removed aseptically . Lungs were homogenized , serially diluted in PBS , and plated on MacConkey agar to enumerate ETEC .
We compared the relative abundance of proteins from membrane protein preparations between WT ETEC H10407 and an isogenic strain deficient in T2SS function ( ΔgspE ) using a mass spectrometry-based approach . GspE provides an ATP hydrolysis-dependent conformational change to the T2SS that drives protein secretion [49] . We isolated proteins from WT and ΔgspE ETEC and fractionated the bacterial proteins to cytoplasmic , secreted , and membrane fractions . We digested proteins with trypsin and differentially labeled the peptides with isotopologs of formaldehyde . We mixed equal quantities of protein from each sample and resolved them by in-solution isoelectric focusing ( IEF ) . IEF fractions were further analyzed by nLC-MS/MS using an LTQ-Orbitrap mass spectrometer . We used the Mascot search engine to compare the measured fragment spectra against an in silico translation of the ETEC H10407 genome . We computed the relative abundance ratios of the proteins that were isolated and identified from both WT and ΔgspE ETEC . We identified 402 proteins from 1 , 901 sequenced peptides and refined this list to 24 proteins that were enriched in the membrane fractions of WT , but not ΔgspE ETEC . Our analysis revealed several membrane proteins , flagellar subunits , and putative invasins ( Table 2 ) . Of these 24 ETEC proteins identified , we cloned , expressed , and purified 9 antigens for further characterization , primarily based upon the solubility of recombinant forms of the proteins . We amplified the genes encoding the identified proteins in such a way to exclude predicted signal-peptides . PCR products were generated from chromosomal ETEC H10407 DNA and cloned as fusions to glutathione-S-transferase ( GST ) . These plasmids were expressed in E . coli BL21 ( DE3 ) and GST-recombinant proteins were purified using GST-Bind Resin ( Fig 1A ) . We used ELISAs to quantify antibody titers in pooled mouse sera . The IgG responses were variable , with the FlgE flagellar hook protein , yielding a 560-fold response ( Fig 1B ) . Other antigens induced a 3- to 85-fold increase in IgG concentrations ( Fig 1B ) . Several other proteins were not studied further in the context of this work , as they were partitioned into inclusion bodies and not easily maintained in a soluble form . It is conceivable that their fusion to other epitope tags or their purification in the presence of urea might overcome these challenges in future studies . We quantified the extent to which the mouse antisera would subsequently protect against ETEC H10407 adherence to HCT-8 cells . Of the 9 antigens tested , 3 antigens ( Skp , MipA , and ETEC_2479 ) showed a significant ability to inhibit H10407 adherence ( Fig 2A ) . To determine the specificity of these antisera for their respective antigens , we deleted individually the genes encoding Skp , MipA , and ETEC_2479 and then re-evaluated the ability of the respective antisera to protect against the adherence of these ETEC mutants to HCT-8 cells . The antisera inhibited the WT strain and the 2 heterologous ETEC mutants , but no longer inhibited the adherence of the corresponding homologous ETEC mutant to as great an extent , suggesting that the phenotypes observed are largely attributable to antisera specificity ( Fig 2B ) . Antisera raised against Skp , MipA , and ETEC_2479 were further examined for their protective efficacy in reducing the adherence to HCT-8 cells of a panel of other ETEC strains that differ in CF-type ( Table 1 ) . Antisera raised against ETEC_2479 provided the highest degree of protection against the panel of ETEC strains . Antisera raised against Skp protected against most of the strains , except ETEC M421C1 . Antisera raised against MipA protected against many strains , but failed to reduce the adherence of ETEC strains M421C1 , 27845–1 , and E20738A ( Fig 2C ) . We conducted an intranasal challenge assay to evaluate the efficacy of immunizing mice against Skp , MipA , and ETEC_2479 in protecting mice against an otherwise lethal challenge with ETEC H10407 . Mice ( 22-29/group ) were immunized three times at two-week intervals with individual antigens combined with cholera toxin . Mice were inoculated intranasally with ETEC H10407 and then evaluated for clinical signs of disease over a 7-day period . All three antigens were protective against the infectious challenge to variable degrees ( Fig 3A ) . ETEC_2479 was the most effective , providing protection for 22 of 25 ( 88% ) of the immunized mice . Immunization with Skp or MipA yielded lesser , though still significant degrees of survival , with 17/25 ( 68% ) and 16/25 ( 64% ) of the mice surviving the infection . By contrast , only 2/29 ( 7% ) of control mice treated with PBS and 2/20 ( 10% ) of mice immunized with a GST purification control survived longer than 76 h after infection . We observed high loads of ETEC ( ~109−10 CFUs/g ) in the lungs of mice that were euthanized due to their presentation of clinical signs of disease ( Fig 3B ) . By contrast , relatively little ETEC ( ~103−5 CFUs/g ) was cultured from the lungs of mice that survived the infection ( Fig 3B , compare open to closed symbols ) . ETEC loads were inversely related to the concentrations of IgA obtained from the feces of infected mice ( Fig 3C ) . Fecal IgA responses were highly variable , but significantly correlated with mouse survival ( Fig 3C , compare open to closed symbols ) . Mice that did not develop significant fecal IgA responses against the antigens did not survive the infectious challenge .
Bacterial surface proteins mediate adhesion and invasion to host cells and are often important vaccine candidates . The T2SS is considered important to ETEC virulence , as it is involved in secretion of proteins to the bacterial surface [50 , 32] . We compared the difference in abundance between membrane protein preparations from ETEC H10407 strains possessing or lacking a functional T2SS system . In vitro adherence studies using the antisera produced in mice refined our study towards three antigens that showed a protective effect against bacterial adhesion to host cells . Similar approaches could be used to identify ETEC OMPs that are secreted by pathways other than the T2SS . One limitation of our work is that many of the proteins we identified are somewhat unexpected and may not be entirely consistent with what might be expected from mutating the T2SS . The molecular basis for this result is unclear . However , we evaluated only the differential abundance of the proteins between the WT and the ΔgspE mutant and did not formally evaluate the secretion of these proteins by the T2SS . We rather chose to focus on the extent to which antibodies developed against these proteins might protect both against ETEC adherence in vitro and in the intranasal challenge model . Studying the mechanisms by which the ETEC antigens described are secreted may be interesting topics for future research . The antigens identified in this study were effective in preventing ETEC attachment to HCT-8 cells , as well as in reducing lethality in an ETEC infectious challenge in mice . Skp is a molecular chaperone involved in outer membrane protein biogenesis that is believed to rescue misdirected OMPs [51] but may also function as a general chaperone [52] . The utility of using chaperones as vaccine candidates has been shown with other bacteria such as Clostridium difficile [53] and Brucella melitensis [54] . In addition , previous studies have suggested that Skp may be a component of the outer membrane [55] . MipA belongs to the MltA-interacting protein superfamily involved in remodeling peptidoglycan . Although few studies have focused on this protein as a vaccine candidate , it was previously identified as an ETEC immunoreactive protein [56] . MipA of Salmonella paratyphi A has been used as a vaccine candidate with moderate protection [57] . The ETEC_2479 long chain fatty acid transporter OMP is predicted to function as a multifunctional outer membrane porin essential for transport of long chain fatty acids and as a receptor for T2 bacteriophage [58] . Studies conducted in other Gram-negative bacteria have shown that porin proteins can be used as potential vaccine candidates [59] . Several other proteins identified in our initial proteomic analysis merit brief mention with regard to their potential use in vaccine studies . The OmpW protein is required for resistance to phagocytosis and is protective against intraperitoneal infectious challenge with E . coli [60] . ClpX was shown to have potential use in a vaccine to protect against the nasopharyngeal colonization of Streptococcus pneumoniae in mice [61] . OmpA is a major immunoreactive E . coli outer membrane protein that has been studied for its role in bacterial invasion of brain microvascular endothelial cells [62] . TibA is a 104 kDa outer membrane protein present on the tib locus of the ETEC chromosome that contributes to ETEC invasion [63] and shares similarity with several autotransporter adhesins of mucosal pathogens [4] . The three antigens we characterized in detail reduced the attachment of different ETEC strains to HCT-8 cells , independent of their CF-type , and might thus be useful in future vaccine preparations . The in vivo infectious challenge study with ETEC H10407 in mice showed that all the antigens were protective . Mice immunized with ETEC_2479 showed the best protection ( 88% ) , while mice immunized with Skp or MipA showed moderate protection ( 68 and 64% , respectively ) . Mice that survived the infectious ETEC challenge had higher fecal IgA concentrations , supporting the notion that the antigens elicited mucosal immune responses to prevent bacterial colonization . The inverse relationship of bacterial loads in the lungs , as well as the survival of mice , correlated well with the fecal IgA levels , as mice that survived the challenge had reduced ETEC loads , as compared with the mice that died . Investigating ETEC intestinal colonization in mice requires pretreating the animals with antibiotics to remove other bacteria and the exact mechanism by which ETEC colonize the small intestine of mice is not fully understood [64] . One of the drawbacks of using the murine intestinal model for ETEC challenge is that mice do not develop diarrhea even when they are exposed to high doses of ETEC [21] . We were primarily interested in conducting survival assays to determine the efficacy of our vaccine candidates . We therefore elected to use an intranasal inoculation model . This model has been used previously as a model for enteric infections [47 , 48] . The mouse intestinal epithelium shares similarity to the bronchus of the lung , as the lymphoid follicles present in the bronchial wall are similar to the Peyer’s patches of the intestine . Mucosal immunity is an important protective mechanism , as attachment to mucosal surfaces is often the first step in establishing infection . This protection is mainly provided by locally secreted sIgA , which neutralizes the pathogens at the point of infection [65] . If antigen is delivered at the mucosal surface , it can induce a strong immune response in the locally associated lymphoid tissue which , in turn , can travel to distal mucosal surfaces [66] . For vaccine development , it is important to target conserved antigens , yet , due to the diversity in ETEC serotypes and the large number of CFs identified , it is difficult to choose any one serotype-specific CF [67] . Finding broadly conserved , protective antigens is of primary importance . Our approach has resulted in finding three antigens that show heterologous protection against a variety of ETEC strains that are diverse in CF type . Preliminary analysis of a panel of 89 strains isolated from Bangladesh showed that MipA , Skp , and ETEC_2479 were prevalent in 89 ( 100% ) , 89 ( 100% ) , and 83 ( 93% ) of the strains , respectively ( D . Rasko , personal communication ) . However , a potential drawback also exists , as these antigens are conserved among other pathogenic and commensal E . coli strains ( Table 2 ) , although the extent to which they are expressed and surface-exposed in these other strains is unclear . The potential impact of using a vaccine developed against antigens that are encoded by commensal organisms warrants further investigation . Future studies could conceivably quantify the extent to which vaccination with ETEC_2479 , Skp , and MipA protect against other ETEC strains that have caused diarrheal disease in humans . Given the lack of full protection observed using each antigen independently , an additional plan for future work is also to evaluate the efficacy of a cocktail of the three antigens administered simultaneously .
|
Diarrheal disease is an endemic health threat in underdeveloped nations . One of the major causative agents of diarrheal disease is a group of bacteria collectively known as enterotoxigenic Escherichia coli ( ETEC ) . These organisms can cause disease symptoms ranging from mild diarrhea to a more severe , cholera-like form . We were interested in characterizing ETEC proteins that can generate a protective immune response as the first step in identifying potential new vaccine candidates . We used proteomics to identify a subset of ETEC proteins and then characterized this subset for their ability to inhibit ETEC binding to cultured intestinal epithelial cells . We then vaccinated mice with the most promising antigen candidates and were able to identify three proteins that protected mice from clinical signs of disease normally caused by ETEC infection . We suggest that future characterization of these proteins may potentially improve our collective efforts to create safe , effective , and broadly protective ETEC vaccines .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Protective Enterotoxigenic Escherichia coli Antigens in a Murine Intranasal Challenge Model
|
Psoriasis is a common inflammatory skin disease characterized by thickened scaly red plaques . Previously we have performed a genome-wide association study ( GWAS ) on psoriasis with 1 , 359 cases and 1 , 400 controls , which were genotyped for 447 , 249 SNPs . The most significant finding was for SNP rs12191877 , which is in tight linkage disequilibrium with HLA-Cw*0602 , the consensus risk allele for psoriasis . However , it is not known whether there are other psoriasis loci within the MHC in addition to HLA-C . In the present study , we searched for additional susceptibility loci within the human leukocyte antigen ( HLA ) region through in-depth analyses of the GWAS data; then , we followed up our findings in an independent Han Chinese 1 , 139 psoriasis cases and 1 , 132 controls . Using the phased CEPH dataset as a reference , we imputed the HLA-Cw*0602 in all samples with high accuracy . The association of the imputed HLA-Cw*0602 dosage with disease was much stronger than that of the most significantly associated SNP , rs12191877 . Adjusting for HLA-Cw*0602 , there were two remaining association signals: one demonstrated by rs2073048 ( p = 2×10−6 , OR = 0 . 66 ) , located within c6orf10 , a potential downstream effecter of TNF-alpha , and one indicated by rs13437088 ( p = 9×10−6 , OR = 1 . 3 ) , located 30 kb centromeric of HLA-B and 16 kb telomeric of MICA . When HLA-Cw*0602 , rs2073048 , and rs13437088 were all included in a logistic regression model , each of them was significantly associated with disease ( p = 3×10−47 , 6×10−8 , and 3×10−7 , respectively ) . Both putative loci were also significantly associated in the Han Chinese samples after controlling for the imputed HLA-Cw*0602 . A detailed analysis of HLA-B in both populations demonstrated that HLA-B*57 was associated with an increased risk of psoriasis and HLA-B*40 a decreased risk , independently of HLA-Cw*0602 and the C6orf10 locus , suggesting the potential pathogenic involvement of HLA-B . These results demonstrate that there are at least two additional loci within the MHC conferring risk of psoriasis .
Psoriasis ( Ps ) is a relatively common , T cell-mediated , inflammatory skin disease . Ps is typically manifested as thickened scaly red plaques , characterized by epidermal hyperplasia , increased vascularity in the dermis , and infiltration of inflammatory cells into the dermis and epidermis [1] , [2] , [3] . Both genetics and environment play a role in the etiology of Ps . Although its pathogenic mechanism is not completely understood , investigations have strongly suggested that a susceptibility locus ( PSORS1 ) located within the human leukocyte antigen ( HLA ) class I region on the short arm of chromosome 6 , is the major genetic determinant of psoriasis [4] . However , the exact location of the PSORS1 gene had been controversial , due to the extended and complicated linkage disequilibrium ( LD ) pattern of the region [5] . Early studies had indicated the existence of two major PSORS1 locations suggested by various fine mapping studies , one a region ∼150 kb telomeric to HLA-C [6] , [7] , [8] , [9] , harboring HCG27 , POU5F1 , TCF19 , CCHCR1 and CDSN , and the other HLA-C itself or very close [10] , [11] . Recently , a combined sequencing and haplotype mapping study found that within the 298 kb homologous region between the two proposed risk haplotypes , only HLA-C encoded variants unique to these haplotypes at the level of translated protein , which at the same time conferred increased risk of psoriasis , strongly suggesting that HLA-C is the Ps susceptibility gene , and excluded the telomeric region [12] . Similarly , two other family-based association studies , one in a white population and the other in a Chinese population , confirmed the direct involvement of HLA-C in psoriatic susceptibility [13] , [14] . In terms of specific risk alleles , HLA-Cw*0602 has been consistently reported in numerous populations , while the results have been controversial for HLA-Cw*1203 , showing modest positive association with psoriasis in some studies [14] , [15] , or no association [12] , [16] , or even significantly lower frequency in psoriasis patents than in controls in others [17] . Located 85 kb centromeric to HLA-C , HLA-B has also repeatedly exhibited association with psoriasis [18] . However , the over-represented B serotypes , B*13 and B*57 , are in tight LD with HLA-Cw*0602 [19]; therefore , the association of HLA-B is thought to be attributable to HLA-C , or the HLA-Cw*0602 harboring haplotypes 13 . 1 ( Cw6-B13 ) and 57 . 1 ( Cw6-B57 ) , which are named according to the B allele [20] . Likewise , genes located farther from HLA-C at the centromeric end , including TNF-α ( tumor necrosis factor-alpha ) [21] , [22] , AGER ( receptor of advanced glycosylation end product-specific receptor ) [23] , HLA-DRB1 , HLA-DQA1 and HLA-DQB1 [24] , [25] , [26] , [27] , have also been found to be associated with psoriasis . Although some of these genes are as far away from HLA-C as 1 . 38 megabases , the extended haplotype pattern of the HLA region still makes it probable that the association of these genes can be explained by LD with PSORS1 , i . e . HLA-C [11] , [28] . Other findings have suggested the existence of another susceptibility gene in the major histocompatibility complex ( MHC ) in addition to HLA-C . A study found that the octamer transcription factor-3 ( OTF3 , also named POU5F1 ) B allele was more prevalent in patients than in controls , even within the HLA-Cw*0602-negative subset of samples [29] . Moreover , less than 25 kb from this gene , two single nucleotide polymorphisms ( SNPs ) in the SEEK1 ( PSORS1C1 ) gene retained association with psoriasis upon stratification for HLA-Cw*0602 status ( positive/negative ) [30] . However , these analyses did not consider HLA-Cw*1203 , nor did they account for the increased risk associated with HLA-Cw*0602 homozygotes [31] . Thus , analyses conditional on HLA-Cw*0602 only , or upon stratification on HLA-Cw*0602 positive/negative status , may not completely remove the confounding by HLA-C . Moreover , in contrast to the telomeric end , the centromeric end of HLA-C has rarely been investigated conditional on PSORS1 . Recently , within the framework of the Genetic Association Information Network ( GAIN ) we performed a multi-center collaborative genome-wide association study ( GWAS ) , which identified seven Ps susceptibility loci at a genome-wide level of significance [32] . In this study , the most significantly associated SNP was rs12191877 ( p = 3×10−53 ) , which is in strong LD with HLA-Cw*0602 ( r2 = 0 . 63 ) . In addition to this SNP , other MHC SNPs that reached genome-wide significance spanned a 4 Mb region , centering on HLA-C . Considering the widely scattered physical locations of these associated SNPs , the density of immune or inflammation related genes , and the above-mentioned multiple-susceptibility-genes in this region hypothesized to be associated with psoriasis , we searched for other psoriasis loci in the GAIN dataset by examining the MHC region in more detail . First , we used the CEPH phased data as a reference to derive a method of determining the HLA-C genotypes based on the SNPs genotyped as part of the GAIN projects [33] . This was followed by a stepwise search for other susceptibility genes within the MHC region conditional on the predicted HLA-Cw*0602 . The findings from these analyses were then replicated in an independent case-control dataset from a Chinese population . These results demonstrate that within the MHC region , there are at least two susceptibility loci for Ps in addition to HLA-C .
Using the phased HLA and SNP genotypes contained in the HapMap CEU panel and additional CEPH samples [33] as a reference set , we imputed in each GAIN subject the HLA-Cw*0602 and HLA-Cw*1203 genotypes , represented as the predicted number of HLA-Cw*0602 or HLA-Cw*1203 alleles . Since the SNP combinations used in the imputation were in complete linkage disequilibrium with the HLA-C allele of interest in the reference samples , the calculated uncertainties in the imputation only arose from haplotype reconstruction . In all GAIN samples , this uncertainty level was very low , indicated by the small average difference between the imputed and the most likely genotypes ( 0 . 002 and 0 . 0001 for HLA-Cw*0602 and HLA-Cw*1203 , respectively ) . In comparison with true HLA-C genotypes obtained from direct sequencing in a subset of our samples ( n = 420 ) , there were no discrepancies observed for HLA-Cw*0602 , and only 2 for HLA-Cw*1203 , leading to a discordant rate of less than 0 . 5% ( Table 1 ) . In a logistic regression analysis , the imputed HLA-Cw*0602 allele was clearly associated with psoriasis . Both the significance level and the magnitude of association were higher than those observed for the most significant genotyped SNP , rs12191877 ( p = 8×10−61 vs . p = 3×10−53; per allele OR = 3 . 85 [3 . 25–4 . 55] vs . OR = 2 . 92 [2 . 54–3 . 37] ) ( Figure 1A ) . Another suggested risk allele of HLA-C , HLA-Cw*1203 was significantly associated in logistic regression adjusted for HLA-Cw*0602 ( p = 0 . 002 , OR = 1 . 44 [1 . 14–1 . 81] ) , and in an analysis of the HLA-Cw*0602-negative subset of samples ( p = 0 . 004 , OR = 1 . 44 [1 . 12–1 . 85] ) . Imputation of other HLA-C major alleles did not show any association with psoriasis ( data not shown ) . These provide further confirmation that HLA-C is the major susceptibility gene at the PSORS1 locus , and that HLA-Cw*0602 is the allele associated with the highest risk . To examine the association of other loci within the MHC region with psoriasis , we applied logistic regression analyses to all other genotyped SNPs in the MHC region ( from 29 . 3 Mb to 33 . 7 Mb on chromosome 6 ) adjusting for the imputed HLA-Cw*0602 genotype . As expected , given the patterns of linkage disequilibrium across this region , the levels of significance for association of the vast majority of SNPs dropped dramatically ( Figure 1B ) . The top SNP in the unadjusted analyses , rs12191877 , was no longer even nominally significantly associated with Ps after adjustment for HLA-Cw*0602 . However , as seen in Figure 1B , we observed other SNPs that remained significant after Bonferroni correction for multiple testing . The SNP exhibiting the highest significance level , rs2073048 , is located at position 32 . 4 Mb within an open reading frame , C6orf10 , 27 kb telomeric to BTNL2 and 144 kb centromeric to NOTCH4 . The minor allele ( G ) at this SNP had a frequency of 15% in controls and was associated with an adjusted odds ratio of 0 . 66 ( 95% confidence interval = [0 . 56–0 . 78] , p = 2×10−6 ) . To better understand the relationship between this locus and HLA-C , we examined the effect of rs2073048 in a stratified analysis in which strata were defined by carriage of HLA-Cw*0602 . The test of homogeneity of effect between strata showed no evidence of heterogeneity ( p = 0 . 78 ) . In the subset of samples that does not contain an HLA-Cw*0602 allele , rs2073048 was still significantly associated with psoriasis ( OR = 0 . 64 [0 . 53–0 . 78] , p = 1×10−5 ) . These results suggest that the association of this locus with psoriasis is independent of HLA-Cw*0602 . Another cluster of SNPs exhibiting significant p-values after Bonferroni correction in Figure 1B were between HLA-B and MICA . The most significant genotyped SNP of these , rs13437088 , is located 30 kb centromeric of HLA-B and 16 kb telomeric of MICA . The minor allele of this SNP , T , had an allele frequency of 0 . 26 in controls , and was associated with an increased risk of psoriasis ( OR = 1 . 32 [1 . 17–1 . 49] , p = 9×10−6 ) . In analyses adjusted for HLA-Cw*0602 and rs2073048 ( Figure 1C ) , the association of this SNP was even stronger ( OR = 1 . 38 [1 . 22–1 . 57] , p = 3×10−7 ) , suggesting that the observed association is not due to the LD with HLA-C or C6orf10 . In stratified analyses , there was no evidence of heterogeneity when samples were stratified by HLA-Cw*0602 , rs2073048 , or both ( p = 0 . 21 , 0 . 26 , and 0 . 58 , respectively ) , showing that the association of this locus is independent of HLA-C and of the C6orf10 locus . When both of the newly identified putative susceptibility loci and HLA-Cw*0602 were included in a logistic regression model , each of them was significantly associated with disease ( p = 3×10−47 , 6×10−8 and 3×10−7 , for HLA-Cw*0602 , rs2073048 and rs13437088 , respectively ) . Individuals carrying risk genotypes at HLA-C , rs2073048 and rs13437088 were estimated to be at a nine-fold increased risk of psoriasis compared to those carrying low risk genotypes at all three loci ( Table 2 ) . To assess whether HLA-B is responsible for the association of rs13437088 with psoriasis risk , we imputed all HLA-B serotypes with CEPH population frequencies >5% , as well as those serotypes previously suggested to be associated with psoriasis ( B*13 , B*57 and B*58 ) . In a logistic regression analysis adjusted for HLA-Cw*0602 and rs2073048 , two B serotypes ( B*40 and B*57 ) were significantly associated with psoriasis after Bonferroni correction for testing multiple serotypes: One additional copy of B*57 conferred a 1 . 7 fold elevated risk of disease , while B*40 was associated with a 40% reduced risk . In individuals who did not carry HLA-Cw*0602 , B*40 remained significantly associated , while B*57 did not ( Table 3 ) . This can be explained by the fact that B*57 is in tight LD with HLA-Cw*0602 , while B*40 is not in LD with HLA-Cw*0602 [19] . The SNP rs13437088 was in high LD with B*57 ( D' = 1 ) . To further confirm our findings of the two novel loci for psoriasis within MHC , we tested them in an independent Han Chinese sample that was included in another GWAS of psoriasis [34] . Using the Han Chinese and Japanese HapMap data as a reference , HLA-Cw*0602 genotypes were imputed using strategies identical to those for the US sample . The predicted HLA-Cw*0602 allele was strongly associated with psoriasis ( p = 1×10−206 ) , which is quite comparable to the top SNP rs1265181 ( p = 2×10−208 ) [34] , whose genotypes were 99 . 6% identical to the predicted HLA-Cw*0602 , showing that HLA-Cw*0602 is the main susceptibility allele within the PSORS1 region . On the other hand , HLA-Cw*1203 was not associated with psoriasis in this population , after adjustment for HLA-Cw*0602 or within the HLA-Cw*0602-negative subset of samples . To assess the two loci we observed in the US samples , we first tested all genotyped SNPs within 100 kb of rs2073048 , the first locus we identified in the US study . The results indicated several SNPs that were significantly associated , the most significant of which was the SNP rs28732201 , which had a minor allele frequency of 0 . 01 , and odds ratio of 2 . 85 [1 . 81–4 . 50] with a p-value of 7×10−6 ( Figure 1D ) . This SNP is located between C6orf10 and BTNL2 , 11 kb upstream to the transcription start site of C6orf10 , and 12 kb downstream to the transcription end site of BTNL2 . After adjustment for both HLA-Cw*0602 and rs28732201 , the locus of HLA-B/MICA ( within 100 kb of the SNP identified in the GAIN dataset ) also exhibited significant association , shown by the SNP rs2442719 ( OR = 1 . 66 [1 . 36–2 . 03] , p = 8×10−7 , Figure 1E ) , located only 1 kb from the telomeric end of HLA-B . When HLA-Cw*0602 , rs28732201 and rs2442719 were all included in a logistic regression model , they all remained significantly associated ( p = 2×10−102 , 1×10−5 , and 8×10−7 , respectively ) . We also imputed the B*40 and B*57 serotypes in these Chinese samples and tested their associations with psoriasis controlling for HLA-Cw*0602 . Interestingly , similar to the US data , B*57 was significantly associated with an increased risk of psoriasis ( OR = 2 . 71 [1 . 81–4 . 06] , p = 1×10−6 ) , and B*40 with a reduced risk ( OR = 0 . 74 [0 . 57–0 . 97] , p = 0 . 03 ) . These associations remained nominally significant when analyses were further adjusted for the C6orf10 locus ( OR = 0 . 75 [0 . 58–0 . 98] , p = 0 . 04 for B*40 and OR = 1 . 98 [1 . 05–3 . 73] , p = 0 . 03 for B*57 ) .
The association of HLA-C with psoriasis was first proposed as early as 30 years ago [35] . However , until recently doubts remained as to whether HLA-C or a nearby gene was the locus responsible for the observed association . One of the difficulties contributing to this is the fact that the HLA region has a complicated and extended linkage disequilibrium pattern , while harboring many immune response genes in high density . Recently , sequencing and haplotype analyses studies have concluded that HLA-C is the major risk determinant of psoriasis within the HLA region , and HLA-Cw*0602 is the main risk allele . However , one immediate question arose: is HLA-C the only susceptibility gene in this region ? This question turns out to be challenging because any other putative psoriasis predisposing genes would have a weaker effect than HLA-Cw*0602 , and the analyses would be complicated by potential linkage disequilibrium with HLA-C . To control for the effects of HLA-C , it would be optimal to use the HLA-C genotype per se; however , in a large multi-center study such as the present one , molecular typing of the HLA alleles would not be readily available . In this paper , we used the CEPH phased data which contains the phased alleles at HLA-A , -B , -C –DR and –DQ as well as thousands of SNPs in this region as a reference sample to identify SNPs that can accurately predict the HLA-C alleles . These genotypes can then be used to perform analyses to search for other susceptibility genes . As our validation study illustrated , we were able to accurately predict the HLA-C alleles in the GAIN samples , in spite of it being composed of a diverse ( though all white ) mixture of individuals of differing European backgrounds . In addition to the present study , others have found similar utility of this approach across European-derived populations . For example , a validation assessment of the imputation method carried out on Dutch , UK , Spanish , and Italian samples showed high sensitivity and specificity in imputing DQA1 , DQB1 and DRB1 alleles [33] , [36] . Moreover , the genomic control [37] parameter of our samples was 1 . 03 , suggesting that population stratification has negligible impact on our association results [32] . Through this imputation , we confirmed that HLA-Cw*0602 is the major psoriasis risk determinant within the HLA region , which has a much stronger association with psoriasis than the most significantly associated SNP , rs12191877 . This has reinforced the importance of using HLA-C risk allele per se in the analyses of MHC , since adjusting for the surrogate SNP cannot fully control for HLA-C . Our analyses also provided further evidence that HLA-Cw*1203 is associated with psoriasis , although it is not clear whether HLA-Cw*1203 is a risk allele itself , or is in LD with the risk allele of another susceptibility gene near HLA-C . On the other hand , HLA-Cw*1203 was not associated with psoriasis in the Han Chinese after adjustment for HLA-Cw*0602 . These may imply that HLA-Cw*1203 does not confer risk to psoriasis; the discrepancies in its association with psoriasis in different studies may be due to the different LD patterns among populations . To search for additional loci for psoriasis in the MHC region , we conducted logistic regression analyses adjusting for the imputed HLA-Cw*0602 , and identified two loci within 1 . 2 Mb of HLA-C , one around the C6orf10 gene , and one between HLA-B and MICA . Both of these loci were significantly associated with disease risk after Bonferroni correction for the number of SNPs considered in the analyses . Furthermore , data analyses demonstrate that they are not simply reflecting linkage disequilibrium with HLA-C , since 1 ) the associations are not secondary to HLA-Cw*0602 as shown in the analyses adjusted for HLA-Cw*0602 or within the HLA-Cw*0602-negative subset of samples , 2 ) further adjustment for HLA-Cw*1203 in analyses did not substantially change our results , and 3 ) no other HLA-C major allele showed association with psoriasis in our data . More importantly , both associated loci were replicated in an independent Han Chinese dataset after adjustment for HLA-Cw*0602 , even though the LD patterns in Chinese are quite different from those in the GAIN dataset . All these observations imply the existence of other psoriasis risk-determining genes within the MHC . When the putative susceptibility loci and HLA-Cw*0602 were included in a logistic regression model , they all remained significantly associated with disease , showing that these loci are not attributable to each other; therefore , within the MHC there are at least three genes conferring risk of psoriasis . The first locus we identified is located 1 . 1 Mb centromeric of HLA-C , indicated by the SNP rs2073048 . It is noteworthy that in our previous paper , using a forward selection technique , we also found some evidence of another SNP ( rs2022544 ) close to this locus that was associated with Ps , but this study did not control for the HLA-C risk allele , but rather for its surrogate SNP [32] , rendering the results subject to residual association of HLA-C . The SNP rs2073048 is located within the 4th intron of an open reading frame C6orf10 , 27 kb telemetric to BTNL2 and 144 kb centromeric to NOTCH4 . There have been previous reports of association of psoriasis with AGER , which is 183 kb telomeric of rs2073048 , and with HLA-DRB1 that is 211 kb centromeric of rs2073048 . In the present study , analyses of imputed serotypes of HLA-DRB1 did not support the involvement of HLA-DRB1 in disease pathogenesis ( data not shown ) . Furthermore , as noted before , there is a recombination hot spot centromeric to NOTCH4 [38] , reducing the possibility that the association at this locus is attributable to those genes located telomeric to the recombination hot spot ( NOTCH4 , AGER , etc . ) , although they still cannot be completely excluded . Thus , C6orf10 and BTNL2 are better candidate genes for this locus . The most associated SNP found in the GAIN dataset ( rs2073048 ) after correction for HLA-Cw*0602 is located within C6orf10 , and the SNP found in the Chinese dataset ( rs28732201 ) is close to the transcription start site of C6orf10; therefore , C6orf10 is one of the most important candidate genes at this locus . It has been observed that the transcription of C6orf10 in keratinocytes can be triggered by TNF-α ( Gene Expression Omnibus dataset number: GDS1289 ) [39] , an important proinflammatory cytokine in the pathogenesis of psoriasis , although the function of the C6orf10 product is not known . Nevertheless , other genes cannot be excluded; more haplotype and sequencing analyses will be required to pinpoint the risk-conferring variants at this locus . The second locus we observed after adjustment for HLA-C and C6orf10 is between HLA-B and MICA , located 117 kb centromeric to HLA-C , suggesting the potential association of HLA-B or MICA with psoriasis risk . In the Chinese samples , this locus was also indicated by logistic regression analyses adjusted for HLA-C and C6orf10 , by a SNP located only 1 kb telomeric to HLA-B . More importantly , although the linkage disequilibrium patterns and the HLA-B tagging SNPs were different between Chinese and white populations , the same HLA-B serotypes were associated with psoriasis: B*57 with an increased risk and B*40 with a reduced risk . These are suggestive of the involvement of HLA-B in psoriasis etiology . The role of HLA-B in psoriasis immuno-pathogenesis might be similar to that of HLA-C , which has been shown to bind peptide motifs that are shared between the streptococcal M proteins and the wound-healing-associated keratins k16 and k17 , thereby clonally expanding the pool of skin-directed autoreactive CD8+ T cells [40] . Another candidate gene for this locus , MICA , is a distant homolog of the MHC class I protein . It can be induced by cellular or metabolic stress in the epithelia , acting as ligands for the activatory T-cell receptor , NKG2D . In psoriasis , it has been shown that MICA is down-regulated in lesional skin compared with non-lesional skin ( p = 0 . 007 , Gene Expression Omnibus dataset number: GDS2518 ) [41] . The under-expression of the MICA protein might allow the unwanted cells to escape the cytolysis by NK or CD8+ T-cells , resulting in keratinocyte proliferation and the enhanced inflammation inherent to lesions of psoriasis . Other genes within this region still cannot be excluded by our analyses; more detailed studies of HLA-B serotypes and MHC haplotypes are required to further elucidate the association of this locus with psoriasis . The study of associations in the MHC region is notoriously difficult due to the presence of many genes involved in immune and inflammatory processes as well as the extensive and complex patterns of linkage disequilibrium . Our use of a genome-wide panel of SNPs that included nearly 2000 SNPs within the MHC , a validated prediction method to determine with high probability the presence of known HLA-C risk alleles for Ps , and a large sample of psoriasis cases and controls allowed us to begin to tease out different effects on psoriasis risk within the MHC region . Our discoveries are replicated in independent samples from another race , reinforcing the evidence of our findings . In combination with the loci reported in our previous work [32] , [42] , and those yet to be identified from large-scale replication studies of thousands of loci arising from our own and other genome-wide association studies , we anticipate that substantial progress will be made in the coming months in explaining the genetic basis of psoriasis . Perhaps more relevantly , we anticipate that some of the genes identified will prove to be attractive therapeutic targets , leading to improved treatment for this disease . In conclusion , we provide evidence that two loci within the HLA region in addition to HLA-C , one near C6orf10 and one near HLA-B , are significantly associated with psoriasis , suggesting that within MHC there are at least three genes moderating susceptibility to psoriasis . However we fully recognize that additional studies including re-sequencing and detailed haplotype analysis will be required to elucidate the causal variants .
The initial genome wide association scan involved 1409 psoriasis patients and 1436 controls recruited from the University of Utah , the University of Michigan , and the Washington University at St . Louis , USA . All cases and controls were of European descent . Informed consent was obtained from each participant . In total , 1359 cases and 1400 controls with 447 , 249 SNP genotypes passed the quality control process . The average age at onset of psoriasis was 24 . 3 years with the majority of patients ( 1127 , 82 . 9% ) developing psoriasis before age 40 . Additional details on subject characteristics and recruitment can be found in Nair et al [32] . The samples in the replication analyses were 1139 cases and 1132 controls used in the initial GWAS of psoriasis conducted at the Anhui Medical University , Hefei city , Anhui province , China . These samples were recruited from Han Chinese populations by multiple hospitals in China . More details about the studied samples are described elsewhere [34] . The local institutional review board at each site approved the study protocol . We used the phased HLA and SNP genotypes contained in the HapMap CEU panel ( 30 trios ) and an additional set of 90 CEPH samples [33] to search for SNP combinations in linkage disequilibrium with specific HLA alleles , using an approach similar to that taken by de Bakker et al [33] , except that whenever possible , a combination of 3 to 4 SNPs was used . In each of our GAIN samples , we inferred the haplotypes of these chosen SNPs and the corresponding haplotype probabilities using the PHASE program version 2 . 1 [43] , [44] , for case and control populations separately as suggested by Mensah et al [45] . The imputed HLA genotype containing an allele of interest was represented as the estimated number of copies of each specific allele , and was calculated by summing the probability of having that allele given a specific haplotype , weighted by the corresponding haplotype probability:where g is the imputed number of HLA alleles , Ai• denotes an event that the haplotype i• contains the HLA allele of interest , hi• is one of the two haplotypes of the haplotype assignment i from PHASE , and p ( hi ) is the haplotype probability . The p ( Ai•|hi ) was directly obtained from the reference samples , and the p ( hi ) was calculated by the PHASE program . Thus , the uncertainties from the incomplete LD between the haplotype and the HLA allele , and from the haplotype reconstruction , were both integrated into the imputation; although not all uncertainty can be estimated due to the relatively small size of the HapMap and CEPH sample . To gain more power in association test , we minimized the overall uncertainty level ( estimated by the averaged difference between the imputed and the most likely genotype ) , by maximizing the LD between the haplotype and the HLA allele , and by maximizing the haplotype probabilities by inclusion of nearby SNPs in low LD with the HLA imputing SNPs in haplotype reconstruction . These additional SNPs were selected by the HAPLOVIEW program [46] , with a threshold of r2<0 . 1 . For better imputation accuracies , a more stringent quality control strategy than the GWAS was applied , where those SNPs with <99 . 5% genotype call rates , or with evidence for departure from Hardy-Weinberg equilibrium at p<0 . 001 among controls , were not considered . To validate the imputed HLA-C genotypes , a sub-sample of 420 Utah psoriasis patients were genotyped using direct sequencing at Atria Inc . ( South San Francisco , CA ) . e conducted trend tests to assess the association between SNPs and psoriasis , using the PLINK program [47] . To examine the association of a specific HLA allele with psoriasis , we used logistic regression analysis on the imputed HLA genotype , weighted according to genotype probabilities as suggested by Mensah et al [45] , i . e . , in our analyses of HLA , we used the imputed number of alleles ( real number between 0 and 2 ) , rather than the most likely number of alleles ( 0 , 1 or 2 ) in logistic regression . To examine other SNPs within the HLA region , logistic regression was performed adjusted for HLA-C and associated SNP genotype , using the PLINK program [47] . Averaged odds ratio and the corresponding 95% confidence intervals for each additional number of minor allele of the studied SNP were calculated . Linkage disequilibrium plots and recombination rate plots were produced using the HapMap phase II data [48] by a C++ program written by the authors .
|
Psoriasis ( Ps ) is a chronic inflammatory disease of the skin , affecting approximately 2% of Europeans . The HLA-C gene , located within the major histocompatibility complex ( MHC ) region on chromosome 6 , is the major genetic determinant of psoriasis . However , multiple susceptibility genes within MHC are also hypothesized . Recently , we carried out a genome-wide association scan ( GWAS ) on psoriasis with 1 , 359 patients and 1 , 400 healthy controls , which identified seven psoriasis loci in the human genome and confirmed the effect of HLA-C . This dataset contains densely distributed genetic variations , single nucleotide polymorphisms ( SNPs ) , which were then further analyzed in search for additional susceptibility genes within the MHC region . Using the SNP data , we imputed in all samples the HLA-C risk allele with high accuracy . Adjusting for the HLA-C , two additional loci , one near C6orf10 and one near HLA-B/MICA , have significant associations with psoriasis , which were also observed in an independent Han Chinese dataset , suggesting that within the MHC there are at least three genes moderating susceptibility to psoriasis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/gene",
"discovery",
"dermatology/psoriasis",
"and",
"other",
"inflammatory",
"diseases",
"genetics",
"and",
"genomics/complex",
"traits",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"genetics",
"and",
"genomics/medical",
"genetics"
] |
2009
|
Multiple Loci within the Major Histocompatibility Complex Confer Risk of Psoriasis
|
One-third of the human population is infected with parasitic worms . To avoid being eliminated , these parasites actively dampen the immune response of their hosts . This immune modulation also suppresses immune responses to third-party antigens such as vaccines . Here , we used Litomosoides sigmodontis-infected BALB/c mice to analyse nematode-induced interference with vaccination . Chronic nematode infection led to complete suppression of the humoral response to thymus-dependent vaccination . Thereby the numbers of antigen-specific B cells as well as the serum immunoglobulin ( Ig ) G titres were reduced . TH2-associated IgG1 and TH1-associated IgG2 responses were both suppressed . Thus , nematode infection did not bias responses towards a TH2 response , but interfered with Ig responses in general . We provide evidence that this suppression indirectly targeted B cells via accessory T cells as number and frequency of vaccine-induced follicular B helper T cells were reduced . Moreover , vaccination using model antigens that stimulate Ig response independently of T helper cells was functional in nematode-infected mice . Using depletion experiments , we show that CD4+Foxp3+ regulatory T cells did not mediate the suppression of Ig response during chronic nematode infection . Suppression was induced by fourth stage larvae , immature adults and mature adults , and increased with the duration of the infection . By contrast , isolated microfilariae increased IgG2a responses to vaccination . This pro-inflammatory effect of microfilariae was overruled by the simultaneous presence of adults . Strikingly , a reduced humoral response was still observed if vaccination was performed more than 16 weeks after termination of L . sigmodontis infection . In summary , our results suggest that vaccination may not only fail in helminth-infected individuals , but also in individuals with a history of previous helminth infections .
More than 1 billion people are infected with helminths worldwide , predominantly in the tropics and subtropics [1] . To avoid their elimination and to limit pathology , helminths have developed sophisticated strategies to dampen the immune response of their hosts [2] , [3] . This helminth-induced immune suppression also affects the immune response to third-party antigens and thus may interfere with efficient response to co-infecting pathogens and to vaccination [4] . Reduced cellular and humoral responses have been observed in helminth-infected humans after cholera [5] , Bacillus Calmette-Guerin ( BCG ) [6] , [7] , tetanus toxoid [8] , [9] , [10] , [11] , [12] and anti-Plasmodium falciparum vaccination [13] ( reviewed in [14] , [15] , [16] ) . Drug-induced termination of helminth infection improved responses to BCG [6] , [17] and cholera [5] , [18] vaccination . These collective studies emphasize that in addition to the pathology caused by helminth infection itself , helminth-induced interference with vaccination efficacy represent a global health problem . As the nature of human studies is predominantly descriptive , murine models of helminth infection have been established to further investigate these issues . Responses to HIV or BCG vaccination were compromised in mice infected with the pathogenic trematode Schistosoma mansoni [19] , [20]; similarly infection with the gastrointestinal nematode Heligmosomoides polygyrus interfered with humoral and cellular responses to malaria vaccinations [21] , [22] . The nematode Litomosoides sigmodontis is used to model human filarial infections [23] , [24] . L . sigmodontis third stage larvae ( L3 ) are transmitted during the blood meal by the arthropod intermediate host , the mite Ornithonyssus bacoti . The natural definitive host is the cotton rat Sigmodon hispidus; however , some laboratory mouse strains such as BALB/c mice are fully susceptible to infection . L3 migrate during the first three days via the lymphatic system to the thoracic cavity . They moult to fourth stage larvae ( L4 ) within 10 days and to immature adults within 30 days . Mature adults mate around day 55 post-infection ( p . i . ) and females release first stage larvae known as microfilariae ( MF ) into the peripheral circulation by day 60 p . i . Although BALB/c mice eventually control parasite burden by innate and adaptive immunity , they remain infected for several months until the parasites are fully eradicated . Thus , L . sigmodontis-infected BALB/c mice provide a suitable model to analyse the impact of chronic nematode infections on co-infections [25] , [26] , autoimmune diseases [27] , [28] and allergy [29] . We used L . sigmodontis infection to study the effect of concurrent nematode infection on vaccination efficacy . An experimental vaccination against the liver stage of Plasmodium berghei using a single injection of a P . berghei circumsporozoite ( CSP ) fusion protein induced lower numbers of CSP-specific CD8+ cytotoxic T cells , if vaccinated mice were infected with L . sigmodontis [30] . Using semi-resistant C57BL/6 mice we have shown that acute L . sigmodontis infection suppressed humoral responses to thymus-dependent ( TD ) model antigen vaccinations performed at day 14 of infection [31] . Since C57BL/6 mice terminate L . sigmodontis infection around day 60 before patency is established [32] the impact of chronic nematode infection can not be modelled using this mouse strain . In the current study we use the fully susceptible BALB/c mice to analyse the impact of different L . sigmodontis life stages on vaccination efficacy . We report an almost complete suppression of IgG response to TD vaccination in chronically infected mice . Suppression was induced by L4 and by parasitic adults that outcompeted the pro-inflammatory effect stimulated by MF . Suppression was observed even when vaccination was performed several months after termination of L . sigmodontis infection . We provide evidence that suppression of TD response during chronic nematode infection was established by interference with follicular B helper T cell ( TFH ) induction , independent of Foxp3+ Treg .
Animal experimentation was conducted at the animal facility of the Bernhard Nocht Institute for Tropical Medicine ( BNITM ) in agreement with the German animal protection law under the supervision of a veterinarian . The experimental protocols have been reviewed and approved by the responsible Federal Health Authorities of the State of Hamburg , Germany , the “Behörde für Gesundheit und Verbraucherschutz” permission number 98/11 . BALB/c mice , cotton rats ( S . hispidus ) and BALB/c “Depletion of Regulatory T cell” ( DEREG ) mice were bred in the animal facilities of the BNITM and kept in individually ventilated cages under specific pathogen-free conditions . Mice were sacrificed by deep CO2 narcosis . The life cycle of L . sigmodontis was maintained as described [31] . Six- to eight-week-old female mice were naturally infected by exposure to L . sigmodontis-infected mites ( O . bacoti ) that transmit infectious L3 during the blood meal . L4 , adult worms and granulomata were harvested by flushing the thoracic cavity of infected mice with 8–10 mL of PBS . Parasites were counted subsequently . To detect MF in the circulation , blood of infected mice was collected in EDTA tubes; subsequently 20 µL of blood was added to 100 µL of ddH2O , and centrifuged at 10 , 000× g for 5 min . The pellet was resolved in 20 µL of Gentian violet and all MF were counted . L . sigmodontis MF were purified from blood of infected cotton rats by density gradient centrifugation on Percoll . EDTA blood of cotton rats was collected and diluted 1∶2 with PBS . Iso-osmotic Percoll ( Sigma-Aldrich , Munich , Germany ) was prepared by mixing 9 parts of Percoll ( density , 1 . 130 g/mL ) with 1 part of 2 . 5 M Sucrose ( Sigma-Aldrich , Munich , Germany ) . The following dilutions of 90% Percoll in 0 . 25 M Sucrose were made: 25% and 30% and layered with the diluted blood on top . After centrifugation at 400× g for 30 min at room temperature ( RT ) without brakes , MF are located between the 25% and 30% layer . MF were harvested , washed twice with PBS by 30 min centrifugation at 400× g and counted . Mice received 10 , 000 viable MF i . v . Non-infected and L . sigmodontis-infected mice were vaccinated at indicated time points post infection by i . p . injection of either 100 µg alum-precipitated dinitrophenol-keyhole limpet hemocyanin ( DNP-KLH , Sigma-Aldrich , Munich , Germany ) , 100 µg 4-hydroxy-3-iodo-5-nitrophenylacetyl ( NIP ) conjugated to Ficoll ( NIP-Ficoll ) ( Biosearch Technologies , Navato , USA ) , or by s . c . injection of 30 µg alum-precipitated DNP-KLH into the hind footpad . For analysis of serum antibodies , blood was collected from mice by submandibular bleeding of the facial vein 7 , 14 and 21 days after vaccination and allowed to coagulate for 1 h at RT . Serum was collected after centrifugation at 10 , 000× g for 10 min at RT and stored at −20°C for further analysis . For analysis of spatial separated cellular responses , mice were sacrificed at the indicated time point , and spleen and popliteal lymph nodes ( popLN ) were dissected . A total of 2 . 5×105 splenocytes or popLN cells were cultured in 3–5 replicates in 96-well round-bottom plates in RPMI 1640 medium supplemented with 10% FCS , 20 mM HEPES , 2 mM L-glutamine and gentamicin ( 50 µg/mL ) at 37°C and 5% CO2 . The supernatant was harvested after 21 days of culture and DNP- and L . sigmodontis-specific IgG1 , IgG2a and IgG2b were quantified by ELISA . DNP-KLH was chosen as TD model antigen as no cross-reaction between DNP-specific Ig and L . sigmodontis antigen was detected and comparison of DNP7-BSA and DNP38-BSA allows detection of high affinity only as well as high and low affinity Ig . For the detection of DNP- , NIP- and L . sigmodontis-specific Ig , ELISA plates were coated overnight with 1 µg/mL DNP7-BSA , ( Sigma-Aldrich , Munich , Germany ) , 1 µg/mL NIP7-BSA ( Biosearch Technologies , Navato , USA ) , or 4 µg/mL L . sigmodontis extract in carbonate buffer pH 9 . 6 . Low affinity DNP-specific IgG was detected by coating ELISA plates with 1 µg/mL DNP38-BSA ( Biotrend Chemikalien , Cologne , Germany ) . Plates were washed , blocked by incubation with PBS 1% BSA for 2 h and incubated for 2 h with serum or cell culture supernatant . Plates were washed and incubated for 1 h with horseradish peroxidase-labelled anti-mouse IgM , IgG1 , IgG2a , IgG2b ( Life Technologies , Carlsbad , USA ) , or IgG3 ( Southern Biotechnology Associates , Birmingham , USA ) . Plates were washed and developed by incubation with 100 µL tetramethylbenzidine ( 0 . 1 mg/mL ) , 0 . 003% H2O2 in 100 mM NaH2PO4 pH 5 . 5 for 2 . 5 min . Reaction was stopped by addition of 25 µL 2 M H2SO4 , and OD450 was measured . For the more abundantly produced isotypes IgG1 and IgM , titres were calculated by defining the highest serum dilution in a serial dilution ( 1∶1000 to 1∶128 , 000 ) resulting in an OD450 above the doubled background . For the less abundant isotypes IgG2a , IgG2b , and IgG3 , arbitrary units were calculated by subtraction of OD450 of the background from OD450 of one fixed serum concentration ( 1∶100 for IgG2a , 1∶1000 for IgG2b , and 1∶100 for IgG3 ) . Background was generally below OD450 = 0 . 1 . Calculation of titres by serial dilution for random control samples revealed similar differences in DNP-specific IgG2a and IgG2b production as the arbitrary units and thus did not yield additional information ( data not shown ) . For analysis of B and TFH cells , mice were sacrificed at the indicated time points and popLN were dissected . Cells ( 1×106 ) were stained with Live/Dead Fixable Blue Dead Cell Stain Kit ( Life Technologies , Carlsbad , USA ) according to the manufacturer's instructions . For surface staining , cells were stained with anti-CXCR5-Biotin ( clone: 2G8 ) for 30 min at 37°C or with anti-IgG-phycoerythrin-Cy7 ( PE-Cy7 ) ( clone: Poly4053 ) and Biotin-labelled peanut agglutinin ( PNA ) ( Galab Technologies , Geesthacht , Germany ) for 30 min at 4°C . After washing the cells were stained with anti-CD3e-allophycocyanin ( APC ) ( 145-2C11 ) , anti-CD4-APC/Brilliant Violet 510 ( BV510 ) /PE ( clone: RM4-5 ) , anti-CD44-BV421 ( clone: IM7 ) , anti-PD1-fluorescein isothiocyanate ( FITC ) ( clone: 29F . 1A12 ) , anti-ICOS-PE ( clone: 7E . 17G9 ) , anti-CD19-PE ( clone: 6D5 ) , DNP-BSA-AF647 , anti-IgM-FITC ( clone: DS1 ) , Strepavidin-APC and Strepavidin-BV421 for 30 min at 4°C . Foxp3 expression was determined using PE-anti-mouse Foxp3 Staining Set ( clone: FJK-16S , Affymetrix eBioscience , Frankfurt , Germany ) according to the manufacturer's instructions . Foxp3+ Treg depletion was controlled by analysis of eGFP , Foxp3 and CD4 expression . Samples were analysed on a LSR II Flow Cytometer ( Becton Dickinson , Mountain View , USA ) using FlowJo software ( TreeStar , Ashland , USA ) . DNP-BSA was fluorescence labelled using Alexa Fluor 647 Protein Labeling Kit ( Life Technologies , Carlsbad , USA ) according to the manufacturer's instructions . Unless otherwise stated all staining antibodies were purchased from BioLegend ( Fell , Germany ) , BD Biosciences ( Heidelberg , Germany ) or Affymetrix eBioscience ( Frankfurt , Germany ) . Heterozygous BALB/c DEREG mice and non-transgenic littermate control BALB/c mice received 0 . 5 µg diphtheria toxin ( DT ) ( Merck , Darmstadt , Germany ) dissolved in PBS i . p . on three consecutive days , starting either two days prior to L . sigmodontis infection or one day before vaccination . Successful depletion of Foxp3+ Treg was routinely controlled by staining for CD4+Foxp3+ cells in the peripheral blood after the third DT treatment . Statistical analysis was performed by ANOVA with Bonferroni post-test or student's t-test using Prism software ( GraphPad Software , San Diego , USA ) . Results are presented as mean ± SEM; p≤0 . 05 was considered statistically significant .
We used BALB/c mice that are fully susceptible for L . sigmodontis infection to analyse the impact of different parasitic life stages on vaccination efficacy . By choosing different durations between exposure to infected mites and subsequent vaccination we defined the life stage of L . sigmodontis present at the moment of vaccination ( Figure 1A ) . We analysed the impact of L3 migrating via the lymphatic vessels to the thoracic cavity by vaccinating mice at the day of infection . To analyse the impact of L4 and young immature adults on vaccination efficacy , we vaccinated 14 and 30 days after infection . Vaccination at day 60 p . i . resulted in the presence of mature adults that had mated and released MF into the peripheral circulation . We employed alum-precipitated DNP-KLH as a model vaccine for TD humoral response . High affinity DNP-specific IgG responses were analysed in nematode-infected and in age-matched non-infected control mice on three consecutive weeks following vaccination . Simultaneous nematode infection increased early responses to DNP-KLH vaccination ( Figure 1B ) . In contrast , presence of L . sigmodontis L4 during vaccination reduced DNP-specific IgG1 , whereas DNP-specific IgG2a was reduced by trend and IgG2b was unchanged ( Figure 1C ) . Vaccination at later time points of infection , i . e . day 30 or day 60 p . i . , resulted in statistically significant suppression of IgG1 , IgG2a , and IgG2b responses to DNP-KLH , in comparison to age-matched non-infected mice ( Figure 1D , E ) . DNP-specific IgG2 response was almost absent in day 60 infected mice . The capture agent used in these experiments , i . e . BSA coupled to 7 DNP molecules , was suited to detect specifically high affinity IgG . Re-analysis of these sera with BSA coupled to 38 DNP molecules , a setting that will capture IgG displaying a lower affinity to DNP in addition to high affinity DNP-specific IgG , revealed similar results ( Figure S1 in Text S1 ) . Thus , the quantity but not the quality of IgG response to vaccination was reduced by established nematode infection . In line with our previous results [31] , [33] , nematode infection did not interfere with the humoral response to the polyvalent TI model antigen NIP-Ficoll that activates B cells in the absence of T helper cells by strong crosslinking of the B cell receptor . Due to absent T cell co-stimulation NIP-Ficoll induces predominantly IgM responses and limited IgG1 and IgG3 responses [34] . Consequently we did not detect NIP-specific IgG2a or IgG2b in NIP-Ficoll vaccinated mice ( data not shown ) but NIP-specific IgM , IgG1 and IgG3 were produced ( Figure 2 ) . Thereby , NIP-specific humoral responses were similar , or in the case of IgG3 , even increased in NIP-Ficoll-vaccinated non-infected in comparison to day 60 L . sigmodontis-infected mice . Taken together , these results show that the presence of L4 and adult L . sigmodontis , but not of recently transmitted L3 , suppressed humoral response to vaccination specifically in T cell-dependent settings . Intensity of suppression was positively correlated to duration of nematode infection . Chronic nematode infection suppressed both , TH2-associated IgG1 and TH1-associated IgG2 responses to vaccination , thus , inflicting generalized suppression and not polarization towards a type 2 immune response . Despite the apparent suppression of TD humoral response in infected mice , L . sigmodontis-specific IgG1 , IgG2a and IgG2b responses were detectable during infection ( Figure S3 in Text S1 ) . Additional DNP-KLH vaccination did not modulate the L . sigmodontis-specific Ig response , as expected ( Figure S3 in Text S1: black circles and grey squares ) . Between 40 and 60% of infected BALB/c mice developed detectable microfilaraemia by day 60 p . i . , leading to simultaneous presence of two different life stages in mice vaccinated at this late time point . Stratification of DNP-specific IgG response of day 60 infected mice that were positive ( n = 7 ) or negative ( n = 12 ) for MF in the peripheral circulation revealed no differences in suppression ( Figure S2 in Text S1 ) . To differentiate between the impact of adults and MF on response to DNP-KLH vaccination , we injected 10 , 000 purified MF at the day of vaccination to model the recent release of MF by females . Interestingly , presence of isolated MF increased the IgG2a response to vaccination while the IgG1 and IgG2b responses remained unchanged ( Figure 3A ) . Thus , MF displayed a pro-inflammatory effect , increasing TH1-associated Ig responses to third-party antigens . This finding suggests that the anti-inflammatory effect observed in day 60 L . sigmodontis-infected microfilaraemic mice was induced by adults and outcompeted the pronounced pro-inflammatory effect of isolated MF . However , injection of 10 , 000 MF in a bolus may trigger stronger pro-inflammatory signals than the putative effects mediated by MF released gradually by female adults in vivo . To investigate if adults would also suppress the possibly stronger pro-inflammatory signals delivered by isolated MF we injected purified MF into day 60 L . sigmodontis-infected mice ( Figure 3B ) . While injection of MF into non-infected mice increased IgG2a responses compared to naïve mice as observed before , MF-treated L . sigmodontis-infected mice displayed reduced IgG2a responses in comparison to MF-treated non-infected mice . IgG response in MF-treated L . sigmodontis-infected mice was also significantly lower than IgG response in non-infected mice . Taken together , these results show that adults suppress the IgG response to vaccination in the presence of circulating MF , despite pro-inflammatory stimuli transduced by MF . Foxp3+ Treg are central regulators of adaptive immune responses and have been shown to mediate helminth-induced immune suppression [35] . Although our previous study did not indicate a function for Foxp3+ Treg in the suppression of CD4+ T cell proliferation during acute L . sigmodontis infection in C57BL/6 mice [31] , accumulating evidence suggests that the dominance of Treg-mediated regulation differs in different mouse strains [36] , [37] , [38] , [39] . Therefore we evaluated the contribution of Foxp3+ Treg to the suppression of vaccination efficacy in day 60 L . sigmodontis-infected BALB/c mice . To this end , we employed BALB/c DEREG mice that express a fusion protein consisting out of the human diphtheria toxin ( DT ) receptor and enhanced green fluorescent protein ( eGFP ) under the control of the Foxp3 promoter [40] . Injection of DT results into transient depletion of Foxp3+CD4+ T cells in DEREG mice while Treg frequencies in non-transgenic littermates remain unchanged [41] . As Foxp3+ Treg depletion is not permanent in DEREG mice , we investigated the effect of transient Foxp3+ Treg depletion either during initial infection ( Figure 4A ) or during vaccination of chronically infected mice ( Figure 4D ) . Absence of Foxp3+ Treg during the first days of L . sigmodontis infection did not abrogate suppression of IgG response to vaccination in nematode-infected BALB/c mice ( Figure 4C ) . DNP-specific IgG1 , IgG2a and IgG2b responses were equally reduced in nematode-infected mice containing Foxp3+ Treg ( black squares ) or not containing Foxp3+ Treg ( blue squares ) . Successful depletion of Foxp3+ Treg was verified by flow cytometry at the day of L . sigmodontis infection ( Figure 4B ) . Transient Treg depletion at the moment of vaccination ( Figure 4D ) that was confirmed by flow cytometry one day after vaccination ( Figure 4E ) increased the DNP-specific IgG2a response in non-infected mice ( Figure 4F ) . Similar increases in pro-inflammatory IgG2a responses upon Treg depletion were observed in a recent study using DEREG mice in a model of atopic dermatitis [42] . Increased IgG2a responses reflected most likely inefficient regulation due to the absence of Treg and , thus function as internal control for depletion efficacy . Nevertheless , increased DNP-specific IgG2a in Treg-depleted mice was still suppressed upon L . sigmodontis infection . Treg depletion during vaccination did not modulate the more abundant IgG1 or IgG2b responses and did not abrogate nematode-induced suppression of IgG response to vaccination . Taken together , these results rule out a contribution of Foxp3+ Treg to the suppression of IgG response to TD vaccination in nematode-infected BALB/c mice . Natural infection with L . sigmodontis that predominantly dwell in the thoracic cavity induced a systemic immune response . Nematode-specific T and B cell responses are detectable in the draining lymph nodes and in the spleen ( data not shown ) . In the experiments performed so far , mice were vaccinated i . p . , thereby inducing systemic responses to DNP-KLH that are initiated mostly in the spleen . To rule out that suppression of IgG response to vaccination was caused by a simple competition of nematode- and vaccine-specific B and T cells in the same lymphatic organ , we separated the sites of L . sigmodontis-specific and DNP-KLH-specific immune responses ( Figure 5 ) . To this end we vaccinated day 60 L . sigmodontis-infected and age-matched non-infected mice with DNP-KLH subcutaneously ( s . c . ) into the hind footpad . This regimen is suited to induce DNP-KLH-specific T and B cell responses predominantly in the draining popliteal lymph node ( Figure 5AB ) . Fully differentiated plasma cells will predominantly migrate into the bone marrow and secrete Ig into the peripheral circulation . The induction of humoral responses in lymph nodes or in the spleen can be visualized by presence of B cells that secrete Ig spontaneously at low concentrations in cell cultures of these lymphatic organs . Cultured splenocytes derived from vaccinated and L . sigmodontis-infected mice secreted nematode-specific IgG1 , IgG2a and IgG2b but did not secrete any DNP-specific IgG ( Figure 5C ) . Cultured popliteal lymph node cells derived from the same mice produced DNP-specific but not nematode-specific IgG1 , thus visualizing the spatial separation of B cell responses to parasite and vaccine . DNP-specific IgG2a and IgG2b concentrations in the culture supernatant were below the detection limit . Systemic serum titres of DNP-specific IgG1 and IgG2b was still suppressed in day 60 L . sigmodontis-infected mice after s . c . vaccination ( Figure 5D ) , demonstrating that nematode-induced suppression acted on B cell responses that were primed in a local lymph node as well . As nematode-specific lymphocytes were not present in the local lymph nodes that drained the site of DNP-KLH vaccination ( Figure 5C ) , the observed suppression of systemic DNP-KLH-specific IgG response was not mediated by competition within the same site . Mice that were L . sigmodontis-infected but not vaccinated produced no DNP-specific IgG at all ( Figure 1B–E and Figure 5D: black circles ) and mice that were DNP-KLH vaccinated but not L . sigmodontis-infected did not produce any L . sigmodontis-specific IgG at all ( Figure S3 in Text S1: white squares ) . To quantify vaccine-induced B cells , we directly stained DNP-specific CD19+ B cells in the lymph nodes of vaccinated mice . DNP-KLH vaccination into the hind footpad induced DNP-binding CD19+ B cells in the draining lymph nodes ( Figure 6 ) while no DNP-binding B cells were detectable in the non-draining contralateral lymph nodes ( data not shown ) . DNP-specific B cells also bound peanut agglutinin ( PNA ) ( Figure 6A ) , indicating localization in the germinal centre [43] . Chronic nematode infection reduced the numbers of DNP-specific PNA-binding B cells ( Figure 6AB ) . Reduced DNP-specific B cell numbers were observed in the IgM/IgG double positive and in the terminally switched , IgG single positive B cell population . Thus , the reduced quantity of DNP-specific Ig detected in the serum of nematode-infected vaccinated mice was reflected by reduced numbers of DNP-specific B cells in the draining lymph nodes . Accumulating evidence suggests that a specialized T cell subset , the TFH , is responsible for provision of co-stimulation to B cells in the germinal centre [44] , [45] . TFH can be identified by expression of the chemokine receptor CXCR5 and the regulatory receptor programmed death 1 ( PD1 ) , in addition to activation markers such as CD44 [46] . DNP-KLH vaccination into the hind footpad did not change the frequency of total CD4+ T cells in draining compared to non-draining lymph nodes ( Figure 7B ) . However , the frequency of TFH , defined as PD1+CXCR5+ cells within the CD4+CD44+ population ( Figure 7A ) , increased selectively in the lymph node draining the site of vaccination ( Figure 7C ) . The strict correlation of TFH expansion within the stable CD4+ compartment to the site of vaccination in both , non-infected and nematode-infected mice , strongly suggests that these TFH were generated in response to vaccination . Draining lymph nodes of mice that were infected with L . sigmodontis for 60 days at the moment of vaccination displayed a significant reduction in both , TFH frequency within activated CD4+ T cells and absolute TFH numbers in comparison to non-infected vaccinated mice ( Figure 7C ) . Further characterization of TFH induced in non-infected and nematode-infected mice revealed no significant differences in the expression of inducible co-stimulator ( ICOS ) or Foxp3 ( Figure 7DE ) . Thus reduced numbers of DNP-specific B cells in DNP-KLH-vaccinated nematode-infected mice were correlated with reduced numbers of vaccination-induced TFH whereas the phenotype remained comparable . The collective data presented above suggest that vaccinations may fail in individuals carrying chronic nematode infections due to impaired induction of TFH . To evaluate the kinetics of this nematode-induced suppression , immune competent BALB/c mice were first allowed to terminate L . sigmodontis infection and then vaccinated with DNP-KLH ( Figure 8 ) . Immune-mediated control of infection was indicated by clearance of MF from the circulation that was observed between day 180 and day 280 p . i . ( Figure 8B and data not shown ) . Responses to vaccination were still reduced in mice that were previously nematode-infected when vaccination was carried out immediately after termination of microfilaraemia and four or eight weeks after clearance of MF from the circulation ( data not shown ) . Therefore we finally introduced an additional recovery period of 16 weeks after clearance of microfilaraemia before vaccination was performed ( Figure 8A ) . Strikingly both , IgG1 and IgG2b responses to DNP-KLH vaccination were still significantly suppressed in mice that had terminated L . sigmodontis infection at least 16 weeks before vaccination ( Figure 8C ) . Thereby DNP-specific IgG2b responses were almost absent in mice with a history of previous L . sigmodontis infection . It should be noted that although clearance of MF from the circulation at these late time points of infection strongly suggests impaired fitness of the adult parasites due to immune-mediated control and extermination , it does not confer direct information about the presence and the status of adult parasites in the thoracic cavity at the moment of vaccination . No living parasites were detectable in 100% of infected and vaccinated mice three weeks after vaccination ( data not shown ) . We observed limited amounts of remaining dead material in the thoracic cavity of some mice and we cannot formally exclude contribution of this helminth-derived material to suppression of vaccination .
In this study we demonstrate that chronic L . sigmodontis infection prevents humoral responses to bystander antigen vaccination . DNP-KLH-vaccinated and nematode-infected mice displayed reduced titres of DNP-specific IgG in the serum and reduced numbers of DNP-specific B cells in the lymph nodes compared to vaccinated , non-infected mice . L . sigmodontis-induced suppression was restricted to TD vaccination and not observed during T cell-independent B cell vaccination using the TI-2 antigen NIP-Ficoll . In congruence with our previous study performed in semi-permissive , day 14 L . sigmodontis-infected C57BL/6 mice [31] , this study strongly suggests that nematodes suppress antibody-producing B cells indirectly via suppression of accessory T cells . Within the CD4+ T cell population a specialized T cell subset , the TFH , is central for the initiation of classical B cell responses [44] , [45] . Next to PD1 expression , TFH are further characterized by the continued expression of the chemokine receptor CXCR5 that regulates their localisation within the B cell follicle [46] . Separating the sites of nematode- and vaccine-induced immune responses we distinguished between nematode- and vaccine-induced TFH . L . sigmodontis-infected mice displayed significantly reduced numbers and frequencies of vaccine-induced TFH in the lymph nodes draining the site of vaccination . Interestingly , we did not detect differences in the phenotype of TFH regarding the expression of ICOS , a central co-stimulatory receptor for B and T cell interaction that is essential for antibody responses to TD antigens [47] . Follicular regulatory T cells ( TFR ) that arise from thymus-derived Foxp3+ Treg and display a TFH-like , CXCR5+ PD1+ phenotype have been implied in regulation of TD B cells responses via limitation of TFH and B cell numbers in the germinal centre [48] , [49] . As Foxp3+ expression in TFH was unchanged and depletion of Foxp3+ T cells did not abrogate nematode-induced suppression , we ruled out a significant contribution of TFR to nematode-induced immune suppression . We have shown before that transient gastrointestinal nematode infection predominantly suppressed TH1-associated IgG2 responses to vaccination [33] , whereas L . sigmodontis infection of semi-permissive C57BL/6 mice [31] as well as chronic infection of fully susceptible BALB/c mice in the current study induced a generalized suppression of both TH1- and TH2-associated isotypes . We also did not observe differences in the affinity of vaccine-induced IgG in nematode-infected and non-infected mice . Taking these findings into account , we hypothesize that nematode infection interfered with the humoral response to vaccination already at the stage of TFH induction . Reduced numbers of vaccine-induced TFH will result in reduced provision of co-stimulation for vaccine-specific B cells . As a consequence reduced numbers of vaccine-specific B cells expand in the draining lymph node leading to reduced titres of vaccine-specific IgG in the peripheral circulation of nematode-infected mice . This hypothesis is supported by our previous study where we reported reduced proliferation of ovalbumin-specific TCR transgenic OT-II T cells as a simplified model for accessory T cells upon adoptive transfer into L . sigmodontis-infected mice [31] . Regarding the mechanism , we provide evidence that suppression of humoral response did not reflect direct competition between nematode- and vaccine-specific lymphocytes by separating the sites of nematode- and vaccine-specific immune responses . Suppression of OT-II T cell proliferation during acute L . sigmodontis-infection of C57BL/6 mice was shown to be Treg-independent and partially mediated by IL-10 [31] . Early depletion of CD4+CD25+ Treg improved host defence in L . sigmodontis-infected BALB/c mice , suggesting an implication of Treg in immune evasion for this genetic background [36] . Since we recently described a central role for Foxp3+ Treg in gastrointestinal nematode-induced immune evasion in BALB/c mice that was not functional in C57BL/6 mice [39] , it was conceivable that Foxp3+ Treg would contribute to the observed suppression of TD vaccination in L . sigmodontis-infected BALB/c mice . However , in the current study we show that suppression of IgG response during chronic infection of BALB/c mice was clearly established in the absence of Foxp3+ Treg . This suggests that immune modulation during acute and chronic L . sigmodontis infection in C57BL/6 and BALB/c mice is established by similar mechanisms such as IL-10 induction [31] . Potential mediators of suppression in addition to Foxp3+ Treg are IL-10 producing Foxp3− T cells [50] , IL-10 producing regulatory B cells [51] , [52] , alternatively activated macrophages [53] , and tolerogenic dendritic cells [54] that have been shown to mediate suppression during nematode infection in several murine systems . We are currently testing the function of these regulatory cell populations in suppression of OT-II T cell proliferation during acute L . sigmodontis-infection of C57BL/6 mice . Using fully susceptible BALB/c mice we dissected the impact of different L . sigmodontis life stages on vaccination efficacy . Suppression was induced by L4 , immature and mature adults , but not by recently transmitted L3 . Injection of isolated MF , in contrast , elevated IgG2a responses to vaccination , thus delivering a pro-inflammatory stimulus . A comparable pro-inflammatory stimulation was mediated by isolated MF in a model of LPS-induced sepsis [55] . In line with one previous study [56] , we demonstrate that this pro-inflammatory effect of MF was dominated by the anti-inflammatory effect exerted by L . sigmodontis adults . By contrast , the MF-mediated aggravation of LPS-induced sepsis was not rescued by presence of adults but also induced by implantation of mature adults releasing MF [55] . The different outcome may reflect the impact of long exposure ( i . e . 60 days ) of the host to L . sigmodontis before MF occurred in the circulation in our study in contrast to sudden implantation of MF releasing adults . Prolonged exposure may be needed for complete immune modulation to silence the strong pro-inflammatory effect of MF . In line with this reasoning we observed that suppression of DNP-specific IgG response to vaccination clearly increased with duration of infection . As our study focused on the first response to vaccination we used the early humoral and cellular response as indicator of efficacy . In order to model vaccination efficacy for the human situation more precisely , also the magnitude of memory responses several months after initial vaccination in absence and presence of helminth infection will be compared in future studies . Suppression of vaccination responses , once established , was observed several months after immune-mediated termination of infection . We cannot exclude that remaining helminth-derived material in the thoracic cavity contributed to suppression in the mice that had cleared microfilaria from the circulation 16 weeks before vaccination was performed . However , comparable remnants of large parasites are likely to be present in humans with a history of previous filarial infection as well . Thus , the prolonged suppression of vaccination-induced responses reported in this study may have implications for health policy . While some murine studies suggest that drug-induced termination of helminth infection may improve vaccination efficacy after 1–3 weeks of recovery [21] , [22] , our results show that the immune status does not immediately return to normal responsiveness in a setting modelling chronic infection . Prolonged suppression of vaccination responses after drug-induced termination of infection have been described in other mouse models . However , in these studies responsiveness was eventually achieved after recovery periods of 8 and 16 weeks , respectively [57] , [58] . Regarding the human population in the tropics where nematode infections are endemic a re-infection during such a recovery period is likely to occur . Despite the limits of murine models to reflect every aspect of the human situation , combined evidence gained in different mouse models for helminth infection can be informative . Since first murine studies demonstrated successful vaccination despite concurrent nematode infection by improved vaccination strategies [21] , [30] , [59] , [60] , we suggest that development of vaccination regimes that are functional despite pre-existing nematode infection would be more promising and should be considered in addition to deworming programs before vaccination .
|
Parasitic worms , called helminths , infect one-third of the world population . Despite exposure to their host's immune system many helminths establish chronic infections and survive several years within their host . They avoid elimination by dampening the immune response of their hosts . This immune suppression also affects immune responses to third-party antigens such as vaccines . Indeed , accumulating evidence suggests that helminth-infected humans display impaired responses to vaccination . Thus , anthelminthic treatment before vaccination is discussed . Here , we use helminth-infected mice to analyse kinetics and mechanism of helminth-induced interference with vaccination efficacy more precisely . We show that chronic helminth infection completely suppressed antibody responses to a model vaccine . Thereby helminths suppressed the antibody-producing B cells indirectly via suppression of accessory T helper cells . The suppression was more pronounced at later time points of infection and still observed in mice that had terminated the helminth infection for more than 16 weeks . In summary , our results suggest that vaccination may not only fail in helminth-infected individuals , but also in individuals with a history of previous helminth infections . Thus , our report highlights the importance to develop vaccination strategies that are functional despite concurrent helminth infection rather than deworming humans before vaccination .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"antibody-producing",
"cells",
"white",
"blood",
"cells",
"immune",
"evasion",
"immune",
"cells",
"cell",
"biology",
"animal",
"cells",
"b",
"cells",
"parasitology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"immunology",
"immune",
"suppression",
"immunomodulation",
"vaccination",
"and",
"immunization"
] |
2014
|
Nematode-Induced Interference with Vaccination Efficacy Targets Follicular T Helper Cell Induction and Is Preserved after Termination of Infection
|
A population of dynamic apical actin filaments is required for rapid polarized pollen tube growth . However , the cellular mechanisms driving their assembly remain incompletely understood . It was postulated that formin is a major player in nucleating apical actin assembly , but direct genetic and cytological evidence remains to be firmly established . Here we found that both Arabidopsis formin 3 ( AtFH3 ) and formin 5 ( AtFH5 ) are involved in the regulation of apical actin polymerization and actin array construction in pollen tubes , with AtFH3 playing a more dominant role . We found that both formins have plasma membrane ( PM ) localization signals but exhibit distinct PM localization patterns in the pollen tube , and loss of their function reduces the amount of apical actin filaments . Live-cell imaging revealed that the reduction in filamentous actin is very likely due to the decrease in filament elongation . Furthermore , we found that the rate of tip-directed vesicle transport is reduced and the pattern of apical vesicle accumulation is altered in formin loss-of-function mutant pollen tubes , which explains to some extent the reduction in pollen tube elongation . Thus , we provide direct genetic and cytological evidence showing that formin is an important player in nucleating actin assembly from the PM at pollen tube tips .
The pollen tube is the passage for two non-motile sperm cells and its proper growth is essential for successful reproduction in flowering plants [1–3] . Pollen tube growth is tightly regulated , and this raises many fascinating questions . Numerous studies suggest that the actin cytoskeleton is the core of the regulatory network of pollen tube growth , presumably by coordinating with various cellular events , such as the trafficking , docking and fusion of vesicles and the construction of the cell wall [4–7] . Actin filaments are arranged into distinct structures within different regions of growing pollen tubes , and these structures carry out distinct cellular functions [8–13] . Highly dynamic actin filaments within the apical region were demonstrated to be directly associated with the growth and turning of pollen tubes [14–16] . To date , however , we still have an incomplete understanding of how those apical actin filaments are constantly generated in pollen tubes . An essential step of actin polymerization is actin nucleation , which is controlled by various actin nucleation factors in cells . Among the actin nucleation factors identified in the literature [17] , the formins and the Arp2/3 complex are found in plants [18 , 19] . The formin proteins are characterized by the presence of two formin homology ( FH ) domains , FH1 and FH2 , which are capable of nucleating actin assembly from actin or actin bound to profilin [18 , 20–22] . Plant formins are categorized into two classes , designated as class I and class II . Class I formins have a transmembrane ( TM ) domain at their N-terminus followed by the C-terminal FH1 and FH2 domains , whereas class II formins do not have an N-terminal TM domain but carry an N-terminal phosphatase and tensin-related ( PTEN ) -like domain besides the conserved FH1 and FH2 domains [18 , 19 , 22–24] . In vitro biochemical analyses showed that most plant formins have the characteristic formin-mediated actin nucleating and barbed end capping and elongating activities [25–31] . Some plant formins were shown to have actin filament bundling [27–29 , 32 , 33] and microtubule interacting activities [29 , 32–34] , though the details of the mechanisms underlying these properties may vary between the different proteins . As important regulators of actin dynamics , the plant formins have been implicated in numerous physiological cellular processes , such as epidermal pavement cell morphogenesis [35] , cell division [32] , cytokinesis [25] , cell-to-cell trafficking [36] , and interaction with pathogens [37] , as well as the response to auxin signaling [38] . In particular , the formins have been implicated in polarized root hair growth [39–42] and pollen tube growth [4 , 31 , 43] . Specifically , after characterizing the cellular functions of the Arabidopsis formin gene AtFH5 , Cheung et al . [4] proposed that formin nucleates actin assembly from the membrane for the construction of the subapical actin structure . In line with this finding , a recent report showed that the pollen-specific Lilium longiflorum Formin 1 ( LiFH1 ) is involved in constructing the actin fringe structure [43] . However , considering that LIFH1 also has actin filament-bundling activity and was proposed to nucleate actin filaments from the surface of LiFH1-localized vesicles , more work is needed to understand how the formins fit into the apical actin polymerization pathway in general . Nonetheless , it was proposed that actin assembly mediated by formin-profilin modules may be a major pathway for actin polymerization from the apical membrane in the pollen tube [44] . The notion of formin acting as the major player in nucleating actin assembly is actually consistent with the scenario in which actin monomers are predicted to be buffered by an equal amount of profilin in pollen [45–47] . Considering that multiple formin isovariants exist in pollen , it is important to carefully document their precise intracellular localization and dynamics , their mechanism of action and their functional coordination . Here , we showed that two class I formins , AtFH3 and AtFH5 , localize to endomembrane systems and the plasma membrane ( PM ) . However , they have distinct PM distribution patterns: AtFH3 is localized evenly throughout the entire pollen tube while AtFH5 is concentrated at pollen tube tips . We demonstrated that both AtFH3 and AtFH5 are involved in the regulation of membrane-originated actin polymerization within the growth domain of the pollen tube and they have overlapping function in this aspect . Loss of function of AtFH3 and AtFH5 reduces the velocity of tip-directed vesicle transport and alters the apical vesicle accumulation pattern in the pollen tube , further supporting the active role of apical actin filaments in regulating vesicle traffic . Thus , we provide strong evidence that class I formins control membrane-originated actin polymerization to enable the construction of the apical actin structure in the pollen tube .
We previously showed that RNAi-mediated downregulation of AtFH3 impairs the formation of shank-oriented longitudinal actin cables in the pollen tube [31] . However , more work is required to determine whether and how AtFH3 may be involved in the regulation of actin polymerization within the apical region of the pollen tube . To better understand the mechanism of action of AtFH3 in regulating actin polymerization in pollen cells , we sought to analyze stable T-DNA insertion mutants of AtFH3 . In addition , considering that AtFH5 was previously shown to nucleate actin assembly from the plasma membrane for the construction of the subapical actin structure within the apical dome of the pollen tube [4] , we also sought to determine whether there is functional coordination of AtFH3 with AtFH5 in regulating apical actin polymerization in pollen tubes . To this end , we analyzed T-DNA insertion mutants for AtFH3 ( fh3-1 and fh3-2 ) and AtFH5 ( fh5-2 and fh5-3 ) as well as the double mutants ( fh3-1 fh5-2; fh3-2 fh5-3 ) ( Fig 1A ) . The results showed that fh3-1 , fh3-2 , fh5-2 and fh5-3 are knockout alleles ( Fig 1B–1E ) . We found that the pollen germination percentage was slightly but significantly reduced in pollen derived from fh3-1 , fh3-2 , fh5-3 , fh3-1 fh5-2 and fh3-2 fh5-3 mutant plants ( Fig 1F and 1G ) . In addition , we found that formin loss-of-function mutant pollen tubes grew significantly more slowly than WT pollen tubes ( Fig 1H ) . Furthermore , we found that the width of pollen tubes was increased slightly but significantly in fh3-1 fh5-2 and fh3-2 fh5-3 double mutants compared to WT ( Fig 1I ) . Interestingly , we found that growing fh3-1 fh5-2 pollen tubes were more curved than WT , fh3-1 or fh5-2 pollen tubes ( Fig 1J ) . This was supported by measurements showing that the ratio of the actual length of the pollen tube to the linear length was significantly increased in fh3-1 fh5-2 pollen tubes ( Fig 1K ) . Thus , these data suggest that AtFH3 and AtFH5 are required for normal polarized pollen tube growth , and AtFH3 has a more dominant role in this process . We next examined the organization of the actin cytoskeleton in WT and formin mutant pollen tubes using fluorescently labeled phalloidin . We found that the fluorescent signal from apical actin filaments was weaker in fh3 , fh5 and fh3 fh5 mutant pollen tubes than in WT ( Figs 2A and 2B and S1A ) . This is consistent with previously published work showing that both AtFH3 and AtFH5 are bona fide actin nucleation-promoting factors [25 , 31] . The reduction in the level of filamentous actin was also noted in formin loss-of-function mutant pollen grains when they were compared to WT pollen grains ( S2 Fig ) . It was previously proposed that within the pollen tube , actin filaments are arranged into two distinct arrays based on their origin [15] . Actin filaments originating from the apical membrane are organized into a unique apical actin structure which , in Arabidopsis pollen tubes , has its base 4–5 μm away from the tip [15] . Our results showed that the reduction in the level of actin filaments is very prominent within the region of the pollen tube that corresponds to the apical actin structure ( Figs 2A and 2B and S1A ) . We evaluated the quantitative contribution of AtFH3 and/or AtFH5 to apical actin polymerization by measuring the amount of actin filaments within the region that is occupied by the apical actin structure . The results showed that loss of function of AtFH3 causes a comparatively more severe reduction in the actin filament level than loss of function of AtFH5 ( Figs 2C and S1B ) . In addition , loss of function of both AtFH3 and AtFH5 caused an even more severe reduction in the actin filament level ( Figs 2C and S1B ) , suggesting that AtFH3 and AtFH5 have overlapping function in regulating apical actin polymerization . We found that , although loss of function of AtFH3 or AtFH5 also causes reduction in the amount of actin filaments in the shank region , loss of function of both AtFH3 and AtFH5 does not have overt effect on the amount of actin filaments in the shank region of pollen tubes ( Figs 2D and S1C ) . The reduction in the level of apical actin filaments was directly visualized by generating 3D plots of the 2D distribution of actin filament staining ( Fig 2E ) . We found that , although the level of apical actin filaments in pollen tubes is not significantly different between single AtFH3 loss-of-function mutants and AtFH3 and AtFH5 loss-of-function double mutants ( Figs 2C and S1B ) , actin filaments appear more fragmented and disorganized within the apical region of pollen tubes from double mutants than from fh3 single mutants ( Figs 2A and S1A ) . To quantify the degree of actin filament disorganization , we measured the angles formed between the apical actin filaments and the growth axis of the pollen tube . We noticed that the angles were significantly higher in fh3-1 and fh3-1 fh5-2 mutant pollen tubes ( Fig 2F ) , indicating that the apical actin filaments were relatively disorganized in these mutants . The increase in the angle is greater in fh3-1 fh5-2 pollen tubes than in fh3-1 pollen tubes ( Fig 2F ) . The increase in the angle was also noticed for actin filaments within the shank region of formin loss-of-function mutant pollen tubes ( Figs 2G and S1D ) . Thus , the data showed that both AtFH3 and AtFH5 are involved in the regulation of the polymerization and organization of apical actin filaments in the pollen tube , and AtFH3 has a more dominant role in this respect . To reveal how loss of function of AtFH3 and/or AtFH5 leads to the reduction in pollen tube growth , we examined the distribution of vesicles in pollen tubes . Transport vesicles were decorated with YFP-RabA4b as described previously [48] . We found that YFP-RabA4b-decorated transport vesicles accumulated in an inverted “V” cone shape in the WT pollen tube ( Fig 3A ) . By contrast , the inverted “V” cone shape created by accumulation of apical vesicles is not very obvious in formin loss-of-function mutant pollen tubes ( Fig 3A ) . Considering that the amount of apical actin filaments is reduced in formin loss-of-function mutant pollen tubes ( Figs 2A–2C and 2E and S1A and S1B ) , this result is consistent with the notion that apical actin filaments spatially restrict vesicles within the apical region of the pollen tube [6 , 15] . Surprisingly , we found that the region of vesicle accumulation was enlarged in fh3-1 fh5-2 pollen tubes compared to WT pollen tubes ( Fig 3A and 3B ) , which is presumably because the abnormally organized apical actin filaments cannot physically restrict the vesicles within the apical region . In addition , the fluorescence of vesicles is obviously brighter in fh3-1 fh5-2 pollen tubes than in WT pollen tubes ( Fig 3A and 3C ) , which is very likely because the backward movement of vesicles from the tip is severely reduced . To examine the dynamics of YFP-RabA4b-decorated vesicles , we used the technique of fluorescence recovery after photobleaching ( FRAP ) . After bleaching the apical and subapical regions , we found that the recovery rate of vesicles is reduced in fh3-1 and fh3-1 fh5-2 mutant pollen tubes ( Fig 3D and 3E and S1 , S2 and S4 Movies ) . By comparison , the recovery rate of vesicles in fh5-2 pollen tubes is only slightly slower than WT pollen tubes ( Fig 3D and 3E and S1 and S3 Movies ) . The extent of the alteration in vesicle recovery rate within the apical and subapical regions of the pollen tube correlates well with the extent of the reduction in apical actin filaments ( Figs 2A–2C and 2E and S1A and S1B ) . Thus , the results showed that loss of function of AtFH3 and/or AtFH5 alters the apical vesicle accumulation pattern and reduces the rate of vesicle turnover in the pollen tube . To determine the precise intracellular localization of AtFH3 and AtFH5 , we generated green fluorescent protein ( GFP ) fusion constructs of AtFH3 and AtFH5 driven by their own promoters , AtFH3pro:AtFH3-eGFP and AtFH5pro:AtFH5-eGFP , and transformed them into fh3-1 and fh5-2 to generate the transgenic plants AtFH3pro:AtFH3-eGFP;fh3-1 and AtFH5pro:AtFH5-eGFP;fh5-2 , respectively . We found that transformation of those constructs rescued the defects in the amount and organization of apical and subapical actin filaments ( S3 Fig ) , suggesting that the GFP fusion constructs are functional . Confocal microscopy revealed that both AtFH3-eGFP and AtFH5-eGFP form punctate structures in the cytoplasm of pollen grains and pollen tubes ( Fig 4A , 4B , 4D and 4E ) . This suggests that the fusion proteins are associated with the endomembrane system , which is consistent with previous characterization of AtFH5 [4] . In addition , both AtFH3 and AtFH5 are able to localize to the PM but exhibit distinct patterns: AtFH3 is localized quite evenly on the PM along the entire pollen tube ( Fig 4A and 4B and S5 Movie ) while AtFH5 is concentrated on the PM within the apical dome of the tube ( Fig 4D and 4E and S6 Movie ) . After plasmolysis , AtFH5-eGFP retained its association with apical membranes while AtFH3-eGFP was retained on the PM along the pollen tube ( S4 Fig ) . AtFH3 and AtFH5 also had distinct localization patterns in ungerminated pollen grains , with AtFH3 exhibiting obvious PM localization ( Fig 4A ) whereas AtFH5 does not exhibit obvious PM localization ( Fig 4D ) . The endomembrane and PM localization of AtFH3 and AtFH5 in pollen tubes were further confirmed by staining with FM4-64 dyes , which showed that AtFH3-eGFP and AtFH5-eGFP overlapped with FM4-64 dyes on the cell membrane and punctate structures within the cytoplasm ( Fig 4C and 4F ) . The subcellular localization data are consistent with the presence of a transmembrane ( TM ) domain in AtFH3 and AtFH5 [4 , 25 , 31] . We found that depolymerization of the actin cytoskeleton ( S5A Fig ) does not prevent the PM and endomembrane targeting of AtFH3 and AtFH5 ( S5B and S5C Fig ) . This suggests that their targeting to the PM and endomembranes in pollen tubes does not require their interaction with the actin cytoskeleton . Furthermore , we found that the N-terminus of AtFH3 , which contains the signal peptide ( SP ) and TM , is sufficient for targeting of AtFH3 to the PM and endomembranes ( S6A Fig ) . This is consistent with previous observations that the membrane localization of LiFH1 is determined by its N-terminus , which also contains the SP-TM domain [43] . Strikingly , we found that replacement of the TM domain of AtFH3 with that of AtFH5 endows AtFH3 with a PM localization pattern similar to that of AtFH5 ( S6B Fig ) . Thus , our study suggests that both AtFH3 and AtFH5 are able to localize to the PM and endomembrane system , and the membrane localization pattern is dictated by their TM domains . To understand the defective actin filament organization in formin loss-of-function mutant pollen tubes , we traced the dynamics of individual actin filaments decorated with Lifeact-eGFP as described previously [16 , 49] . We found that actin filaments are continuously polymerized from the apical membrane during WT pollen tube growth , and consequently form a bright apical actin structure ( Fig 5A ) [15] . However , we found that apical actin polymerization was impaired in formin loss-of-function mutant pollen tubes , and this affected the formation of the apical actin structure ( Figs 5A and S7 and S7–S10 Movies ) . We next traced the dynamics of individual actin filaments and determined the parameters associated with them . Given that apical actin polymerization is severely impaired in fh3-1 fh5-2 mutant pollen tubes and it is hard to select individual apical membrane-originated actin filaments for measurement , we only traced the dynamics of individual actin filaments in WT , fh3-1 and fh5-2 pollen tubes and carefully compared their dynamic parameters . We found that the elongation rate of actin filaments originating from the apical membrane is reduced significantly in fh3-1 and fh5-2 pollen tubes when compared to WT pollen tubes ( Fig 5B ) . This explains to some extent the impairment in the formation of the apical actin structure . In addition , we found that although there is no overt difference in the severing frequency of actin filaments between fh3-1 and fh5-2 pollen tubes and WT pollen tubes ( Fig 5D ) , the depolymerization rate of actin filaments is reduced significantly in fh3-1 and fh5-2 pollen tubes when compared to WT pollen tubes ( Fig 5C ) . Furthermore , no overt difference was detected in the maximal filament lifetime of apical actin filaments in fh3-1 and fh5-2 pollen tubes ( Fig 5F ) , but the maximal filament length of apical actin filaments is reduced significantly in fh3-1 pollen tubes ( Fig 5E ) , which is very likely due to the reduction in the filament elongation rate of apical actin filaments in fh3-1 pollen tubes . Given that AtFH3 and AtFH5 are bona fide actin nucleation-promoting factors [25 , 31] , the reduction in the filament depolymerization rate in formin mutant pollen tubes is very likely indirect , and is presumably related to the reduction in the actin filament level in formin mutant pollen tubes . Together , these data suggest that the impairment in the formation of the apical actin structure is mainly caused by the defects in formin-mediated actin filament elongation .
We found that AtFH3 and AtFH5 are involved in the regulation of actin polymerization in pollen cells ( Figs 2 and 5 and S1 , S2 and S7 ) . Within pollen tubes , the reduction in the level of filamentous actin resulting from loss of function of AtFH3 and AtFH5 is comparatively more severe at the tip ( Figs 2 and 5 and S1 and S7 ) , and this is very likely because the activity of formin is strictly required at the pollen tube tip where active actin polymerization occurs [15] . Correspondingly , we found that formin is relatively concentrated on the PM at the tip of the pollen tube ( Fig 4 ) . In support of the role of formin in regulating apical actin polymerization , a recent study showed that a class I formin , LiFH1 , is involved in the construction of the actin fringe in the pollen tube [43] . However , LiFH1 is biochemically distinct from AtFH3 and AtFH5 since it has actin filament-bundling activity [43] . If there is a formin that behaves like LIFH1 in Arabidopsis pollen tubes , it will be interesting to explore if and how it coordinates with AtFH3 and AtFH5 to regulate apical actin polymerization . In terms of the effect of loss of function of AtFH3 on the organization of actin filaments in the shank of pollen tubes , our results differ slightly from a previous report that RNAi-mediated downregulation of AtFH3 causes severe defects including disorganized shank-localized actin cables and depolarized pollen tube growth [31] , as we noticed the disorganization of actin filaments in the shank of fh3 pollen tubes and inhibition of pollen tube growth but not that severe ( Figs 1 , 2 and 5 ) . We do not currently know the reason for this , but it could be due to off-target effects derived from the RNAi-mediated downregulation approach . For instance , this RNAi construct might target to other pollen-expressed formins . If this is indeed the case , it explains why loss of function of AtFH3 and/or AtFH5 causes weak phenotype in term of pollen tube growth . Certainly , we also cannot rule out the possibility that Arp2/3 complex might take a partial role in nucleating actin assembly to compensate for the loss of AtFH3 and/or AtFH5 in pollen tubes , although Arp2/3 complex and formins nucleate actin assembly using different biochemical mechanisms . Nonetheless , we convincingly demonstrate that AtFH3 and AtFH5 contribute to actin polymerization at the pollen tube tip and AtFH3 plays a more dominant role than AtFH5 in this process ( Figs 2 and 5 and S1 and S7 ) . This is actually consistent with the fact that the expression of AtFH3 is more abundant than AtFH5 in pollen ( https://www . genevestigator . com/gv/index . jsp ) . Considering these results along with observations showing that actin filaments are mainly generated from the apical membrane ( Figs 5A and S7 ) [15 , 16 , 44 , 50 , 51] , it is fair for us to propose that the membrane-anchored class I formins , AtFH3 and AtFH5 , drive actin polymerization by utilizing profilin-actin complexes in the cytoplasm within the apical region of the pollen tube ( Fig 5G ) . We showed that AtFH3 and AtFH5 are abundant within the cytoplasm of the growth domain and are presumably localized on vesicles ( Fig 4B and 4E ) . These observations suggest that the activity of formins on the surface of vesicles is maintained at a very low level since no obvious actin polymerization was detected from the surface of vesicles in pollen tubes . Furthermore , it was reported that apical actin polymerization occurs concurrently with and is required for pollen tube growth [15] . The mechanism that activates formins on the PM during pollen tube growth is of great interest . Compared to non-plant formins , plant formins lack the GTPase-binding domain ( GBD ) and the diaphanous autoregulatory domain ( DAD ) that are crucial for the regulation of their actin nucleation activity [18 , 19 , 24] . The molecular mechanisms that tightly regulate the activity of plant formins remain a mystery . We found that AtFH3 and AtFH5 both localize to the PM , and exhibit distinct PM localization patterns ( Fig 4 ) . Considering that the intracellular localization of AtFH3 and AtFH5 is determined by their TM domains ( S6 Fig ) , the distinct PM localization pattern of AtFH3 and AtFH5 suggests that their TM domains have distinct functions . In support of this notion , we found that substitution of the TM domain of AtFH3 with that of AtFH5 enables AtFH3 to exhibit a PM localization pattern similar to that of AtFH5 ( S6B Fig ) . Our data suggest that , although AtFH3 and AtFH5 belong to the same subclass , functional divergence of their TM domains has endowed them with distinct PM localization patterns . The two proteins might consequently perform distinct roles in regulating membrane-originated actin polymerization within the pollen tube . The function of formins , as regulators of actin dynamics , is achieved through their FH1FH2 domain [18] . It remains to be determined whether the C-terminal FH1FH2 domain of AtFH3 and AtFH5 might have distinct actin regulatory functions . Previous studies revealed that both AtFH3 and AtFH5 are bona fide actin nucleation factors [25 , 31] , but no side-by-side comparison has been performed . Given that actin monomers were predicted to be buffered by an equal amount of profilin in pollen [9] , AtFH3 and AtFH5 might differ in their ability to utilize profilin-actin complexes in the pollen tube . There are at least five actin isovariants and two profilin isovariants in Arabidopsis pollen , and they were reported to be distributed uniformly in the cytoplasm of the pollen tube overall [44 , 52 , 53] . It is possible that they may form different profilin-actin complexes within the cytoplasm . In this regard , AtFH3 and AtFH5 might have preference for certain profilin-actin complexes in the pollen tube . Consequently , the combination of different actin , profilin and formin isovariants may fine-tune the actin polymerization machinery to meet the demands of rapid pollen tube growth . In support of this speculation , a previous report showed that AtFH4 interacts specifically with profilin 2 ( PFN2 ) rather than PFN3 [41] .
The Arabidopsis plants were cultured at 22°C under a 16-h light/8-h dark cycle . The T-DNA insertion mutants , fh3-1 ( Salk_150350 ) , fh5-2 ( Salk_044464 ) , and fh5-3 ( Salk_152090 ) were obtained from Nottingham Arabidopsis Stock Center on the Columbia-0 ecotype ( Col-0 ) background . They were backcrossed with Col-0 three times before the subsequent phenotypic analyses . fh3-2 ( CSHL_GT24923 ) was obtained from Cold Spring Harbor Laboratory and backcrossed with Col-0 three times before the phenotypic characterization . The genotyping of fh3-1 , fh5-2 and fh5-3 was performed using primer pairs fh3-1 salk_150350-LP/fh3-1 salk_150350-RP and fh5-2 salk_044464-LP/fh5-2 salk_044464-RP , and fh5-3 salk_152090-LP/fh5-3 salk_152090-RP ( S1 Table ) , respectively , in combination with Salk_LB 1 . 3 ( S1 Table ) . The genotyping of fh3-2 was performed using primer pair fh3-2 CSHL_GT24923-LP/fh3-2 CSHL_GT24923-RP along with Ds3-1 ( see S1 Table ) . The T-DNA insertion mutant fh5-2 has been characterized previously [25] . To determine the functional coordination between AtFH3 and AtFH5 , fh3-1 fh5-2 and fh3-2 fh5-3 double mutants were generated by crossing fh3-1 with fh5-2 or fh3-2 with fh5-3 . To complement fh3-1 and fh5-2 and indicate the intracellular localization of AtFH3 and AtFH5 , GFP fusion constructs of AtFH3 and AtFH5 driven by their own promoters were generated . To generate the AtFH3-eGFP fusion construct , the nucleotide sequence containing the promoter and genomic region of AtFH3 were amplified from Arabidopsis genomic DNA with the primer pair AtFH3pg-PstI-F/AtFH3pg-KpnI-R ( see S1 Table ) and eGFP was amplified from pCAMBIA1301 carrying eGFP with eGFP-SacI-F/eGFP-EcoRI-R ( see S1 Table ) . The PCR products were subsequently moved into pCAMBIA1301 to generate pCAMBIA1301-gFormin3-eGFP . To generate the AtFH5-eGFP fusion construct , the promoter sequence and the genomic sequence of AtFH5 were amplified with the primer pairs AtFH5pro-F/AtFH5pro-R and AtFH5genomic-F/AtFH5genomic-R ( see S1 Table ) , respectively . Given that no suitable restriction enzyme sites were available , the AtFH5 genomic sequence was mutated to disrupt an internal SacI restriction site so that SacI could then be used for the subsequent cloning . The AtFH5 genomic sequence was amplified with primers gAtFH5-Mut-F/gAtFH5-Mut-R using the AtFH5 genomic sequence as the template . The product was subsequently moved into pCAMBIA1301 to generate the final pCAMBIA1301-gFormin5-eGFP construct . The constructs gFormin3-eGFP-pCAMBIA1301 and gFormin5-eGFP-pCAMBIA1301 were transformed into fh3-1 and fh5-2 to generate the transgenic plants , gFormin3-eGFP-pCAMBIA1301;fh3-1 and gFormin5-eGFP-pCAMBIA1301;fh5-2 , respectively , using the agro bacteria-mediated flower-dipping method [54] . The transgenic plants at T3 were used for the subsequent analysis . To determine whether the intracellular localization pattern of AtFH3 is determined by its N-terminus , which contains signal peptide ( SP ) and transmembrane ( TM ) domain , we amplified the sequence containing both SP and TM of AtFH3 ( AtFH3-SP-TM ) using primer pair AtFH3-SPTM-F/AtFH3-SPTM-R ( S1 Table ) . The PCR product of AtFH3-SP-TM , along with the Lat52 promoter amplified with pair Lat52-F/Lat52-R ( S1 Table ) , was moved into pCAMBIA1301 to generate pCAMBIA1301-Lat52pro-AtFH3-SP-TM-eGFP . The construct was subsequently transformed into WT Arabidopsis plants . Pollen derived from the transgenic plants was germinated on solid GM for 2 h , then observed under an Olympus FV1000MPE multiphoton laser scanning confocal microscope equipped with a 100× objective ( numerical aperture of 1 . 4 ) . Samples were excited under a 488-nm argon laser with the emission wavelength set at 505–605 nm . To replace the TM domain of AtFH3 with that of AtFH5 , overlap PCR was performed to amplify the sequence of the promoter of AtFH3 ( AtFH3pro ) and AtFH5-SP-TM with primer pairs 3+5TM F2/3+5TM R2 and 3+5TM F3/3+5TM R3 ( S1 Table ) using AtFH3pro and AtFH5-SP-TM as the template , respectively . Subsequently , the overlap products were amplified specifically with primer pair 3+5TM F1-XbaI/3+5TM R1-KpnI ( S1 Table ) . Given that no suitable restriction sites were available , the AtFH3pro-AtFH5-SP-TM genomic sequence was subsequently mutated using PCR with primer pair 3+5TM-Mut-F/3+5TM-Mut-R ( S1 Table ) to disrupt an internal PstI restriction site in order to facilitate subsequent cloning . The sequences of AtFH3pro-AtFH5-SP-TM and AtFH3 FH1FH2 were then amplified with primer pairs 3+5TM F2-PstI/3+5TM R2-XbaI and AtFH3 FH1FH2-F/AtFH3 FH1FH2-R ( S1 Table ) , respectively . The error-free PCR products were subsequently moved into pCAMBIA1301 to generate pCAMBIA1301-AtFH3pro-AtFH5-SP-TM-AtFH3-FH1FH2-eGFP . The construct pCAMBIA1301-AtFH3pro-AtFH5-SP-TM-AtFH3-FH1FH2-eGFP was finally transformed into fh3-1 to generate the transgenic plants , pCAMBIA1301-AtFH3pro-AtFH5-SP-TM-AtFH3-FH1FH2-eGFP;fh3-1 . The transgenic Arabidopsis plants at T3 were used for the subsequent analysis . qRT-PCR was performed to determine the transcript levels of AtFH3 and/or AtFH5 in the formin T-DNA insertion mutants . Total RNA was isolated from pollen derived from WT ( wild-type ) , fh3-1 , fh3-2 , fh5-2 , fh5-3 , fh3-1 fh5-2 and fh3-2 fh5-3 plants using TRIzol reagent ( Invitrogen ) according to the manufacturer’s instructions , and cDNA was subsequently synthesized using MMLV reverse transcriptase ( Promega ) with oligo-d ( T ) 18 . To determine the AtFH3 transcript levels , partial coding region sequences of AtFH3 were amplified with primer pairs AtFH3-F1/AtFH3-R1 and AtFH3-F2/AtFH3-R2 ( see S1 Table ) . To determine the AtFH5 transcript levels , the partial coding region of AFH5 was amplified with the primer pair AtFH5-F/AtFH5-R . To determine the AtFH3 and AtFH5 transcript levels in the complementation plants , the primer pairs AtFH3-F2/AtFH3-R2 and AtFH5-F/AtFH5-R ( see S1 Table ) were used , respectively . The internal control was eIF4A , which was amplified with the primer pair q-eIF4A-F/q-eIF4A-R ( see S1 Table ) . The real-time PCR data were analyzed with the method of Livak ( 2-ΔΔCt ) [55] . In vitro Arabidopsis pollen germination was performed according to previously described methods [56] . Briefly , pollen was isolated from newly opened flowers and placed on pollen germination medium [GM: 1 mM CaCl2 , 1 mM Ca ( NO3 ) 2 , 1 mM MgSO4 , 0 . 01% ( w/v ) H3BO3 , and 18% ( w/v ) sucrose solidified with 0 . 8% ( w/v ) agar , pH 6 . 9~7 . 0] . The plates were cultured at 28°C under moist conditions . After 2 h of culture , the pollen germination rate was quantified by observing pollen grains and pollen tubes under an IX71 microscope ( Olympus ) equipped with a 10× objective . Images were collected with a Retiga EXi Fast 1394 CCD ( charge-coupled device ) camera using Image-Pro Express 6 . 3 software . To calculate the pollen germination percentage , a minimum of 300 pollen grains was counted in each experiment . At least three experiments were performed . To accurately calculate the pollen tube growth rate , we developed a new method based on calculating the slope of a kymograph of a single growing pollen tube . Briefly , after the pollen tube grew to an average length of approximately 200–300 μm , the solid pollen germination medium containing the germinating pollen was moved to a circular plate ( Cat# D35-20-1-N , In Vitro Scientific ) under an IX71 microscope ( Olympus ) equipped with a 4× objective . A microscope field containing at least 15~20 pollen tubes was identified , and the growth of individual pollen tubes was monitored by collecting time-lapse images ( about 15–20 images in total ) at time intervals of 1 min . A kymograph was created along the growth direction at the center of the growing pollen tube and the growth rate of the pollen tube was calculated from the slope of the kymograph . The experiments were repeated at least three times . Pollen tubes were stained with the lipophilic dye FM4-64 ( Invitrogen ) . The loading of pollen tubes with FM4-64 was achieved by direct addition of FM4-64 dye ( 5 μM in liquid pollen germination medium ) on the surface of solid pollen germination medium . After incubation with FM4-64 solution for 15 min , images were captured with an Olympus FV1000MPE multiphoton laser scanning confocal microscope as described above . FM4-64 dye was excited with an argon laser at 546 nm , and the emission wavelength was set in a range of 600–650 nm . To reveal the organization of the actin cytoskeleton in pollen grains and pollen tubes , pollen grains were germinated for 2 h on solid GM , then subjected to fixation and staining with Alexa-488/568 phalloidin as described previously [52 , 57] . Actin filaments were observed with an Olympus FV1000MPE multiphoton laser scanning confocal microscope equipped with a 100× objective ( numerical aperture of 1 . 4 ) . The fluorescent phalloidin was excited with an argon laser at 488 nm and 560 nm , and the emission wavelength was set in the range of 505–605 nm and 650–700 nm , respectively . The relative amount of actin filaments in pollen grains and pollen tubes was quantified by measuring the fluorescence pixel intensity using ImageJ software ( http://rsbweb . nih . gov/ij/; version 1 . 46 ) . At least three experiments were performed . The organization of apical actin filaments or bundles was quantified by determining the angles formed between each apical actin filament or bundle and the growth axis of pollen tubes , which was performed with ImageJ roughly as described previously for the quantification of the angles formed between longitudinal actin cables and the growth axis of the pollen tube in the shank region [56] . To ensure that each apical actin filament or bundle was analyzed only once , three to four optical sections were excluded for analysis in each pollen tube . Since we do not know the polarity of apical actin filaments or bundles , we only selected the small angles . More than 200 apical actin filaments or bundles from 10 pollen tubes were measured for each genotype . In order to visualize the dynamics of actin filaments in pollen tubes , the actin marker Lifeact-eGFP was introduced into formin loss-of-function mutants ( fh3-1 , fh3-2 , fh5-2 , fh5-3 , fh3-1 fh5-2 and fh3-2 fh5-3 ) by crossing the mutants with transgenic WT plants harboring Lat52:Lifeact-eGFP [16] . Time-lapse Z-series images were collected every 2 s using MetaMorph software with the step size set at 0 . 7 μm . The dynamics of individual actin filaments were quantified by measuring dynamic parameters , including the elongation rate , depolymerization rate , severing frequency , maximum filament length and maximum filament lifetime as described previously [16 , 57] . At least ten pollen tubes for each genotype were analyzed . A kymograph taken along the growth direction at the center of the pollen tube was created to analyze the F-actin intensity along the growing pollen tube as described previously [16] . To observe the tip-directed vesicle transport in pollen tubes , YFP-RabA4b was introduced into the formin loss-of-function mutants by crossing the mutants with transgenic WT plants expressing Lat52:YFP-RabA4b [48 , 58] . Pollen from the resulting plants was germinated on solid GM at 28°C , and when the pollen tubes reached about 150 μm , they were imaged under an Olympus FV1000MPE multiphoton laser scanning confocal microscope equipped with a 100× objective ( numerical aperture of 1 . 4 ) . Samples were excited under a 488-nm laser with the emission wavelength set at 505–605 nm . Optical sections were scanned with the step size set at 0 . 7 μm . For the fluorescence recovery after photobleaching ( FRAP ) experiments , apical regions were bleached for 3 s using a 488-nm laser at 100% power and a 405-nm laser at 45% power . Fluorescence recovery was recorded at 2 s intervals for 200s with a 488-nm laser at 10% power . To determine the recovery rate , the mean gray value of the apical region ( 0–5 μm away from the tip ) was measured using ImageJ software and plotted against the elapsed time as described previously [52 , 58] . Experiments were repeated at last 20 times and the values of YFP-RabA4b fluorescence were averaged and used for subsequent exponential curve fitting as described previously [52] .
|
Actin polymerization has been implicated in the regulation of rapid polarized pollen tube growth . The important role of actin polymerization is well appreciated , but the mechanisms that regulate rapid actin polymerization in pollen tubes remain incompletely understood . It was postulated that one of the major actin polymerization pathways in pollen tubes involves formin/profilin modules . However , direct genetic and cytological evidence is still required to support the role of formin in this framework . Using state-of-the-art live-cell imaging in combination with reverse genetic approaches , we demonstrate here that two class I formins , Arabidopsis formin 3 ( AtFH3 ) and formin 5 ( AtFH5 ) , are involved in the regulation of apical actin polymerization and actin array construction in pollen tubes . In support of the role of AtFH3 and AtFH5 in regulating membrane-originated apical actin polymerization , we found that both of them are localized to the plasma membrane ( PM ) at pollen tube tips . Live-cell imaging revealed that the reduction in filamentous actin is very likely due to the decrease in elongation of actin filaments originating from the apical membrane . We also found that AtFH3 and AtFH5 exhibit distinct PM localization patterns in the pollen tube , suggesting that they might have distinct roles in regulating actin polymerization in pollen tubes . Our study provides direct genetic and cytological evidence that formins act as important players in regulating apical actin assembly in pollen tubes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"plant",
"anatomy",
"cell",
"motility",
"actin",
"filaments",
"vesicles",
"cell",
"processes",
"pollen",
"plant",
"science",
"cellular",
"structures",
"and",
"organelles",
"dynamic",
"actin",
"filaments",
"contractile",
"proteins",
"actins",
"actin",
"polymerization",
"proteins",
"pollen",
"tube",
"cell",
"membranes",
"biochemistry",
"cytoskeletal",
"proteins",
"cell",
"biology",
"biology",
"and",
"life",
"sciences"
] |
2018
|
Arabidopsis class I formins control membrane-originated actin polymerization at pollen tube tips
|
Dietary restriction extends longevity in organisms ranging from bacteria to mice and protects primates from a variety of diseases , but the contribution of each dietary component to aging is poorly understood . Here we demonstrate that glucose and specific amino acids promote stress sensitization and aging through the differential activation of the Ras/cAMP/PKA , PKH1/2 and Tor/S6K pathways . Whereas glucose sensitized cells through a Ras-dependent mechanism , threonine and valine promoted cellular sensitization and aging primarily by activating the Tor/S6K pathway and serine promoted sensitization via PDK1 orthologs Pkh1/2 . Serine , threonine and valine activated a signaling network in which Sch9 integrates TORC1 and Pkh signaling via phosphorylation of threonines 570 and 737 and promoted intracellular relocalization and transcriptional inhibition of the stress resistance protein kinase Rim15 . Because of the conserved pro-aging role of nutrient and growth signaling pathways in higher eukaryotes , these results raise the possibility that similar mechanisms contribute to aging in mammals .
Calorie restriction ( CR ) , which usually refers to a 20–40% reduction in calorie intake , can effectively prolong life span in taxonomically diverse organisms ranging from yeasts to mammals [1] . It is also known that selective restriction of carbohydrates or proteins , as well as alternate day fasting without an overall restriction of calories , can also extend longevity [2]–[3] , suggesting that reduced levels of specific macronutrients in the diet ( Dietary Restriction , which also includes CR ) , can achieve at least some of the effects of CR [3] . On the other hand , studies on model organisms focusing on genes and pathways involved in aging have identified Ras , the yeast ortholog of S6 kinase , Sch9 and the target of Rapamycin ( TOR ) as key regulators of longevity and stress resistance [4]–[5] . However , although the stress resistance genes as well as the transcription factors Msn2 , Msn4 , and Gis1 play key roles in the effects of CR on life span extension [6] , the connection between the availability of each component of the diet , stress resistance and aging genes has remained only partially understood [7] . Depletion of glucose , the best characterized nutrient , causes calorie restriction-associated phenotypes including life span extension [8]–[9] while its addition to starved yeast cells alters the expression of almost one third of the yeast transcriptome [10]–[11] . These effects are largely due to altered activity of the protein kinase A ( PKA ) through Ras-cAMP and Sch9-dependent pathways [12]–[13] . Regarding the other macronutrients , many genes have been identified for their role as sensors and transporters for nutrients different from glucose [14]–[19] , but little is known about the molecular cascades activated by specific nutrients . In yeast , amino acid scarcity increases replicative life span [20] possibly by affecting protein synthesis [21] . In flies and rodents , changes in amino acid or dietary composition can have profound effects on life span [22]–[26] and in human cell cultures the availability of amino acids affects gene expression profiles [27]–[29] , but the effect of each amino acid on aging is largely unknown . Evidence based on the effect of amino acid withdrawal and repletion points to the TORC1 complex as a major amino acids transducer in mammalian cells [30]–[31] . In CHO-IR mammalian cells , amino acid withdrawal results in the selective inhibition of S6K1 and dephosphorylation of 4E-BP , rendering these targets unresponsive to insulin [32] . On the contrary , amino acids replenishment or just the addition of leucine and arginine , in the absence of serum or insulin , restores 4E-BP phosphorylation , S6K1 activity and insulin sensitivity [32] . The AGC kinase Sch9 ( ortholog of the mammalian S6 and possibly AKT kinases ) has been reported to be a major target of the activated TORC1 complex [33] . The latter phosphorylates Sch9 at multiple residues within the hydrophobic motif ( HM ) whereas the Sch9 T loop is the target of the Phosphoinositide Dependent protein Kinase 1 ( PDK1 ) orthologs Pkh1–3 [34]–[35]: the main downstream effectors of the PI3kinase that in yeast are activated by sphingolipids [36]–[37] . It has therefore been suggested that Sch9 integrates nutrient signals coming from Tor1 with stress signals coming from sphingolipids [35] , [38] . Interestingly , a physical interaction between the general amino acid permease ( Gap1 ) and three components ( Pis1 , Lip1 , Tsc13 ) of the sphingolipids biosynthetic pathway has recently been demonstrated [39] raising the possibility that sphingolipids metabolism may be affected by amino acids availability . However , increasing evidences suggest that few molecular switches may integrate all nutrient signals . Flo11 , one of the genes activated by the lack of nitrogen [40] , is also affected by PKA activation state [41] integrating nitrogen signaling and glucose signaling . Other reports have related the activity of the General Amino-acids Permease ( GAP1 ) to the PKA activation state [42]–[43] and even though they reached conflicting conclusions , both suggest a dependency of the amino acid transport system on the PKA activation state , thus connecting amino acid to glucose response . Pkh1–3 phosphorylate both PKA catalytic subunit and Sch9 T loop , linking sphingolipids to glucose signaling [34] . Moreover , Bcy1 , the regulatory subunit of the PKA , is modulated by cAMP concentration , which transduces glucose availability , but is also phosphorylated by TORC1 [44] . Here , we investigated the effect of media composition on aging and stress resistance and identified the connection between specific nutrient and the well-known and conserved pro- and anti-aging genes .
Although nutrient-dependent pathways in Saccharomyces cerevisiae have been studied extensively [21] , [45]–[48] , the role of each nutrient on cellular protection and aging is poorly understood . Because our previous studies and those of many other laboratories indicated that resistance to oxidative stress is tightly linked to longevity , we developed a protocol to measure oxidative stress resistance variations in post-diauxic yeast cells exposed to various nutrient mixtures ( nutrient response assay , figure S1 , see materials and methods section for details ) . Experiments were performed using the DBY746 laboratory wild type yeast strain ( carrying the Leucine , Histidine , Tryptophan and Uracil auxotrophies; see table S1 for the list of strains used and for the complete relevant genotype ) and repeated with the corresponding isogenic prototrophic yeast strain , as well as with a natural yeast strain commonly used by winemaking industries . In agreement with previous observations [49]–[50] , we found increased stress resistance by reduction of dextrose concentration from 2% to 0 . 5% both in auxotrophic and in prototrophic DBY746 and in winemaking yeast ( Figure 1a and S2b , S3a ) and synergistic effects of carbohydrates and all amino acids addition on cellular sensitization to either peroxide treatment or heat shock ( Figure 1a S2a , b and S3a ) . To rule out the possibility that the effect of amino acids on stress resistance was merely due to pH changes , experiments were repeated at pH 3 . 7 and 6 . Essentially the same results were obtained ( Figure S2b , c ) . In addition , all the experiments were performed in non-limiting conditions for nitrogen source to separate the effects of limited nitrogen source from that of amino acid restriction . To avoid the possibility that liquid cultures , such as those used in standard chronological life span assay , may lead to toxic levels of certain metabolites ( e . g . acetic acid and ethanol [51]–[52] ) or that regrowth phenotype may affect the interpretation of the results , we repeated the experiments using in situ chronological life span , in which about 200 cells are maintained isolated from others on a plate containing 2% glucose [53] . This assay confirmed the strong association between stress resistance and life span measurements ( Figure 1b ) . In fact , consistently with the stress resistance assay , the presence of the complete amino acids mixtures significantly shortened the life span of wild type yeasts with respect to a medium containing only the essential amino acids . To identify the downstream effectors of glucose- and amino acids-dependent sensitization , we monitored stress resistance in isogenic yeast strains lacking key mediators of nutrients signaling: ( the RAS2 and/or SCH9 genes ) in the presence of different nutrient mixtures ( Figure 1c ) . Both gene products are known to modulate PKA activity , a known stress response inhibitor , by different means . Ras2 is essential for glucose-dependent PKA activation in nutrient starved yeast cells ( Figure 1c ) , while Sch9 phosphorylates the PKA regulatory subunit in response to Tor1 activation [44] . Notably , the latter has been implicated in dextrose , nitrogen and amino acid metabolism in different organisms [54] . Our results ( Figure 1c ) confirm previous observations [17] on the ability of amino acids to increase stress sensitivity in glucose-derepressed cells . In addition , we confirmed the major role of Ras2 in glucose response and the role of Sch9 as a major amino-acid response transducer . It must be noted that pure amino acid treatment , even for prolonged time ( 24 h , data not shown ) , was completely ineffective in sensitizing cells to peroxide treatment in wild type , ras2Δ , sch9Δor ras2Δsch9Δ genetic backgrounds ( Figure 1c ) . These experiments confirmed previous results [15] , but also provide new insights suggesting that amino acids and dextrose act on separate pathways and that glucose de-repression is necessary for amino acids-dependent cell sensitization . Considering the central role of the AGC kinase Sch9 in amino acid response , stress resistance and aging , we studied its connection with upstream kinases in an attempt to identify the mechanisms linking specific nutrients to Sch9 . AGC kinases are regulated by a general scheme mainly based on the phosphorylation of two amino acid residues which in yeast Sch9 are threonine 570 ( located within the T loop ) and threonine 737 ( located within the hydrophobic motif , HM , Figure 2a ) . Phosphorylation of these residues is accomplished by specific kinases [55]–[59] . In yeast , the PDK1 orthologs Pkh1/2 phosphorylates the T-loop T570 , while the TORC1 complex phosphorylates the HM T737 ( Figure 2a ) [60]–[61] . Western blots using an anti-P570 confirmed the disappearance of the Sch9 immuno-reactive band in protein extracts from strains with deficiencies in both Pkhs ( Pkh1ts , pkh2Δ , see Figure 2b ) ( antibody specificity was confirmed using the T570A mutant ( Figure 2b ) ) , thus confirming previous observations obtained using the Pkhs inhibitor drug myriocin or mutants with impaired sphingolipid biosynthesis ( Lcb1 mutants ) [35] . Site directed mutagenesis , leading to T570A and T737A amino acid substitutions , which abolish T-loop and HM phosphorylation sites , respectively , resulted in increased stress resistance and longevity ( Figure 2c , d ) . Surprisingly , the comparison of the relative effect of the two amino acidic substitutions T570A and T737A uncovered a more important role for the Pkh1/2 T-loop phosphorylation site in stress resistance and aging . T570A increased longevity and stress resistance to an extent similar to that caused by SCH9 deletion ( Figure 2c , d ) whereas the T737A substitution only caused a partial increase in stress resistance and longevity . Cells , with impaired Pkh function , showed increased survival in both the DBY746 and the W303 genetic backgrounds ( Figure 2e , S3b ) . Consistent with this result , resistance to multiple stresses , in both genetic backgrounds , increased when Pkh1/2 function was reduced ( Figure 2e , S3c ) . These results confirmed the role of sphingolipid metabolism on stress resistance and longevity [35] . We then evaluated the stress resistance of DBY746 yeast cells bearing the allele coding for the sch9-T570A or the sch9-T737A in response to different nutrient mixtures ( Figure 2f ) . The results point to a critical role for Sch9 as an integration point of Pkhs- and Tor-mediated amino acid responses ( Figure 2f ) . In addition , the comparison of mutations within the T-loop and the HM domains points to the T-loop and its activators Pkh1–2 as a major pro-aging amino-acid response pathway . To understand the connection between specific amino acids and Sch9 activation by phosphorylation of the T-loop and HM domain , we measured stress resistance in wild type yeast cells treated with mixtures containing minimal medium plus one single non-essential amino acid . The result , shown in Figure 3a , identified the amino acids serine , threonine and valine as the most effective cell-sensitizing amino acids . Previous studies , based on survival in media lacking specific substances , identified glutamic acid as a pro-aging amino acid [38] , [62]–[63] whereas methionine has been described as a pro-aging amino acid in other eukaryotic systems [23] , [64]–[65] . We then tested the ability of combinations between the identified amino acids to increase stress sensitivity . The results confirmed the stress enhancing capability of serine , threonine and valine and to a lesser extent of glutamic acid and methionine ( Figure S4a ) . Since Pkhs , the kinases known to phosphorylate the T-loop of the Sch9 protein , can be inhibited by the drug myriocin , which blocks sphingosine biosynthesis [66] , we compared single amino acid-dependent stress sensitization in the presence/absence of myriocin . Surprisingly , the drug rescued serine sensitization but had no effect on methionine/threonine/valine treatment ( Figure 3b ) . The involvement of the Pkh sphingolipid pathway in serine response was confirmed by a dose-response assay with increasing myriocin concentration at various serine levels ( Figure 3c ) . The results confirmed the myriocin-dependent rescue of the stress resistance , even at very high serine concentrations ( 5× ) . The dependency of serine sensitization on Pkhs function was further confirmed by assessing modulation of serine sensitization in the presence/absence of functional Pkh alleles ( Figure S4b ) . In addition , overexpression of the human ortholog of the Pkhs kinases , PDK1 , increased serine sensitivity suggesting the existence of a conserved role of PDK orthologs in lower and higher eukaryotes ( Figure S4c ) . Chronological life span assay confirmed that serine is a pro-aging amino acid since serine supplementation , at the standard concentration , significantly shortened the life span of wild type yeast strains ( Figure 3d ) . Finally , western blot analysis using the anti phosphothreonine 570 Sch9 specific antibody , confirmed that serine addition was capable of increasing the phosphorylated moiety of the threonine 570 amino acid residue ( Figure 3e ) . In addition , removal of serine from the amino acid mixture was sufficient to significantly decrease the P570 level , pointing to serine as the major amino acid regulator of Sch9T570 phosphorylation ( Figure 3e ) . The possibility that the amino acid administration influences the amount of Sch9 protein rather than its phosphorylation status was ruled out by tagging the Sch9 protein with the hemoagglutinin epitope and using a commercial anti-HA antibody to monitor the amount of the Sch9 protein in the various conditions ( Figure 3e ) . To understand whether Tor1 may serve as the link between the other pro-aging amino acids , Sch9 activation and cellular sensitization , we examined the effect of the three amino acids , identified as the most effective in stress sensitization ( Figure 3a ) , in the presence or absence of the TORC1 drug inhibitor rapamycin . Rapamycin was capable of suppressing the sensitization caused by threonine and valine , but was completely ineffective in suppressing the serine-dependent effects ( Figure 4a ) in a pH-independent manner ( Figure S5a , b ) . In addition , treatment with different nutrient mixtures , in the presence/absence of rapamycin , suggested that TORC1 has a negligible role in dextrose–dependent sensitization and confirmed its capability in suppressing threonine and valine ( MTV ) -dependent sensitization ( Figure 4b ) . Chronological lifespan , obtained in the presence of minimal medium supplemented with either threonine or valine at the standard concentration , confirmed the pro-aging role of these amino acids ( Figure 4c ) . On the other hand , incubation of yeast with standard synthetic medium , restricted ( 1∶10 of the standard concentration ) for only one of the identified pro-aging amino acids , was sufficient to increase the life span and to reduce the rate of spontaneous point mutations ( Figure 4 d , e ) . Finally , the insertion of the point mutation T737A within the hydrophobic motif of the Sch9 protein kinase was capable to abolish the sensitization to oxidative stress due to threonine or valine administration while was ineffective against serine administration ( Figure 4f ) . These results reveal the connection between specific amino acids and two different amino acid-dependent pro-aging pathways: the Tor1-dependent one and the newly discovered Pkhs-dependent one [35] , both converging on Sch9 but on two different phosphorylation sites . Notably , we also demonstrate that restriction of specific non-essential amino acids increases life span and decreases the mutation rate . To understand the connection between specific amino acids and aging , we investigated the localization of the protein kinase Rim15 in response to the presence of the key pro-aging amino acids . Rim15 is a serine/threonine protein kinase whose function is central in G0 arrest [67] and in cellular aging [6] , [68] . Tor1 , Sch9 and PKA control its cellular localization and activity [69] . Observing the fluorescence obtained using a plasmid coding for Rim15-GFP fusion protein , we first confirmed the role of the Sch9 protein kinase in Rim15 cellular localization . In fact , while wild type cells localize Rim15 outside the nucleus in log phase cultures and after nutrients addition to stationary phase cultures , the lack of SCH9 caused Rim15 nuclear localization in day 2 yeast cells ( Figure 5b ) and after nutrient re-feeding ( Figure S6c ) . In addition , wild type post-diauxic yeast cells ( day 2 ) were exposed to nutrient re-feeding and Rim15-GFP fluorescence was monitored overtime . We observed an initial re-localization ( appearance of the granules ) followed by a time-dependent decrease ( 2–6 hrs interval ) of the Rim15-GFP fluorescence ( Figure 5a upper panel ) but only with wild type Sch9 protein ( Figure S6c ) . To characterize the nature of the granules , appearing after nutrient re-feeding , we co-expressed , in wild type cells , the plasmid coding for Rim15-GFP fusion protein together with the plasmid coding for the Pab1-RFP fusion protein , the latter being a known marker for stress granules [70] . Fluorescence analysis , 2 hours after nutrient replenishment in day 2 cultures , showed co-localization of the two fusion proteins indicating Rim15 localizes within stress granules after nutrients supplementation ( Figure 5c ) . To determine if the Rim15-GFP disappearance , observed after longer nutrient incubation , was due to increased protein degradation or silenced transcription , we performed quantitative PCR of Rim15 transcript on RNA extracts , obtained at different time points , after nutrient replenishment . The results confirmed the role of reduced RNA levels in Rim15 activity regulation ( Figure 5d ) . We then monitored if the addition of the identified pro-aging amino acids was sufficient to obtain Rim15 cellular re-localization . Addition of single amino acids to the otherwise minimal medium revealed the central role of threonine , serine and valine in activating the Rim15 re-localization pathway ( Figure 5a lower panel ) but only in glucose de-repressed cells ( not shown ) , thus suggesting that the Tor1 and Pkh pathways may directly or indirectly regulate Rim15 protein activity . On the other hand Rim15 is reported to activate stress resistance transcription factors Msn2/4 and Gis1 . It is known that Msn2/4 transcription factors bind to the STRE ( STress Responsive Elements ) motif contained in the promoter of stress responsive genes [71]–[73] whereas Gis1 transcription factor binds to the PDS ( Post Diauxic Shift ) motif contained in the promoter of several genes activated at this metabolic transition [74]–[76] . We tested the effect of nutrients on isogenic yeast strain carrying single , double or triple deletions of these transcription factors ( Figure 6a ) . Our experiments revealed a central role for Rim15 and Msn2/4 but a much more modest effect of the deletion of Gis1 transcription factors alone on the glucose- and amino acids-dependent stress sensitization ( Figure 6a ) . However , the unexpected high stress resistance and STRE element activation in the gis1Δ strain is likely due to a compensatory Msn2/4 activation in the presence of a gis1 null allele . In fact , the Msn2/4-dependent STRE/beta galactosidase activity was increased in the gis1Δ null cells ( Figure S6a ) but deletion of both Msn2/4 and Gis1 caused a sensitization similar to that observed in rim15 deletion mutants ( Figure 6a ) . Quantitative PCR of the prototype stress response gene SOD2 confirmed the role of amino acid administration in controlling stress response gene expression but only in the presence of the Rim15 protein kinase ( Figure 6b ) . Beta-galactosidase assay using PDS ( Gis1-dependent ) or STRE ( Msn2/4 ) gene reporters in various nutrient conditions indicated the contribution of amino acids to inhibition of both STRE and PDS-dependent gene transcription ( Figure 6c ) . The usage of single , as well as , mixtures of the most sensitizing amino acids confirmed the role of serine and threonine/valine dependent pathways in regulating mainly PDS-driven gene expression ( Figure 6c ) .
The understanding of the mechanisms linking DR to its anti-aging effects in higher eukaryotes has been hindered by their complexity . S . cerevisiae provides a very simple organism in which the effect of each major pro-aging nutrient can be dissected . In the present work we describe the cooperation between glucose and the amino acids threonine , valine and serine , in sensitizing yeast cells to stress and promoting aging via two major pathways . These results enhance our understanding of previously poorly understood roles and interactions between specific nutrients to promote aging but also point to specific amino acids and their effect on different pathways previously established to cause aging . By contrast , the effect of methionine restriction in extending the life span of fruit flies [23] and rodents [77]–[80] does not appear to be as important for the protection of yeast cells . We also show a less potent than serine , valine and threonine but detectable role of glutamate in increasing stress sensitivity in agreement with data on the ability of its deficiency to extend yeast survival [62]–[63] . Our genetic and biochemical analysis revealed that the yeast amino acid response relies on at least two different pathways: the well-characterized TORC1-S6K pathway , which has been described as an integrator of different nutrient and energy signals in organisms including humans [81]–[82] and the sphingolipid-dependent Pkh1/2 pathway , also shown to promote aging in yeast ( Figure 7 ) [35] . Threonine and valine activated the TORC1 pathway and promoted cellular sensitization that could be reversed by the well-established anti-aging drug rapamycin , whereas serine specifically activated Pkhs and promoted cellular sensitization by a mechanism , which could be reversed by the Pkh inhibitor myriocin . Considering that L-serine is the substrate of the serine palmitoyltransferase , the α-oxamine synthase enzyme that catalyzes the condensation reaction of L-serine and palmitoyl-CoA to form 3-ketodihydrosphingosine , we propose that serine may activate Pkh by enhancing the first and rate-determining step of the sphingolipid biosynthesis pathway [83] . Thus , serine administration may be equivalent to enhance sphingolipid biosynthesis , which is known to activate Pkhs and promote aging [35] . This hypothesis is supported by the overlap between the treatment with the sphingolipid biosynthesis inhibitor myriocin and Pkhs impairment . Our analyses demonstrate that Sch9 plays a critical role in these nutrient response mechanisms since it integrates signals from the two pathways adapting its phosphorylation status and activity accordingly . This may explain the dominant role of the Tor/Sch9 pathway in promoting aging in yeast and possibly in higher eukaryotes . Rim15 subcellular localization also primarily relied on specific amino acid availability ( threonine , serine and valine ) . Thus , although activated Rim15 is clearly central in nutrient-dependent cellular protection , its regulation by amino acids appears to be complex and to depend on both the Pkh and Tor/S6K pathways . In agreement with our hypothesis partial depletion of each of the amino acids described in this work was capable of increasing the life span . As expected from our previous studies , transcription factors Msn2/4 and Gis1 are downstream mediators of the anti-aging effects of Rim15 ( Figure 6 ) . In summary , these results shed new light on a nutrient response network in which different genes are linked to specific components of the diet . Because orthologs of many genes in this network are known to affect aging in higher eukaryotes , these results are likely to point to similar mechanisms in mammalian cells .
Strains and plasmids used in this study are listed in table S1 . Gene Knockouts were generated by one-step gene disruption [84] . The sch9T570A and sch9T737A mutants were constructed using plasmids PFR82 and pAM202 ( kindly provided by Dr . Thorner , University of California , Berkeley [57] ) by homologous integration at the SCH9 locus . Cells were grown in YPD ( 1% yeast extract , 2% peptone , 2% glucose ) , minimal medium SDC ( 0 . 17% yeast nitrogen base , 0 . 5% ammonium sulfate , 0 . 08% amino acids , pH 6 ) or selective media , with appropriate amino acids content ( see table S2 for a complete list ) to maintain selection for plasmids , containing 2% glucose as carbon source . Cells were grown at 30°C . Overnight SDC cultures were diluted ( 1∶10 ) into flasks covered with aluminum foil caps with fresh SDC medium to a final volume of 10 ml ( with a flask to culture ratio of 5∶1 ) and kept at 30 °C with shaking ( 200 rpm ) to ensure proper aeration . This dilution time point was considered day 0 . Every other day , aliquots from the culture were properly diluted and plated onto YPD plates . The YPD plates were incubated at 30 °C for 2–3 days . Viability was assessed by Colony Forming Unit ( CFUs ) count . The CFUs obtained at day 2 were considered to be the initial survival ( 100% ) . In situ viability was done as previously reported [53] . Briefly , aliquots of a two days old liquid culture of a trp- strain are plated on many plates of synthetic medium containing the indicated mixture of amino acids but lacking tryptophan . Plates are incubated at 30 °C and no growth is observed due to the lack of tryptophan . Every two days , plates are supplemented with the appropriate amount of tryptophan and put back in the incubator where cells start now dividing . Colony forming units are registered after two more days of incubation and scored as the percentage of CFU with respect to CFU at day 2 , the latter considered as 100% of survival . Mutation frequency was evaluated , as previously described [53] , by monitoring the percentage of cells that become resistant to canavanine treatment during chronological life span experiment . Heat shock resistance was measured by spotting serial dilutions of cells removed from SDC cultures onto YPD plates and incubating at either 55 °C ( heat-shocked ) or 30 °C ( control ) for 60 minutes to 120 minutes . After the heat-shock , plates were transferred to 30 °C and incubated for 2–3 days . For oxidative stress resistance assays , aliquots of cells were diluted in K-phosphate buffer 0 . 1M , pH 6 , and treated with different concentrations of hydrogen peroxide for 30 minutes . Serially diluted cells were then spotted onto YPD plates and incubated at 30 °C for 2–3 days . Overnight SDC cultures were diluted 1∶10 into fresh SDC medium and were maintained at 30°C with shaking until day 2 or 3 of growth , then cultures were disposed into 24-multiwell plates and centrifuged at 3500 rcf for 20 minutes , pellets were re-suspended into different fresh media , that contained different mix of amino acids and/or different glucose concentrations , cells were incubated at 30°C shaking for 4 hours ( amino acids used for the mixes and their concentrations are listed in table S2 ) . After incubation cells were pelleted and re-suspended in K-phosphate buffer 0 . 1M , pH 6 and treated with hydrogen peroxide for 30 minutes . Cells from each well were then spotted onto YPD plates and incubated at 30 °C for 2–3 days ( for a scheme of the nutrient response assay see fig . S1 ) . Protein extract were prepared by glass-bead disruption in a protein extraction buffer ( 50 mM MES KOH pH 6 . 2 , 0 . 05 mM EDTA , 0 . 1 mM MgCl2 , 0 . 5 mM DTT , 1× Protease inhibitor mix ( Sigma ) , 1 mM PMSF , 25 mM NaF , 10 mM NaN3 , 10 mM sodium beta-glycero-phosphate , 10 mM Na2H2P2O7 ) . The lysate was spun down at 5000 rpm for 15 minutes and the supernatant tested . Protein concentrations were determined using Bradford Assay . Proteins were separated by SDS-PAGE [85] . Resolved proteins on gels were transferred onto nitrocellulose membrane ( Schleicher & Schuell ) using 192 mM glycine , 25 mM Tris , 20% methanol in a Mini Trans-Blot Electrophoretic Transfer Cell ( Biorad ) . Blots were then blocked with 1% bovine serum albumin ( BSA ) in 20 mM Tris , 0 . 5M NaCl , pH 7 . 5 ( TBS ) , washed in TBS with the addition of 0 . 05% Tween-20 ( TBST ) and incubated over night with primary antibody . Membranes were washed in TBST and incubated for 2 hours with secondary antibody ( AP or HRP conjugated ) , then washed in TBS and labeled with BCIP/NTB or chemiluminescent ECL liquid substrate system ( Promega and Invitrogen respectively ) . The primary antibodies used were: anti-BCY1 ( goat polyclonal , Bcy1 [yN19] sc-6765 , Santa Cruz Biotechnology ) , anti-HA ( mouse monoclonal , HA probe sc-7392 Santa Cruz Biotechnology ) , anti-Sch9P570 ( a kind gift of Robby Loewith ) . Cells expressing Rim15-green and/or Pab1-red fluorescent fusion proteins were grown to stationary phase , treated with different nutrients mixtures for various times , then cells were used for fluorescence microscopy directly without fixation . Nuclei were stained with 0 . 5 ug/ml of Hoechst 33342 ( Invitrogen ) for 15 minutes before cells were watched . Cells were viewed with an Olympus BX50 fluorescence microscope using the appropriate filters . Total RNA was isolated using RiboPure- Yeast kit ( Ambion ) according to the kit's instructions . RNA was treated with RNase-free DNase I ( Promega ) to remove contamination of genomic DNA . 0 . 5 µg of total RNA was reverse transcribed into cDNA using ImProm-II Reverse Transcriptase ( Promega ) with sequence specific primers ( ACT1 , GAATCCAAAACAATACCAGTAG; SOD2 , AGCTGCTAATTTAACCAAGAAG; RIM15 , TTATCGTACTTTCATCGTCAC ) . Quantitative PCR experiments were performed on StepOne Real-Time PCR instrument ( Applied Biosystems ) using Fast SYBR Green Master Mix ( Applied Biosystems ) and the gene specific primers: ACT1 , fw-TCGTGCTGTCTTCCCATCTATC and rev-GTAGAAGGTATGATGCCAGATC; SOD2 , fw-CTCCGGTCAAATCAACGAAT and rev-CCTTGGCCAGAAGATCTGAG; RIM15 , fw-GGAGCTGGAACTGGACGGCAAG and rev-AGCATGTCTGTGGCCTTTTGAA . Thermo-cycling conditions were as follows: 95 °C for 20 seconds followed by 40 cycles of 95 °C for 3 seconds and 60 °C for 30 seconds . Relative gene expression was calculated using the 2−ΔΔCT method and normalized to ACT1 mRNA levels . Cell pellet from 1 ml of culture was lysed with low salt buffer ( 50 mM Tris pH 7 . 5 , 0 . 1× protease inhibitor cocktail ( Sigma ) , 100 mM NaCl , 2 mM EDTA , 2 mM EGTA , 50 mM NaF ) . The protein concentration of the lysate was determined by Bradford assay . 55 µl of appropriately diluted samples of lysate was mixed with 85 µl of substrate solution ( 1 . 1 mg/ml ONPG in 60 mM Na2HPO4 , 40 mMNaH2PO4 , 10 mM KCl , 1 mM MgSO4 , 50 mM 2-mercaptoethanol , pH 7 . 0 ) . Absorbance at 420 nm was read every 5 minutes until 30 minutes after the initiation of reaction . Percentage of activity at every condition was determined respect of a control condition fixed as 100% of activity . Statistical significance was evaluated by 2 tailed T-test for groups with unequal variants , control versus individually all the others conditions .
|
Calorie restriction ( CR ) , but also the restriction of specific components of the diet , has been known for decades to affect longevity . However , the understanding of how each component of the macronutrients affects longevity and stress resistance is poorly understood , in part because of the complexity of many of the model organisms studied . Here we studied how each amino acid and glucose cooperate to activate cell sensitizing pathways and promote aging . We identified specific amino acids in the diet that affect cellular protection and aging , describe how different pathways mediate these pro-aging effects , describe the effect of glucose and specific amino acids on the levels/activity of stress resistance kinases and transcription factors , and identify specific nutrient depletions capable of increasing longevity and stress resistance . Because of the conserved pro-aging role of orthologs of many of the genes in the signaling network described in this paper , these results are likely to serve as a foundation for the elucidation of similar nutrient-dependent pro-aging mechanisms in mammals .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"signal",
"transduction",
"signaling",
"in",
"cellular",
"processes",
"model",
"organisms",
"cellular",
"stress",
"responses",
"adenylyl",
"cyclase",
"signaling",
"pathway",
"mechanisms",
"of",
"signal",
"transduction",
"genetics",
"tor",
"signaling",
"ras",
"signaling",
"signaling",
"pathways",
"biology",
"microbiology",
"molecular",
"cell",
"biology",
"gene",
"networks"
] |
2014
|
Serine- and Threonine/Valine-Dependent Activation of PDK and Tor Orthologs Converge on Sch9 to Promote Aging
|
Accuracy of transcript quantification with RNA-Seq is negatively affected by positional fragment bias . This article introduces Mix2 ( rd . “mixquare” ) , a transcript quantification method which uses a mixture of probability distributions to model and thereby neutralize the effects of positional fragment bias . The parameters of Mix2 are trained by Expectation Maximization resulting in simultaneous transcript abundance and bias estimates . We compare Mix2 to Cufflinks , RSEM , eXpress and PennSeq; state-of-the-art quantification methods implementing some form of bias correction . On four synthetic biases we show that the accuracy of Mix2 overall exceeds the accuracy of the other methods and that its bias estimates converge to the correct solution . We further evaluate Mix2 on real RNA-Seq data from the Microarray and Sequencing Quality Control ( MAQC , SEQC ) Consortia . On MAQC data , Mix2 achieves improved correlation to qPCR measurements with a relative increase in R2 between 4% and 50% . Mix2 also yields repeatable concentration estimates across technical replicates with a relative increase in R2 between 8% and 47% and reduced standard deviation across the full concentration range . We further observe more accurate detection of differential expression with a relative increase in true positives between 74% and 378% for 5% false positives . In addition , Mix2 reveals 5 dominant biases in MAQC data deviating from the common assumption of a uniform fragment distribution . On SEQC data , Mix2 yields higher consistency between measured and predicted concentration ratios . A relative error of 20% or less is obtained for 51% of transcripts by Mix2 , 40% of transcripts by Cufflinks and RSEM and 30% by eXpress . Titration order consistency is correct for 47% of transcripts for Mix2 , 41% for Cufflinks and RSEM and 34% for eXpress . We , further , observe improved repeatability across laboratory sites with a relative increase in R2 between 8% and 44% and reduced standard deviation .
RNA-Seq has established itself as a popular alternative to microarrays for the quantification of RNA transcripts . In contrast to microarrays , which measure the quantity of an RNA transcript by hybridization to a transcript specific oligonucleotide , RNA-Seq generates cDNA for fragments of the RNA transcript , which are sequenced by a next generation ( NGS ) sequencer . One advantage of RNA-Seq over microarrays is that it does not require prior knowledge of the nucleotide sequence of the RNA transcript , which is needed to produce a transcript specific hybridization probe , and that it can therefore detect and quantify novel RNA transcripts . In addition , quantification by RNA-Seq covers a wider dynamic range since microarrays suffer from signal saturation resulting in the truncation of abundance estimates for highly abundant transcripts [1] . Despite these advantages , obtaining accurate transcript quantification measurements from RNA-Seq has proven difficult . One of the main reasons for the inaccuracy is the failure of the statistical models used in the derivation of the measurements to properly represent biases inherent in RNA-Seq data . The statistical model of the original version of Cufflinks [2] , for instance , assumes that the cDNA fragments generated by RNA-Seq are uniformly distributed along the transcripts . In reality , however , this assumption is rarely fulfilled and quantification measurements by this version of Cufflinks are therefore often inaccurate . One type of bias affecting transcript quantification from RNA-Seq data is the result of a preference of the fragmentation , i . e . the process that generates cDNA fragments from RNA transcripts , to produce fragments at certain positions within the transcript , e . g . at the start and/or at the end of the transcript [3] . Hence , this type of bias is referred to as positional bias [4] . Positional bias can also be caused by a bias in the RNA itself , for instance , due to RNA degradation which results in a shortening of the RNA . Another kind of bias in RNA-Seq is introduced during ligation , amplification and NGS sequencing [5] . This bias is correlated to the RNA sequence of a transcript and is therefore called sequence specific bias [4] . The present article focuses on the first type of bias , i . e . the positional bias , and develops a model , Mix2 ( rd . “mixquare” ) , which learns the positional bias in RNA-Seq data . In our experiments we compare Mix2 to Cufflinks [2 , 4] , eXpress [6] , RSEM [7] and PennSeq [8] both on synthetic data and on real RNA-Seq data [9] generated from the Universal Human Reference ( UHR ) and Human Brain ( HBR ) samples of the Microarray Quality Control ( MAQC ) experiment [10] . The inclusion of bias models into the statistical models of RNA-Seq data has been investigated before . In [11] a model is proposed to account for the variability in read counts depending on the sequence surrounding the start of a fragment . The intention is similar to that of the fragment specific bias model [4] , which has been implemented as an extension to Cufflinks [2] . In addition , [4] includes a non-parametric positional bias model , which can theoretically be trained with the EM algorithm . However , due to the large number of variables , this is only feasible for few transcript length dependent classes , for which statistics are collected in a small number of positional bins . As a result , [4] implements a positional bias model depending exclusively on the length of a transcript . Similar to [4] the generative model of RSEM [7 , 12] uses a hidden variable for the positional bias , where the latter is estimated from the global bias observed in the complete RNA-Seq data set . Also in RSEM , therefore , does the positional bias model depend exclusively on the transcript length . The generative model of eXpress [6] differs from Cufflinks mainly in the order of fragment-length selection and implements an online rather than a batch EM algorithm . The implementation described in [6] is further restricted to a sequence specific bias , with a uniform positional bias similar to Cufflinks . The model developed in PennSeq [8] is again non-parametric and the large number of variables makes its training computationally prohibitive . For this reason , the bias model of PennSeq [8] is not included in the parameter update but is approximated by the overall bias in a gene locus and by the transcript specific reads . The method described in [13] is a model for gene read counts , which models bias by exon specific weights , which are estimated both for the complete data set and for individual genes . In [14] the authors focus on RNA-Seq data with 5’ bias which is the result of RNA degradation and use an exponential model for the fragment distributions . The model proposed in [15] is , again , a model for the read counts of a gene . Here the read counts are modelled by a quasi-multinomial distribution with a parameter that can be adapted to account for over and under dispersion . Mix2 is , similar to [2 , 4 , 6–8 , 12] , a generative model for the probability of a fragment in an RNA-Seq data set . In comparison , however , the model for the positional fragment bias in Mix2 is parametric . This considerably simplifies its implementation removing the need for any restrictions of the non-parametric methods . At the same time , the model of the positional fragment bias in Mix2 is very versatile since mixtures of probability distributions can approximate distributions of arbitrary complexity . Section ( Materials and methods ) develops the theory of Mix2 in greater detail and shows how its parameters can be updated with the EM algorithm leading to simultaneous estimates for transcript abundances and transcript specific positional fragment biases . Section ( Experiments on artificial data ) optimizes the number of mixture components of the Mix2 model and compares it with Cufflinks , RSEM and eXpress on artificial data sets . Sections ( Experiments on the Microarray Quality Control ( MAQC ) data ) and ( Experiments on the Sequencing Quality Control ( SEQC ) data ) on the other hand , discuss experiments on two publicly available real RNA-Seq data sets with Mix2 , RSEM , eXpress , Cufflinks and PennSeq . These experiments show that , in comparison to the other methods , Mix2 leads to better correlation between estimated and measured transcript concentrations , correct recovery of mixing ratios and yields consistent titration orders . In addition , we show that the Mix2 concentration estimates are repeatable across lanes and laboratory sites and lead to a more accurate detection of differentially expressed transcripts . In addition , Section ( Types of bias in the MAQC data ) shows that Mix2 can be used as an explorative tool to detect positional biases present in RNA-Seq data . Mix2 has been implemented as an Octave script with readable code and as a closed source C++ implementation . We used the latter for the majority of our experiments . Both versions can be downloaded from https://www . lexogen . com/mix-square-scientific-license . While the C++ version is considerably faster , we show at the end of Section ( Experiments on artificial data ) and in Fig E in S2 Appendix that quantification results for both implementations are virtually identical . Hence , either version can be used to evaluate the accuracy of Mix2 .
An essential part of Next Generation Sequencing ( NGS ) is the library preparation . This process takes an RNA sample and produces a library of short cDNA fragments , each corresponding to a section of an RNA transcript . The cDNA fragments are sequenced by an NGS sequencer resulting in single or paired end reads which are mapped to a reference genome . Hence , the probability p ( r ) of a fragment r can be interpreted as the probability of its genomic coordinates . In a genomic locus the probability p ( r ) is the superposition of the fragment distributions p ( r|t = i ) for the N transcripts in the locus , i . e . p ( r ) = ∑ i = 1 N α i p ( r | t = i ) ( 1 ) where αi is the relative abundance of transcript t = i , i . e . the probability that transcript t = i generates any fragment , and p ( r|t = i ) is the probability that transcript t = i generates fragment r . Hence p ( r|t = i ) models the transcript specific fragment bias . An estimate for the concentration of transcript t = i is obtained by normalizing the relative abundance αi , yielding the RPKM [16] or FPKM values [2] . Mix2 uses a mixture model for p ( r|t = i ) , i . e . p ( r | t = i ) = ∑ j = 1 M β i j p ( r | t = i , b = j ) ( 2 ) where p ( r|t = i , b = j ) are the M components of the mixture and the βij are the non-negative component weights . Hence , p ( r|t = i , b = j ) is a probability distribution over r and ∑ j = 1 M β i j = 1 . ( 3 ) Since p ( r ) is itself a mixture of the p ( r|t = i ) with weights αi , this implies that p ( r ) is a mixture of mixtures motivating the name of Mix2 . For the p ( r|t = i , b = j ) we use Gaussians placed equidistantly along the transcript t = i ( Materials and methods ) . An example for such a Mix2 model can be found in Fig 1 . Fig 1 ( a ) shows the mixture weights βij whereas Fig 1 ( b ) shows the weighted Gaussians , βijp ( r|t = i , b = j ) , and the sum of the weighted Gaussians , p ( r|t = i ) . The distributions in Fig 1 ( b ) are given in transcript coordinates for a transcript of 2000 bp length , while the longer dashed curve in Fig 1 ( c ) shows p ( r|t = i ) in genome coordinates . The locus in Fig 1 ( c ) contains two transcripts which share a common junction . The shorter of the transcripts in Fig 1 ( c ) has the same set of βij as in Fig 1 ( a ) but is only 1000 bp long . The relative abundances of the long and short transcript are 0 . 7 and 0 . 3 , respectively , which results in the overall distribution p ( r ) given by the solid curve in Fig 1 ( c ) . In comparison , Cufflinks [2] can , for this locus , only model fragment start distributions p ( r|t = i ) as visualized by the dashed curves in Fig 1 ( d ) and is therefore inappropriate for 5’ biases as the one in Fig 1 ( c ) . We use the Expectation Maximization ( EM ) algorithm ( Materials and methods ) to learn the parameters of Mix2 from RNA-Seq data . This results in simultaneous estimates of the transcript abundances αi and the mixture weights βij , hence the transcript specific fragment distributions p ( r|t = i ) . Mix2 is identifiable in most cases and can otherwise be easily made identifiable ( see Section 1 . 2 in S1 Appendix ) . It can therefore always be ensured that the EM algorithm converges to the unique maximum likelihood solution . It should be pointed out that we did not check for identifiability in our implementation and , like Cufflinks , returned the maximum likelihood solution produced by our method . The mixture weights βij determine the shape of the fragment distribution of transcript t = i . Thus , if transcripts have a similar distribution they should share the same βij . This results in their fragment distributions being identical and reduces the number of parameters in Mix2 making it less prone to over-fitting . Consider , for instance , the fragment start distribution of the Cufflinks model in Fig 2 ( a ) . Here the distributions are similar for transcripts with 2000 bp and 3000 bp length and for transcripts with 700 bp and 1000 bp length . In this situation therefore , these four transcripts can be separated into two groups where the transcripts within each group share the same mixture weights βij . In general , this leads to the scenario where each transcript t = i has an associated group g = k and the distributions p ( r|t = i ) of transcripts within this group share the same βij . Multiple factors might influence the similarity of fragment start distributions . Here , we investigate gene membership and , as the bias correction methods in [4 , 12] , transcript length . The rationale for choosing these two properties is that even if fragments are uniformly distributed immediately after fragmentation , fragment size selection introduces the transcript length dependent bias in Fig 2 ( a ) . On the other hand , transcripts belonging to the same gene can share a substantial part of their sequence and exhibit therefore potentially similar fragmentation properties . In the following , we refer to tying between all transcripts within a gene as global tying and to tying between all transcripts within a gene and the same length range as group tying ( Materials and methods ) . Hence , Fig 1 ( c ) shows an example for a Mix2 model with global tying since both transcripts share the same set of weights βij . In our experiments we do not treat multi-mapping and uniquely mapping reads differently and instead consider each read mapping to be associated with a unique fragment . Multi-mapping reads and sequence specific bias can , however , be integrated into Mix2 as shown in Section 1 . 1 in S1 Appendix . This section pursues two goals . First , we find a sensible number of mixture components for each of the variants of Mix2 . Second , we investigate the performance of Mix2 under conditions favoring other quantification methods . The number of mixture components derived in this section is used as a default by Mix2 . While this might be suboptimal in some cases it is a necessary compromise since reference measurements of isoform concentrations on which to optimize the number of mixture components are usually not available in RNA-Seq data . In our first experiments we studied the 4 transcript length dependent biases in Fig 2 . These resemble the biases we detected in our experiments on real RNA-Seq datasets MAQC and SEQC . The most dominant biases in the MAQC data are visualized in Fig 3 . The 5’ biased fragment distributions in Fig 3 ( a ) and 3 ( b ) , for instance , resemble the biases in Fig 2 ( b ) and the biases for short transcripts in Fig 2 ( a ) . Similarly , the fragment distributions concentrated on the 3’ side in Fig 3 ( e ) resemble the biases in Fig 2 ( c ) . While the biases in Fig 3 do not depend on transcript length as strongly as the biases in our artificial data , we see with the exception of Fig 3 ( d ) an increase in the average transcript length with increasing 3’ bias . Overall , around 20 . 16% of transcripts have a 5’ bias in the MAQC data , 26 . 34% have a 3’ bias and 26 . 92% a uniform fragment distribution . Hence , the biases in our artificial data represent a sensible starting point for evaluating the accuracy of quantification methods under real-life conditions . Fig 2 also illustrates how an incorrect choice of bias model can affect the accuracy of quantification estimates . If the 5’ bias in Fig 2 ( b ) is observed for a transcript of length 3000 bp but the bias model in Fig 2 ( a ) is used for quantification then the incorrect model will try to explain the bias in Fig 2 ( b ) with a shorter transcript , thus potentially leading to an overestimate of the concentration of the short transcript . In our experiments we used a set of 7 test genes from the GRCh37/hg19 Ensembl annotation v75 containing between 4 and 15 transcripts as well as the main variants of differential splicing , see Table A in S2 Appendix . While this set of genes might seem small , considering all parameter combinations resulted in 538k experiments . The difference between true and estimated relative isoform abundances is measured with the L1 distance , which is the sum of the absolute differences between true and estimated relative isoform abundances . This value lies between 0 and 2 . In our experiments in Fig 4 ( a ) we addressed the question if the optimal number of mixture components for Mix2 depends on the gene , the number of fragments in the gene or the fragment bias . In the first three bars of Fig 4 ( a ) we optimized the number of mixture components for each combination of bias , gene and sample size separately and plotted the average L1 distance over all experiments for the three types of Mix2 . In the second group of bars we optimized the number of mixture components for each combination of gene and sample size ignoring the bias , while in the third group we optimized for each gene ignoring bias and sample size . Finally , for the last three bars in Fig 4 ( a ) we selected a single number of mixture components for all experiments independent of gene , bias and sample size . Overall , Fig 4 ( a ) shows that little is gained by optimizing for each gene , sample size and bias separately . The number of mixture components of Mix2 can be chosen independent of these factors . Fig 4 ( b ) shows the influence of the number of mixture components on the quantification accuracy of Mix2 . For each number of mixture components on the x-axis the average L1 distance is given for different sample sizes . Each group of 4 blocks contains results for 500 , 1000 , 5000 and 10000 fragments , where the different groups show results for the different types of Mix2 . Fig 4 ( b ) implies that the minimal L1 distance for Mix2 without tying is obtained for 3 mixture components , while for Mix2 with group and global tying the minimum is obtained with any number between 4 and 10 . In all our subsequent experiments we therefore chose 3 mixture components for Mix2 without tying and 4 mixture components for Mix2 with global and group tying . We further studied the convergence of the fragment start distributions p ( r|t = i ) estimated by Mix2 to the correct start distributions from which the data was sampled . For this purpose , we calculated the average L1 distance between estimated and true distribution for the transcripts within a gene weighted by the estimated relative abundance of the transcript . Fig 4 ( c ) and 4 ( d ) visualize the difference between the L1 distance before ( tall bars ) and after estimation of the fragment start distributions . The initial fragment start distribution in Mix2 is close to uniform and independent of tying structure and read depth ( Materials and methods ) . Fig 4 ( c ) shows that estimates of the fragment start distributions in Mix2 converge to the correct solution since the L1 distance for the 7 test genes decreases during the course of the parameter estimation by between 50% and 65% . Mix2 with group tying and a read depth of 10000 fragments achieves in many cases the smallest final L1 distance . In Fig 4 ( d ) the relative decrease in L1 distance is considerably smaller for the Cufflinks bias as the initial fragment start distribution in Mix2 is already close to the uniform Cufflinks distribution . Overall , the final L1 distance is similar for all 4 biases . Next , we compared the accuracy of Mix2 , Cufflinks , RSEM , and eXpress on the full transcriptome . We generated an exponential transcript expression profile from an exponential gene expression profile by FluxSimulator [17] , as shown in Fig F and Fig G in S2 Appendix , and divided the gene reads between the gene transcripts according to relative abundances drawn from a uniform Dirichlet distribution . In total we generated around 50 mio 100 bps read pairs for each bias in Fig 2 , which we aligned with Tophat2 . Data of this kind correspond to the positional bias models in Cufflinks [4] and RSEM which are derived by scaling a single prototype distribution to the transcript length . Our experiments therefore show whether Mix2 can learn data structures hard-coded into the statistical models of Cufflinks and RSEM . Fig 4 ( e ) shows that on 5’ and 5’+3’ bias RSEM slightly outperforms Mix2 with group tying , while for 3’ bias Mix2 yields the best results . On Cufflinks bias all methods apart from eXpress perform similarly . Fig 4 ( e ) further contains results for Mix2 with initial fragment start distributions before estimation of mixture weights βij ( Mix2 init ) . This shows that adapting the mixture weights improves relative abundance estimates . For Cufflinks bias this improvement is minor since the initial fragment start distributions in Mix2 are already close to the correct solution . The bad performance of eXpress is likely due to the fact that it models only sequence-specific and not positional bias . Fig 4 ( f ) shows that Mix2 without any tying and with group tying perform consistently across biases and slightly better than RSEM with mean L1 distance of 0 . 36 versus 0 . 39 . For Mix2 without parameter tying this is remarkable , as its statistical model is completely assumption free with regards to the nature of fragment bias . We also performed experiments for a smaller read-depth of 5 mio read pairs . These experiments which are summarized in Fig H in S2 Appendix show an increase of L1 distance for all methods but otherwise a similar trend as the experiments with 50 mio read pairs . Finally , we compared the accuracy of the Octave and C++ version of our code , since on larger data sets we used , for efficiency reasons , the closed source C++ version . We found on our 7 test genes that the median L1 distance between the relative abundances estimated by the two versions was between 0 . 01 and 0 . 004 for 500 and 10k reads per gene . The boxplots for these experiments can be found in Fig E in S2 Appendix . Due to the small difference , either version of our code can be used to evaluate the accuracy of Mix2 . This section compares the accuracy of Mix2 , Cufflinks [4 , 18] , PennSeq [8] , RSEM [7] and eXpress [6] on two publicly available real RNA-Seq data sets generated from the Universal Human Reference ( UHR ) RNA and human brain ( HBR ) RNA of the Microarray Quality Control ( MAQC ) data [10] ( Materials and methods ) . Both Cufflinks and RSEM were run with bias correction . In accordance with our experiments on artificial data we used three mixture components for Mix2 with no tying and 4 mixture components for Mix2 with global and group tying . We treat the qPCR measurements as a reference and compare them to the FPKM values generated by the quantification methods . The qPCR concentrations in the MAQC data are unevenly distributed spanning several orders of magnitude . It is therefore customary to compress the concentrations by taking the logarithm . This achieves a more even distribution but leads , on the other hand , to outliers for FPKM values close to zero whose logarithm approaches minus infinity . In order to reduce the influence of these outliers on the quality metrics for the quantification methods it is necessary to either remove transcripts with small FPKM values or to truncate the logarithm of their FPKM values . We chose the latter strategy since the former ignores the fact that some quantification methods fail to detect highly abundant transcripts and furthermore leads to test sets varying significantly in size reducing the comparability of the quality metrics . In our experiments , the logarithm of an FPKM value was truncated if it was below the first quartile minus 1 . 5 times the interquartile range of the logarithms of the method’s FPKM values on the test set . This threshold , which was between 0 . 01 and 0 . 001 in our experiments , is the point below which values are often considered to be outliers in boxplots . We use the logarithm to the basis of 10 in our experiments . In order to evaluate the quantification methods on more recent data than MAQC , we also performed experiments on the SEQC data set [25] . The latter contains 100 bps paired-end RNA-Seq reads generated with Illumina HiSeq 2000 at 6 laboratory sites and from four different RNA samples . Samples A and B correspond to UHR and HBR from the MAQC data set . Sample C and D were created by mixing A and B in 3:1 and 1:3 ratios , respectively . This allows tests for titration order consistency and the correct recovery of mixing ratios . These tests are independent of a “gold standard” such as qPCR which is biased by its own technical limitations . In this section , therefore , we evaluate the accuracy of quantification with respect to the built-in ground truths of SEQC rather than with respect to qPCR measurements . For this purpose , we downloaded RNA-Seq data for samples A to D for laboratory site BGI . To study repeatability across sites we further downloaded RNA-Seq data for sample A and sites AGR , CNL and COH . We compare the three variants of Mix2 with Cufflinks , RSEM and eXpress . PennSeq had to be excluded as it failed to produce any output on the SEQC data . As before , we used GRCh37/hg19 and Ensembl annotation version 75 in our experiments and 3 mixture components for Mix2 without tying and 4 mixture components for Mix2 with global and group tying .
This article introduced Mix2 which uses a mixture of probability distributions to model transcript specific positional fragment distributions in RNA-Seq data . Due to the flexibility of mixture models , Mix2 can adapt to multiple positional fragment biases of arbitrary complexity . The parameters of Mix2 are efficiently trained with the EM algorithm resulting in simultaneous estimates for fragment distributions and relative abundances . In addition , parameters of Mix2 can be tied between transcripts with similar fragment distribution leading to improved estimates of the relative abundances . Even though Mix2 accommodates a sequence specific bias , we currently implement only a model for positional bias . Sequence specific bias can , however , be a prominent feature in RNA-Seq data . In [11] , for instance , it was found that a linear model on the sequences of 3 RNA-Seq data sets accounted for between 40% and 50% of the variance in sequence frequencies . Implementing a model for the sequence specific bias to Mix2 might therefore result in further improvements in the accuracy of transcript quantification . Experiments were conducted on artificial data to determine the optimal number of mixture components of Mix2 . These experiments showed that optimization can be performed independent of gene , bias and sample size and that the optimal number of mixture components is 3 for Mix2 without tying and any number between 4 and 10 for Mix2 with tying . These numbers are fairly small given the biases in Fig 2 . In particular , it seems implausible that Mix2 with only 3 mixture components should be able to accurately model the 5’+3’ bias in Fig 2 ( d ) . This suggests that the potential of Mix2 has yet to be fully exploited . Our experiments , further , showed that the estimate of Mix2 for the transcript specific fragment bias converges to the correct distribution and that these estimates can therefore be used to detect positional bias present in RNA-Seq data . Experiments were also performed on RNA-Seq data generated from Universal Human Reference ( UHR ) RNA and Human Brain ( HBR ) RNA for the Microarray and Sequencing Quality Control ( MAQC , SEQC ) data sets . On MAQC , we obtained improved correlation between qPCR measurements and quantification estimates with Mix2 , while on the SEQC data set , Mix2 produced improved titration order consistency and recovery of mixing ratios . In addition , correlation and standard deviation of Mix2 quantification estimates were superior across lanes in MAQC and across laboratory sites in SEQC , implying reduced sensitivity to technical variance . Furthermore , correlation of qPCR and FPKM fold changes between UHR and HBR on MAQC were noticeably higher for the Mix2 model than for the other methods . We showed in a classification experiment that this leads to higher accuracy in the detection of differential expression . In general , quantification accuracy affects sensitivity and specificity of statistical tests for differential expression . Inconsistent quantification estimates , for instance , increase technical variability and lead to an increase in the variance of the data models underlying statistical tests . This , in turn , leads to a decrease of the chi-square distributed test statistic of the Wald test in DSS [27] , DESeq2 [23] , edgeR [22] and the score test in PoissonSeq [28] resulting in a loss of sensitivity . The same holds for the Wilcoxon test in SamSeq [32] where the test statistic moves closer to the mean of the test distribution under the 0-hypothesis . Likewise , conditional model probabilities in baySeq [29] and conditional probabilities in the exact tests of DESeq [30] and edgeR [31] become more uniform and therefore less distinctive between different conditions . The noise distributions in NOISeq [33 , 34] and DEGSeq [36] become wider with increasing technical variability and fold change thresholds for the detection of differential expression increase , again reducing the sensitivity of these methods . For DEXUS [35] the variance of the major and minor conditions increases resulting in greater overlap . Since the prior of the model probabilities favors a single condition this makes it less likely that minor conditions and therefore differential expression will be detected . Overall , the aforementioned statistical tests benefit both in terms of sensitivity and selectivity from more accurate quantification estimates and we expect therefore to see improved differential expression calls for transcripts when using these tests in conjunction with Mix2 . In terms of resource usage , both RSEM and Mix2 take about the same time to process the 7 lanes of UHR and HBR in MAQC and are both faster than eXpress and Cufflinks . Memory usage of RSEM is slightly smaller than that of Mix2 and Cufflinks but memory consumption of all 4 methods is low given the specifications of current computing environments . In contrast to the experiments on artificial data , experiments on MAQC and SEQC showed a degradation in the performance of Mix2 when tying parameters . This cannot be attributed to a suboptimal choice of parameters in the transcript clustering procedure . Instead , it seems that positional fragment bias does not exclusively depend on gene membership and transcript length . This fact was also highlighted in our experiments on bias types in MAQC and SEQC . In these experiments we found dominant bias types by clustering fragment start distributions estimated by Mix2 . On MAQC we obtained 6 clusters , containing 73 . 43% of the distributions , of which 5 clusters exhibited non-uniform distributions . The cluster with uniform distributions contained only 26 . 92% . On SEQC we also see the majority of distributions located in non-uniform clusters . In addition , clusters are similar across laboratory sites . Contrary to our experiments on artitifical data there is no obvious relationship between bias and transcript length , although correlations between the two do exist . For instance , transcripts whose fragment start distributions are 3’ biased or uniform tend to be longer , whereas 5’ biased transcripts tend to be shorter . A more detailed analysis of biases might reveal relations between positional bias and RNA sequences that will lead to a better tying strategy for the Mix2 model on real RNA-Seq data . In summary , Mix2 can be used as an explorative tool to investigate the positional biases present in RNA-Seq data and thereby study the influence of library preparation , sequencing and data processing on the accuracy of transcript concentration estimates . In addition , and more importantly , our results show that Mix2 yields improved transcript concentration estimates for RNA-Seq data with higher repeatability for technical replicates and leads , furthermore , to improved accuracy in the detection of differential expression .
We factorize p ( r|t = i ) as follows , p ( r | t = i ) = p ( l ( r ) | t = i , s ( r ) ) ∑ j β i j p ( s ( r ) | t = i , b = j ) ( 4 ) where s ( r ) and l ( r ) are the start and length of fragment r and p ( s ( r ) |t = i , b = j ) are Gaussians whose means μij are placed equidistantly along the transcript . If one disregards the dependency on s ( r ) and t = i it is possible , similarly to Cufflinks , to estimate p ( l ( r ) ) from the data . For paired-end data , however , we always set p ( l ( r ) ) to a Gaussian with mean 200 and standard deviation 80 , which is the default fragment length distribution for Cufflinks . The component weights βij are associated with equidistant positions within the transcript t = i and represent therefore the overall shape of p ( r|t = i ) . In transcript coordinates the means μij of the Gaussians are given by μ i j = j · l ( t = i ) M - l ( t = i ) ( 2 M ) ( 5 ) and their standard deviations are , independent of j , set to σ i j = l ( t = i ) ( 2 M ) . ( 6 ) where l ( t = i ) is the length of transcript t = i and M is the number of mixture components . The Gaussians are , furthermore , normalized such that their sum over the possible fragment starts s = 1 , … , l ( t ) equals one . The relative abundances αi in Mix2 can be updated with the EM algorithm in the usual manner , as implemented , for instance , in Cufflinks [2] . This update formula is given in Section 1 . 1 in S1 Appendix . For the transcripts in group g = k the βij = βkj can be updated with the EM algorithm as follows β k j ( n + 1 ) = ∑ r p ( n ) g = k , b = j | r ∑ r p ( n ) g = k | r ( 7 ) where p ( n ) ( g = k | r ) = ∑ i ∈ k p ( n ) ( t = i | r ) ( 8 ) and p ( n ) ( g = k , b = j | r ) = ∑ i ∈ k p ( n ) ( t = i , b = j | r ) ( 9 ) and the sums in Eqs ( 8 ) and ( 9 ) are extended over the transcripts t = i in group g = k . Here β k j ( n + 1 ) and p ( n ) ( · ) are the mixture components and posterior probabilities after the n+1-th and after the n-th iteration , respectively . To calculate p ( 0 ) ( · ) it is necessary to initialize the model parameters , which we do as follows α i ( 0 ) = 1 N , β k j ( 0 ) = 1 M . ( 10 ) where N is the number of isoforms in the gene locus and M is , again , the number of mixture components . Hence , our initial distributions p ( 0 ) ( r|t = i ) are close to uniform . EM iterations are repeated until changes in the model parameters or the overall likelihood of the model fall below a predefined threshold . We initially place the transcripts of a gene into groups according to their length where the 7 transcript length boundaries are equidistantly distributed on the log scale between 300 and 5000 . Subsequently , groups are merged until each group has at least 20 valid reads and there is at most one group containing a single transcript . Groups are merged according to their distance , which is calculated as follows | m e a n ( g = k 1 ) - m e a n ( g = k 2 ) | 100 - min ( m e a n ( k 1 ) , m e a n ( k 2 ) ) ( 11 ) where mean ( g = k ) is the average length of transcripts in group g = k . The two closest groups are merged first . For each of the 7 genes , 200 sets of abundances ( α1 , … , αN ) were sampled uniformly , according to the Dirichlet distribution . Subsequently , for each of the 200 sets of abundances 500 , 1000 , 5000 and 10000 fragments were sampled from the superposition Eq ( 1 ) , where the p ( r|t = i ) belong to one of the 4 bias models in Fig 2 . These biases are referred to as Cufflinks bias ( a ) , 5’ bias ( b ) , 3’ bias ( c ) and 5’+3’ bias ( d ) . The Cufflinks bias is the fragment start distribution of the Cufflinks model for a fragment length distribution with mean 200 bp and standard deviation 80 bp . The other biases in Fig 2 are derived by scaling an initial 5’ , 3’ or 5’+3’ biased distribution to the length of the transcript . Subsequently , the scaled distribution is multiplied by the Cufflinks bias for the transcript length and renormalized . This explains why the 5’ tails of the biases in Fig 2 become increasingly heavy for shorter transcripts . Section 2 in S1 Appendix contains a brief discussion of how the Cufflinks bias is derived from the Cufflinks model and Fig A to Fig D in S2 Appendix show examples for the coverage resulting from sampling the biases in Fig 2 . The fragment lengths l ( r ) were sampled from a Gaussian with mean 200 bp and standard deviation 80 bp and the resulting fragments were then converted into 50 bp paired-end reads and written to a SAM file [37] . Thus , 800 data sets were generated per gene and sample size or , equivalently , 1400 data sets per bias model and sample size resulting in a total of 22400 data sets . On each of these data sets Mix2 was run without tying as well as with group and global tying , where the number of mixture components ranged from 2 to 20 . Hence a total of 537600 experiments were performed with Mix2 on these artificial data . For the experiments on the complete transcriptome we used only genes with multiple transcripts since the estimated and true relative abundance on genes with a single transcript is always one and their L1 distance is therefore zero . Hence , accumulating the L1 distance of genes with a single transcript artificially decreases the average L1 distance . The RNA samples were sequenced on an Illumina GenomeAnalyzer resulting in 7 lanes per sample of 35 bp single-end reads [9] . The RNA-Seq data of the MAQC data set were downloaded from the NCBI read archive under accession number SRA010153 , while the associated qPCR values were downloaded from the Gene Expression Omnibus ( GEO ) under accession number GSE5350 . The reads of all 14 lanes were aligned to GRCh37/hg19 and Ensembl version 75 with Tophat2 [18] . Rather than the RefSeq annotation , Ensembl version 75 was used in the experiments , since the Ensembl annotation contains in many cases more transcripts per gene than RefSeq and therefore yields a more challenging and also larger test set . Since the MAQC data set records the association between qPCR probes and RefSeq annotations it was necessary to select only those qPCR probes mapping to a single RefSeq annotation , which , in turn , has a unique Ensembl equivalent . This resulted in a test set containing 798 transcripts with on average 8 . 6 transcripts per gene . It should be noted that the implementation of PennSeq in PERL has to be considered a proof-of-concept and as such fails to produce an output for around 10% of the test set . Since the RNA-Seq data from MAQC are single-end the fragment length l ( r ) is unknown and was summed out of Eq ( 4 ) . This sets the first term of the right-hand side of Eq ( 4 ) to 1 . As the fragment start the down-stream end of each read was selected . We downloaded the SEQC data from the NCBI read archive under accession number GSE47792 . On the BGI data we studied the ratio between the concentration of a single transcript for samples C and D and compared this to the expected ratio based on the concentration calculated for samples A and B . The latter is given as follows C D = k 1 A + 1 - k 1 B k 2 A + 1 - k 2 B ( 12 ) where k1 = 3z/ ( 3z + 1 ) and k2 = z/ ( z + 3 ) and , according to [25] , z = 1 . 43 . We evaluate on transcripts for which both sides of Eq ( 12 ) are well-defined for all methods . This gives us a test set of 76514 transcripts .
|
RNA-Seq is a powerful tool for detecting and quantifying genes and gene isoforms . However , accurate quantification in genomic loci with multiple isoforms has proven difficult . This is due to the fact that the transcript generating an RNA-Seq fragment cannot be identified if multiple transcripts share the fragment sequence . Due to this ambiguity , transcript concentration is usually determined in a statistical framework by calculating the probability that a transcript generates an RNA-Seq fragment . Accurate estimation of this probability requires an accurate model of the transcript specific distributions of RNA-Seq fragments . However , fragment distributions in statistical models of RNA-Seq data are usually over-simplified . This article introduces the Mix2 ( rd . “mixquare” ) model which uses mixtures of probability distributions to model the transcript specific positional fragment distributions . Mix2 learns the mixture weights and approximates therefore the fragment bias in RNA-Seq data . We compare Mix2 on artificial and real RNA-Seq data to four state-of-the-art quantification methods . Our experiments show that Mix2 yields more accurate and repeatable quantification estimates and that it leads to more accurate detection of differential expression . We further show that the biases detected by Mix2 contradict the common assumption of a uniform fragment distribution .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"sequencing",
"techniques",
"engineering",
"and",
"technology",
"molecular",
"probe",
"techniques",
"probability",
"distribution",
"mathematics",
"statistics",
"(mathematics)",
"industrial",
"engineering",
"quality",
"control",
"forms",
"of",
"dna",
"bioassays",
"and",
"physiological",
"analysis",
"molecular",
"biology",
"techniques",
"dna",
"rna",
"sequencing",
"research",
"and",
"analysis",
"methods",
"probe",
"hybridization",
"rna",
"hybridization",
"molecular",
"biology",
"probability",
"theory",
"complementary",
"dna",
"microarrays",
"biochemistry",
"statistical",
"models",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"statistical",
"data"
] |
2017
|
Mixture models reveal multiple positional bias types in RNA-Seq data and lead to accurate transcript concentration estimates
|
Oral cholera vaccination is an approach to preventing outbreaks in at-risk settings and controlling cholera in endemic settings . However , vaccine-derived herd immunity may be short-lived due to interactions between human mobility and imperfect or waning vaccine efficacy . As the supply and utilization of oral cholera vaccines grows , critical questions related to herd immunity are emerging , including: who should be targeted; when should revaccination be performed; and why have cholera outbreaks occurred in recently vaccinated populations ? We use mathematical models to simulate routine and mass oral cholera vaccination in populations with varying degrees of migration , transmission intensity , and vaccine coverage . We show that migration and waning vaccine efficacy strongly influence the duration of herd immunity while birth and death rates have relatively minimal impacts . As compared to either periodic mass vaccination or routine vaccination alone , a community could be protected longer by a blended “Mass and Maintain” strategy . We show that vaccination may be best targeted at populations with intermediate degrees of mobility as compared to communities with very high or very low population turnover . Using a case study of an internally displaced person camp in South Sudan which underwent high-coverage mass vaccination in 2014 and 2015 , we show that waning vaccine direct effects and high population turnover rendered the camp over 80% susceptible at the time of the cholera outbreak beginning in October 2016 . Oral cholera vaccines can be powerful tools for quickly protecting a population for a period of time that depends critically on vaccine coverage , vaccine efficacy over time , and the rate of population turnover through human mobility . Due to waning herd immunity , epidemics in vaccinated communities are possible but become less likely through complementary interventions or data-driven revaccination strategies .
Vaccination campaigns with sufficiently high efficacy and coverage can ideally achieve herd immunity in the population . Herd immunity emerges when the indirect protection of vaccination reduces the per-case expected number of onward infections in a population ( i . e . , the effective reproductive number , Re ) below one—whether these onward infections are direct person-to-person or indirect environmental transmission pathways . [1–3] While the balance between these direct and indirect pathways will depend on the setting , the role of both pathways had been described for cholera[4] and utilized in modeling literature . [5–7] Herd immunity is not permanent , however , and is expected to wane over time via short-lived vaccine efficacy and an influx of susceptible , unvaccinated individuals . Due to a reliable efficacy profile and high attainable coverage , killed oral cholera vaccines ( kOCV ) can generate powerful herd protection effects . [8 , 9] In 2012 , the World Health Organization ( WHO ) created a kOCV stockpile to facilitate vaccine usage in three settings: ( 1 ) humanitarian crises at high risk of cholera importation and transmission; ( 2 ) high-endemicity “hot spots”; and ( 3 ) cholera outbreaks . [10] As the stockpile approaches its fifth year , evaluation of its management must address uncertainties in sustainability and long-term strategy , particularly regarding the duration of herd immunity ( DHI ) in these three settings . Regarding the first setting , kOCVs can be a quick stopgap measure to protect cholera-prone dynamic populations such as refugee camps , [11] but it remains unclear how much time is “bought” by vaccination before longer-term solutions such as water , sanitation , and hygiene promotion are necessary . Second , feasibility and economic analyses of vaccination in endemic “hot spot” settings are strongly influenced by the frequency of revaccination . [12] Third , it remains to be seen how strongly , and in what direction , population mobility should be considered when prioritizing target populations for reactive vaccination during outbreaks . These are not merely hypothetical concerns . Beginning in October 2016 , the Bentiu Protection of Civilians ( PoC ) Camp in South Sudan sustained a cholera outbreak despite high-coverage two-dose mass vaccination campaigns in both 2014 and 2015 which were intended to preemptively stave-off the risk of cholera . [13 , 14] Consequently , questions have emerged about the utility of vaccination and the expected risk of outbreaks , [15] particularly in dynamic populations where cholera often breaks out . [16] Modeling studies of other diseases ( e . g . , [17–21] ) suggest a suite of factors which may have contributed to the camp’s susceptibility to an outbreak , including waning vaccine efficacy , the influx of susceptible displaced people , an extremely high birth rate , and resettlement of vaccinated individuals . However , the relative contributions of these factors and their implications for vaccination strategy in the future are not clear . Here we examine the implications of vaccine waning and human mobility on herd immunity over time in non-endemic settings , providing new insights related to the risk of outbreaks in vaccinated populations . Using mathematical models , we compare how well several common vaccination strategies sustain herd immunity and we demonstrate the non-monotonic relationship between migration rate and the projected impact of pre-emptive vaccination . We analyze data from the Bentiu PoC Camp to quantify the impact of expected drivers of waning herd immunity and assess whether they are sufficient to explain the vulnerability of the camp to the observed outbreak .
We developed a system of deterministic differential equations to model a well-mixed population that is being targeted with vaccination . The population compartments of principal interest for this study are individuals who are fully susceptible to disease , S , and those who were vaccinated n-months ago , Vn ( Fig 1 ) . To account for the observation that kOCV direct effects do not tend to wane exponentially , [22 , 23] we created an ensemble of n monthly stages ( V1 , V2 , … , Vn ) , which collectively generate an Erlang-distribution for the duration of time in the V-ensemble . [24 , 25] We set the mean time residing in any Vn compartment to 30 . 5 days; therefore , susceptible individuals move after vaccination to compartment V1 for an average of one month , then to V2 for an average of one month , and so forth . See supporting materials for the system of differential equations ( S1 Text ) . The system of ordinary differential equations was solved using the deSolve package[26] in the statistical software program R ( version 3 . 2 . 4 ) . All code used to generate this paper can be freely found at https://github . com/peakcm/cholera . Vaccination is implemented according to two approaches: mass and routine . We model mass vaccination as a large fraction of individuals moving into the V1 compartment on a particular day , possibly recurrently ( e . g . , annually ) . Routine vaccination moves a substantially smaller fraction of individuals into the V1 compartment for many days in a row . In each approach , vaccine priority is given first to susceptible individuals , S , then those who were vaccinated the longest time ago ( i . e . , Vn , then Vn−1 , and so on until reaching the allotted number of vaccines for that day ) . In addition to mass vaccination and routine vaccination , we test a blended “Mass and Maintain” strategy in which one-time mass vaccination is followed by routine vaccination . See supplemental materials for mathematical details on modeling mass vaccination transition rates ( S1 Text ) . Currently , a complete kOCV course includes two doses administered approximately two weeks apart . [10] However , because the timescale of interest for this study is measured in years , not days , we assume mass vaccination campaigns elapse over one day and provide protection instantaneously . Furthermore , for generalizability across disease systems , we focus on the number of vaccine courses rather than the number of actual vaccines . We parameterized the time-varying vaccine efficacy , VE ( t ) , of kOCV ( whole-cell with B-subunit ) using estimates from a large clinical trial in Bangladesh ( see Discussion for more details on choice of vaccine ) . [22 , 27] Recent meta-analysis results found no differences in two-dose efficacy for vaccines with and without the B-subunit . [23] To provide monthly estimates of vaccine efficacy , VE ( t ) , we used published 6-month point estimates with linear interpolations between each . Efficacy after the 5th year is assumed to be zero , as the estimated mean efficacy becomes negative . [22] Therefore , our last Vn compartment before returning to full susceptibility is at 60 months ( V60 ) . We assume individuals emigrate from the population at a rate that is the same regardless of vaccination status . The total population size , N ( t ) , is held constant by offsetting emigration with an equal rate of immigration , unless otherwise noted . Our main results assume that incoming migrants bring neither vaccine-derived nor naturally-acquired immunity into the population . We estimated migration rates from three example settings where kOCVs have been used , including: ( 1 ) a ‘stable’ urban population; ( 2 ) a highly mobile urban population; and ( 3 ) a displaced person setting with intermediate mobility . First , to represent a stable urban population , we estimate a migration rate of 120years ( i . e . , an average residence time of 20 years ) from the observation that only 9% of an OCV study population in Calcutta , India , changed in the two years following vaccination in 2006 . [28] Secondly , to represent a highly mobile urban population , we estimate a migration rate of 12years from the observation that 58% of a study population in Dhaka , Bangladesh , had relocated over two years . [29] Thirdly , to represent a displacement camp with intermediate mobility , we estimate a resettlement rate of 14 . 3years in the Bentiu PoC Camp in South Sudan in the period from February to October 2016 , during which the International Organization on Migration ( IOM ) reports a rather stable population of 104 , 000 people and approximately 2 , 000 monthly individuals both entering the camp and resettling from the camp ( S1 Fig ) [http://www . iomsouthsudan . org/tracking/] . We define the duration of herd immunity ( DHI ) as the amount of time following a vaccination campaign with an effective reproductive number , Re , below one . We calculate Re ( t ) =X ( t ) *R0 ( 1 ) where X ( t ) is the proportion of the population susceptible at time t , X ( t ) =S ( t ) +∑i=1nVi ( t ) ( 1-VE ( i ) ) N ( t ) . ( 2 ) Our modeling framework serves to estimate the key proportion X ( t ) dynamically . From this value , we derive Re ( t ) , DHI , and the probability of an outbreak as follows . Due to the special behavior of deterministic models , we perform the following adjustments . When a simulation asymptotically approaches Re ( t ) = 1 from below , we define DHI as the time until Re ( t ) ≥ 0 . 99 . Because epidemic extinction is possible in reality when R0 > 1 ( and , conversely , epidemic propagation is possible when R0 < 1 ) , we use our calculation of Re ( t ) to estimate the probability of an outbreak . When Re > 1 , the final epidemic size tends to follow a bimodal distribution with a probability of sporadic die-out and a probability of a large epidemic . Using a recent method for computing epidemic final size distributions , [30] we find the threshold of 10 cases is a reasonable cutoff size such that a large outbreak is henceforth very likely for sizeable values of Re ( S2 Fig ) . We therefore define an outbreak as more than 10 cases and , by assuming a Poisson distribution of secondary infections ( mean = Re ) , we can calculate the probability of an outbreak of more than y cases initiated by a single infectious case using the Borel-Tanner distribution:[31 , 32] Pr ( Y>y ) =1-∑i=1y1 ( i-1 ) ! ii-2Rei-1e-iRe . ( 3 ) To assess the role of mobility on the optimal pre-emptive targeting of kOCVs , we simulate a setting with migration rates ranging from zero , representing a closed population , to a very high value of 11year ( i . e . , an average residence time of one year ) . Since we focus here on an at-risk population in a non-endemic setting , our outcome of interest is the cumulative probability ( C ) of sustaining a cholera outbreak that was seeded by an imported case , which equals one minus the probability of having no outbreaks greater than y cases: C=1-∏t=1D ( ( 1-Pr ( Y>y ) ) Imig ) ( 4 ) where D is the duration of follow-up time in days , y is the minimum outbreak size , and Imig is the expected number of infected individuals who migrate into the population in one day . Imig is calculated by: Imig=πN ( em-1 ) , ( 5 ) where π is the probability an incoming migrant is infected , N is the size of the targeted population , m is the daily migration rate , and therefore the daily number of incoming migrants equals N ( emt − 1 ) where t = 1 day . We assume each imported case has an independent probability of starting an outbreak of more than y cases given the effective reproductive number Re ( t ) on that day t . We measure the difference between the cumulative outbreak probability , C , over D days in the absence of vaccination as compared to the first D days following mass vaccination . A larger difference suggests a more impactful vaccination intervention . For our main results , we focus on a setting with moderate transmissibility ( R0 = 1 . 5 ) [6 , 33] and set the probability that a migrant is infected , π , equal to 1N , which simplifies Eq 5 to Imig = ( emt − 1 ) ( S1 Text ) . We examine the suspected drivers of waning herd immunity in a well-described outbreak in the Bentiu PoC Camp in South Sudan . Of the three million persons targeted for health resources in broader South Sudan , including the Bentiu PoC Camp , UNFPA expects 335 deliveries per day , which equates to birth rate of approximately 124 . 4years . [34] We assumed this to be our demographic turnover rate as a conservatively high estimate . We estimated population susceptibility over time , X ( t ) , in six scenarios ( Table 1 ) . In the “observed” scenario , we used empirical measures of four key drivers of waning herd immunity , specifically: the birth/death rate of 124 . 4years; an empirical distribution of efficacy over time , VE ( t ) ; a camp resettlement rate of 14 . 3years ( i . e . , an average camp residence time of 4 . 3 years ) which is balanced by an equal rate of entries for a net-zero impact on N ( t ) ; and a dynamic population size , N ( t ) , driven by net growth or shrinkage through camp entries or exits . We compare this scenario with counterfactual scenarios that eliminate at least one of these drivers and will therefore increase DHI . We constructed a composite counterfactual scenario in which: the birth/death rate was set to zero; vaccine efficacy was held constant at its maximum value ( 70 . 3% ) for all time since vaccination; the camp resettlement rate was set to zero; and the population size was held constant at approximately the level observed during the outbreak ( 100 , 000 ) . To isolate the impact of each driver of waning herd immunity , we ran simulations where one driver is set to the “observed” condition while the other three drivers are set to their counterfactual condition to remove their influence ( Table 1 ) . To assess the relative importance of each driver of waning herd immunity in this case study , we calculate a measure of attributable percent . For a scenario i that isolates one driver , we measure the proportion susceptible ( X ( t ) i ) on October 16 , 2016 , the start of the observed outbreak . To compare scenarios , we calculate the difference between estimates of the proportion susceptible at the start of the outbreak under scenario i with estimates in the composite counterfactual scenario , ΔX ( t ) i=X ( t ) i-X ( t ) composite . ( 6 ) Finally , we calculate the percent of waning herd immunity attributable to each driver ( AR% ) , AR%=100*ΔX ( t ) i/∑ΔX ( t ) i . ( 7 ) In order to estimate the probability of an outbreak given introduction of a cholera case using the population susceptibility over time , X ( t ) , we must estimate the basic reproductive number , R0 . Following frameworks[35 , 36] recently applied to cholera in South Sudan , [37] we retrospectively estimate the time-varying reproductive number using two sources: ( 1 ) daily case reports , which we extract from Cholera Situation Reports from the South Sudan Ministry of Health , [38] and ( 2 ) an expected generation interval distribution , which we assume to follow a discretized gamma distribution with median of 5 days . [37] This method assumes uniform mixing , no imported cases after the first case , and no missing data . [35 , 36] Maximum likelihood estimation procedures were implemented in the statistical software program R using the R0 package . [39]
Following mass vaccination with 100% coverage , population susceptibility , X ( t ) , quickly increases over time in the presence of high migration rates and short-lived vaccine efficacy ( Fig 2A , solid line ) . Even with a hypothetical perfect vaccine that retains complete protection indefinitely , high migration rates can drive population susceptibility near 100% within 9–10 years ( Fig 2B , solid line ) . Between three primary drivers causing herd immunity to wane , namely migration , waning efficacy , and demographic turnover through births and deaths , we find that the first two are substantially more influential than either the birth or death rate , which are each typically much slower processes . As compared to rates of birth and death set to zero , even pessimistic estimates of a life expectancy of 40 years result in negligible differences in the proportion of the population susceptible due to the relatively faster rates of other drivers ( S3 Fig ) . Following kOCV vaccination with 100% coverage in a population with high migration , we estimate the vaccine-derived DHI to be approximately 0 . 47 years when R0 = 2 , 0 . 98 years when R0 = 1 . 5 , and 4 . 06 years when R0 = 1 ( Fig 2C , solid lines ) . These durations increase to 1 . 07 years , 1 . 89 years , and 5 . 16 years , respectively , in the presence of low migration rates instead ( Fig 2C , dashed lines ) . As expected , DHI is reduced when vaccine coverage is less than 100% , and , depending on both the coverage and R0 , herd immunity is sometimes unattainable ( S4 Fig ) . Achieving herd immunity is a key theoretical threshold , but in reality an outbreak is possible below the threshold and is not guaranteed above the threshold . [40] Mass vaccination reduces , but does not eliminate , the probability that an imported case sparks an outbreak for a duration of time that depends critically on the migration rate and how vaccine efficacy wanes over time ( Fig 2E and 2F ) . For example , even though herd immunity is lost within just 0 . 47 years in a high migration setting when R0 = 2 ( Fig 2C , solid red line ) , the outbreak probability is kept below 50% for twice as long ( Fig 2E , solid red line ) . We considered several operational strategies for sustaining herd immunity through vaccination alone . In a hypothetical population of size N with R0 = 1 . 5 and a high rate of migration ( 12years ) , mass vaccination every year or every two years with 100% coverage of susceptible individuals can render herd immunity for 3 . 5 or 2 . 8 years , respectively , before depleting a fixed vaccine allotment of 3N full vaccine courses ( Fig 3A ) . If these vaccines are instead allotted on a daily basis through routine vaccination , DHI can be extended to 4 . 4 years ( Fig 3B ) . However , recurring mass campaigns have diminishing returns per vaccine once herd immunity is achieved; meanwhile routine vaccination alone requires a long period of time to build up herd immunity . We therefore find that a blended “Mass and Maintain” strategy that complements a single mass vaccination campaign with subsequent routine vaccination can maintain herd immunity longer than either strategy alone ( Fig 3C ) , both for this example and for a wide range of settings with various migration rates and R0 values ( S1 Table ) . In addition to the importance of migration on DHI , one may posit that communities with higher migration rates are also more likely to have cholera imported . In order to optimize pre-emptive kOCV impact in at-risk settings , there is a tradeoff between targeting low-mobility communities , where herd immunity may last for a long time but cholera introduction is rare , and high-mobility communities , where the opposite is expected . We find that communities with intermediate levels of migration may experience the largest vaccine-derived decrease in outbreak risk sparked by an imported case ( Fig 4 ) . For example , the migration rate recorded in the Bentiu PoC Camp in mid-2016 is near the optimal condition for maximizing the impact of a single mass vaccination campaign in the 4–6 year time horizon , assuming R0 = 1 . 5 . If one is more interested in shorter time horizons since vaccination , the migration rate that maximizes vaccine impact favors mobile communities , similar to some urban areas in Dhaka , Bangladesh . [29] Sensitivity analyses suggest that intermediate mobility rates ( e . g . , between those observed in Dhaka and Calcutta ) generally maximize vaccine impact , but the optimal migration rate is slower in settings that have a larger population size , a higher transmission potential ( R0 ) , or where a higher fraction of incoming migrants are infected ( e . g . , due to high-burden neighbors ) ( S5 Fig ) . Conversely , settings with small population size , low transmission potential , and whose migrants have a small probability of being infectious require very high migration rates in order to garner much baseline risk of cholera importation and outbreak . The Bentiu PoC Camp grew from 4 , 291 occupants in February 2014 to a peak of 140 , 101 in December 2015 and then converged to approximately 104 , 000 in May 2016 ( Fig 5A ) . Assuming a cholera-naïve population before vaccination , we estimate that only 37% of the camp remained susceptible after the second round of vaccination in June 2015 . By the time that the first cholera case in the camp was detected ( i . e . , October 16 , 2016 ) , the camp susceptibility percentage increased to 81% ( Fig 5B ) . By December 1 , 2016 , we estimate that only 40 . 5% of camp residents had ever been vaccinated , which closely matches a WHO/IOM survey performed that month that reported kOCV coverage of 40% . [38] Using case reports and assuming a fixed generation interval distribution , we estimate the mean effective reproductive number , Re ( t ) , exceeded unity for approximately two months following the first case , with a maximum likelihood estimate of 1 . 45 ( 1 . 18–1 . 75 ) ( S6 Fig ) . Using the population fraction susceptible of 0 . 81 estimated above , we calculate a basic reproductive number , R0 , of approximately 1 . 80 in this setting in the absence of vaccination ( Eq 1 ) . These findings are within the range of estimates derived from South Sudan in 2014 . [37] Assuming this pre-vaccination estimate of R0 = 1 . 80 , we find that after vaccination the probability of an outbreak first exceeded 50% in May 2016 and reached 57% when the outbreak began in October ( Fig 5C , black line ) . Using a “Mass and Maintain” strategy including vaccination of 100% of individuals migrating into the camp after the second mass vaccination campaign , we estimate that only 52% of the population would have been susceptible on the date the first case was reported in the camp , which is low enough to generate herd immunity at the time ( assuming R0 = 1 . 80 ) ( S7 Fig ) . The drivers of waning herd immunity in this population , from strongest to weakest , were short-lived vaccine efficacy , population growth , camp resettlement rate , and demographic turnover via births and deaths ( Table 1 , S2 Table ) . In the counterfactual scenario lacking these drivers , we would expect that as few as 34% of the population were susceptible on the day of the first reported cases in Bentiu PoC , which would render herd immunity even if R0 was as high as 3 .
Vaccination can rapidly protect a population at risk of a cholera outbreak , but the duration of vaccine-derived herd immunity depends critically on vaccine coverage , waning vaccine efficacy , and a net influx of susceptible people through population mobility . In our case study of the Bentiu PoC Camp , we find that quantification of these drivers help explain the vulnerability of this population to an outbreak despite two recent high-coverage vaccine campaigns . Therefore , disease re-emergence does not necessarily imply vaccine failure and can be avoided by data-driven revaccination strategies or by scaling-up long-term , broad-spectrum solutions while under the temporary cover of pathogen-specific vaccination . Our results provide key time windows during a population can expect to resist a cholera outbreak even if the pathogen were to be introduced . One practical implementation of the “Mass and Maintain” vaccination strategy in a camp setting can include a one-time mass vaccination campaign followed by routine vaccination of new members of the population , such as births and new entries . Population sub-groups with high vulnerability and mobility , such as coastal fishing communities , [41] may also benefit from the “Mass and Maintain” vaccination strategy targeted at seasonal influxes of migrants such as new fishermen . In an urban or open population , such as Dhaka or Calcutta , routine identification of new members becomes more challenging , but performance of the WHO Expanded Programme on Immunization in cholera endemic regions like Bangladesh are promising . [42] Recent work has also shown serological triggers for periodic mass vaccination can be an effective alternative method to maintain herd immunity to measles . [43] . For cholera specifically , there is a need for more research into cross-sectional markers of immunity which can inform risk profiling , revaccination timing , and , if stratified by age , the impact of mass vaccination . [44] A growing area of research focuses on the vaccine efficacy , and duration of protection , provided by a single kOCV dose . [5 , 45] Such work will help elucidate the relative merits of revaccination versus ongoing vaccination of new arrivals , for example . Current guidelines for the optimal use of the kOCV stockpile recommend targeting “areas with important population movements . ”[46] Mobility is recognized as an important driver of the performance of vaccination strategies to control ongoing cholera outbreaks . [47] Here , we focus on pre-emptive vaccination of at-risk communities to show the competing effects of high mobility on expected vaccine impact . In order to operationalize the finding that vaccination may be most impactful for populations with intermediate degrees of mobility , data on migration rates from sources such as censuses or mobile phone call data records must be collected to define “intermediate” mobility for a given context . [48] Long-term solutions to cholera , and many other waterborne diseases , include investments in water , sanitation , and hygiene infrastructure . In high-income countries where such systems are established , the reproductive number is expected to be far below unity and therefore the effective integration of new migrants , even from poorer or cholera-prone regions , is not expected to decrease herd protection of the host population . However , in areas such as informal settlements and large low-income cities , incoming migrants can put stress on an already fragile water and sanitation infrastructure , potentially pushing the reproductive number above one and rendering the population at risk of sustained cholera transmission . Our results depend on several simplifying assumptions . By modeling a well-mixed population , we are assuming no heterogeneity in contact patterns or local reproductive numbers . In reality , we expect diseases , especially those like cholera with environmental transmission dynamics , to exhibit substantial spatial heterogeneity in transmission intensity . [49] These differences become crucial if , as we may expect , migration occurs at higher rates into sub-regions with higher transmission potential due to confounders like poverty and temporary housing . Critically , we would expect DHI to decrease , the probability of an outbreak to increase , and the routine vaccination of migrants to become even more crucial . Our model assumes a leaky mode of vaccine action , whereby vaccination reduces the disease susceptibility of each recipient . Our calculation of proportion susceptible , X ( t ) , is robust to other assumptions regarding the method by which vaccine effects wane , namely: time-dependent failure in “take , ” corresponding to an all-or-nothing response; and time-dependent failure in “degree , ” corresponding to a leaky vaccine response ( S8 Fig ) . [50] Our parameterization of a waning leaky vaccine aligns with prevailing interpretations[22] of the clinical trial data , [27] but alternative explanations for changes in vaccine efficacy over time in a clinical trial are difficult to rule-out , such as frailty , loss to follow up , and random variability . [51] The migration rates estimated from Dhaka , Bentiu , and Calcutta are intended for benchmarking purposes and do not imply that migration rates are either constant or generalizable to the whole city or region . Indeed , we would expect to retain herd immunity longer after vaccination for a given migration rate if the rate was calculated in a population which included a stable sub-group of permanent residents and a small , highly mobile sub-group of temporary residents . Cholera vaccine efficacy has been shown to vary by age of recipient , [22 , 23 , 27] however for simplicity and lack of detailed data we do not model this age structure . If children respond poorly to kOCV and are members of a mass vaccination campaign , we would expect herd immunity to wane more quickly , and especially so if children are disproportionate sources of transmission . Furthermore , over the course of an outbreak , we may expect the relative contributions of different age groups to differ , which can have important consequences on vaccine impact and targeting . [52] Currently , little is understood about immunity to cholera , though the waning individual immunity could be derived from the central role of mucosal phenomena . [53] For simplicity , we focus on pre-emptive vaccination of a generalized population without previous exposure to cholera . Although kOCV with the B-subunit is less preferred for vaccine stockpile applications , our primary results present the kOCV efficacy profile with the B-subunit due to the biological plausibility of the estimates in the time-varying analysis[22] and the recent observation that two-dose vaccine efficacy with and without the B-subunit is likely indistinguishable . [23] Though in reality many other forces likely contributed to the Bentiu PoC Camp’s susceptibility to the observed cholera outbreak , our case study shows that known key drivers ( namely waning vaccine efficacy , a net influx of susceptible people through population mobility ) alone are strong enough to produce the observation that the camp population sustained a cholera outbreak despite recent vaccination campaigns . Provided additional data on the cholera outbreak and the camp population , a complete cholera model fit to the epidemic may yield additional insights . The model we present is not limited to cholera or other diseases with only short-duration or leaky vaccines ( e . g . , the typhoid capsular polysaccharide vaccine [54] ) . The phenomenon of waning herd immunity also has strong implications on disease control strategies that include mass vaccination or “mop up” vaccination , such as measles[55] and yellow fever . [56] For yellow fever in particular , fractional vaccine doses have been used to extend vaccine supply under the assumption that vaccine efficacy of fractional doses lasts at least one year . [57] Following the mass vaccination of 25 million people in Angola and the Democratic Republic of the Congo , routine vaccination may be the most efficient way to henceforth sustain herd immunity in these populations , should this be the goal . Human mobility and waning herd immunity are key considerations for when these urban populations should be revaccinated . Herd immunity is a key target for the control of vaccine-preventable diseases and can be monitored over time using information on the vaccine efficacy and population turnover rates . We show this information is essential for optimizing revaccination strategies , targeting vaccine stockpiles , and explaining re-emergence of outbreaks in recently vaccinated populations .
|
Cholera vaccination can be a relatively quick means to temporarily prevent cholera from spreading in an at-risk population . In order to understand how long this temporary protection remains and therefore the timeline for when we need to install longer-term water and sanitation solutions , we must know how long we can expect the vaccine to provide herd protection . To answer this and other related questions , we developed a mathematical model to test different vaccination strategies in a simulated population and in a case study of a displaced-persons camp in Bentiu , South Sudan . We found that the duration of vaccine-derived herd protection can be short ( <1 year ) in settings of moderate transmission potential and high population mobility , but this duration can be extended through a strategy that complements a one-time mass vaccination campaign with ongoing , routine vaccination . We show that short-lived vaccine efficacy and high population turnover in the Bentiu camp can help explain why the camp had a cholera outbreak despite two high-coverage vaccination campaigns in the two previous years . Our results support , and provide timelines for , cholera vaccination as initial protection while longer-term structural interventions can be implemented .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"south",
"sudan",
"vaccines",
"preventive",
"medicine",
"bacterial",
"diseases",
"animal",
"behavior",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"vaccination",
"and",
"immunization",
"zoology",
"africa",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"cholera",
"cholera",
"vaccines",
"behavior",
"epidemiology",
"people",
"and",
"places",
"animal",
"migration",
"immunity",
"biology",
"and",
"life",
"sciences"
] |
2018
|
Prolonging herd immunity to cholera via vaccination: Accounting for human mobility and waning vaccine effects
|
Translesion DNA synthesis ( TLS ) is a DNA damage tolerance mechanism in which specialized low-fidelity DNA polymerases bypass replication-blocking lesions , and it is usually associated with mutagenesis . In Saccharomyces cerevisiae a key event in TLS is the monoubiquitination of PCNA , which enables recruitment of the specialized polymerases to the damaged site through their ubiquitin-binding domain . In mammals , however , there is a debate on the requirement for ubiquitinated PCNA ( PCNA-Ub ) in TLS . We show that UV-induced Rpa foci , indicative of single-stranded DNA ( ssDNA ) regions caused by UV , accumulate faster and disappear more slowly in PcnaK164R/K164R cells , which are resistant to PCNA ubiquitination , compared to Pcna+/+ cells , consistent with a TLS defect . Direct analysis of TLS in these cells , using gapped plasmids with site-specific lesions , showed that TLS is strongly reduced across UV lesions and the cisplatin-induced intrastrand GG crosslink . A similar effect was obtained in cells lacking Rad18 , the E3 ubiquitin ligase which monoubiquitinates PCNA . Consistently , cells lacking Usp1 , the enzyme that de-ubiquitinates PCNA exhibited increased TLS across a UV lesion and the cisplatin adduct . In contrast , cells lacking the Rad5-homologs Shprh and Hltf , which polyubiquitinate PCNA , exhibited normal TLS . Knocking down the expression of the TLS genes Rev3L , PolH , or Rev1 in PcnaK164R/K164R mouse embryo fibroblasts caused each an increased sensitivity to UV radiation , indicating the existence of TLS pathways that are independent of PCNA-Ub . Taken together these results indicate that PCNA-Ub is required for maximal TLS . However , TLS polymerases can be recruited to damaged DNA also in the absence of PCNA-Ub , and perform TLS , albeit at a significantly lower efficiency and altered mutagenic specificity .
Translesion DNA synthesis is a universal DNA damage tolerance mechanism , which enables continuous functioning of replication despite the presence of DNA lesions . While the replisome might be able to bypass lesions that cause minor changes in the structure of DNA , lesions which are bulky or cause significant DNA deformation , block replication . Such lesions are bypassed by specialized low-fidelity DNA polymerases , which are capable of replicating across DNA damage due to a flexible structure and promiscuous active site that allows lesion bypass at the cost of increased mutagenesis . At least 5 specialized DNA polymerases are involved in TLS in mammalian cells , namely DNA polymerases η , κ , ι , ζ and REV1 , however , the number may be as high as 10 . Each polymerase exhibits a certain range of specificity towards various types of DNA lesions , with some overlap [1]–[4] . TLS typically operates in two-polymerase reactions , in which the first polymerase inserts a nucleotide opposite the lesion , and the second polymerase , usually DNA polymerase ζ ( polζ ) , extends beyond the lesion [5]–[7] . The biological importance of TLS is indicated by the essentiality of polζ for mouse development [8] , and the high cancer predisposition caused by germ-line mutations in the POLH gene ( encoding DNA polymerase η; polη ) in humans [9] , [10] . TLS must be tightly regulated to prevent an escalation in mutation rates . Although TLS regulation is not fully understood , it does appear to be regulated primarily at the posttranslational level , and involves the ubiquitination of PCNA , the sliding DNA clamp that tethers DNA polymerases to the DNA [11]–[14] . In addition , TLS is regulated by the p53 tumor suppressor , which exerts its effect primarily via its target p21 protein [15] . The latter is a cell cycle inhibitor , which exerts its regulatory effect on TLS via its interactions with PCNA [15] , and cyclin-dependent kinases [16] . Together p53 and p21 restrain the extent of TLS , but make it more accurate , thereby reducing the mutagenic load of TLS [15] . A key regulatory element in TLS is the monoubiquitination of PCNA at lysine 164 in response to treatment with DNA damaging agents . In S . cerevisiae this reaction is carried out by the Rad6-Rad18 E2-E3 ubiquitinating enzymes ( Figure 1A; [17] , [18] ) , and is critical for the activity of TLS , functioning to recruit TLS polymerases through their ubiquitin-binding domain , and thereby switching from replicative to TLS polymerases [11] , [12] . In higher organisms the involvement of ubiquitinated PCNA ( PCNA-Ub ) in TLS is less clear . Analysis of replication of damaged DNA in chicken DT40 cells suggested that PCNA ubiquitination is involved in filling in of post-replication gaps [19] . However when measured using a plasmid system in DT40 cells carrying the PcnaK164R/K164R mutation which prevent ubiquitination of PCNA , TLS was normal across a TT 6-4 photoproduct ( TT 6-4 PP ) , a common UV DNA lesion [20] . As for mammals , a common model suggests that similar to S . cerevisiae , PCNA-Ub recruits TLS polymerases to the site of DNA damage mediated via their ubiquitin-binding domain [13] , [14] , [21]–[23] . This model was challenged by studies reporting that mutations in the ubiquitin-binding domain of polη had no effect on its activities , and it is the direct binding of polη to PCNA which is critically important for its activities [24] , [25] . In response it was argued that these results can be explained by an effect of artificial overproduction of the mutant polymerase , which suppressed its lower binding affinity [26] . However , later studies reported that complete deletion of the UBZ ubiquitin-binding domain from polη had no effect on its activities , including TLS across a site specific TT cyclobutane pyrimidine dimer ( CPD ) in a replicative plasmid assay system [27] , leaving the controversy unsettled . In an attempt to resolve the controversy and clarify the role of PCNA-Ub in TLS in mammalian cells we used several assays with mouse embryonic fibroblasts in which specific TLS genes associated with PCNA ubiquitination were manipulated . The experiments reported here show that the main TLS pathway requires PCNA-Ub . However , there exists a secondary but significant TLS pathway , which occurs in the absence of PCNA-Ub , with lower efficiency and altered mutagenic specificity .
A Pcna mutant in which Lys164 was replaced by an Arg cannot undergo ubiquitination or sumoylation , and was successfully used to study the role of PCNA ubiquitination in TLS in yeast [17] , [18] and chicken DT40 cells [19] , [28] . The generation of PcnaK164R/K164R mice [29] provides a similarly effective tool for studying the role of PCNA ubiquitination in TLS in mammalian cells . We first examined the UV sensitivity of PcnaK164R/K164R mouse embryo fibroblasts ( MEF ) . As can be seen in Figure 1B , the mutant cells were more sensitive than the Pcna+/+ MEF . This suggests that ubiquitination of PCNA at Lys164 is involved in conferring UV resistance in mammalian cells , although it is possible that the effect was caused not only by the lack of ubiquitin , but also by the mutant form of the PCNA . UV irradiation causes stalling of replication forks and the generation of ssDNA regions in DNA , which may subsequently be broken , thereby forming double strand breaks ( DSB ) . The latter can facilitate a variety of chromosomal rearrangements , causing genomic instability , cancer and cell death . To minimize the formation of DSB , cells employ two major types of DNA damage tolerance mechanisms , namely TLS and HDR ( homology-dependent repair , also termed HRR , homologous recombination repair; reviewed in [4] , [30] ) . Of the two , TLS was reported to be the major mechanism for overcoming UV lesions in MEF [31] . Thus , analysis of UV-induced ssDNA regions during replication can be used as a measure for DNA damage tolerance in general , and TLS in particular . To examine the effect of PCNA ubiquitination on DNA damage tolerance we analyzed the formation and clearance of UV-induced ssDNA regions in PcnaK164R/K164R MEFs compared to Pcna+/+ MEFs . This was done using immunofluorescence staining of endogenous Rpa2 , a subunit of the Rpa ssDNA-binding protein , which is a key protein in DNA replication and repair [32] , using a protocol previously used in our lab [33] . To focus on gaps formed during replication we isolated by centrifugal elutriation MEFs at the G1/S boundary , UV irradiated them , and let them grow in culture . At various time points after irradiation the cells were harvested , washed to remove unbound Rpa , and then fixed and stained for chromatin-bound Rpa using immunofluorescence ( Figure 2A ) . In unirradiated cells a low background level of Rpa foci was observed ( less than 10%; Figure 2B ) . Presumably the transient nature and short patch of Rpa-bound ssDNA at normal replication forks does not allow detection under our assay conditions . Upon UV irradiation Rpa foci were induced in the two cell types , but at different rates . Thus , by 6 hours post-irradiation nearly 40% of the PcnaK164R/K164R cells contained Rpa foci , whereas Pcna+/+ cells contained only the background level of 10% Rpa foci ( Figure 2B ) . The observed fraction of cells with Rpa foci at any given time represents the sum of the rates of their formation and disappearance . Therefore , the higher percentage of Rpa foci at early times in PcnaK164R/K164R cells represents , most likely the sum of a similar rate of formation but a slower rate of disappearance compared to Pcna+/+ cells . The extent of cells with Rpa foci increased for both cell types , reaching its maximum at 18 hours post-irradiation , after which the number of foci declined , indicating a net conversion of the ssDNA regions to double stranded DNA ( dsDNA ) . The disappearance of Rpa foci was significantly slower in the PcnaK164R/K164R compared to Pcna+/+ MEFs ( Figure 2B ) . Thus , Rpa foci accumulate faster in PcnaK164R/K164R compared to Pcna+/+ MEFs following UV irradiation , and disappear slower , consistent with a defect in DNA damage tolerance by TLS . To directly examine the effect of PCNA-Ub on TLS we used a model assay system based on plasmids carrying a gap opposite a defined site-specific DNA lesion . Briefly , cultured cells were transfected with a mixture containing a gapped plasmid with a site-specific lesion in the ssDNA region , a normalizing control plasmid with a gap , but no lesion , and a carrier plasmid ( Figure S1 ) . After allowing time for TLS in the mammalian cells , the plasmid content was extracted under alkaline conditions , and after renaturation it was used to transform an indicator E . coli recA ( TLS-defective ) strain . Under these conditions only plasmids that had been fully filled in and ligated in the mammalian cells were able to efficiently transform the bacterial strain . E . coli transformants were selected on LB plates containing kanamycin , to select for descendents of the gap-lesion plasmid ( kanR ) , and LB containing chloramphenicol , to select for descendents of the normalizing gapped plasmid ( cmR ) . The ratio of kanR/cmR colonies provided a measure of the efficiency of gap filling by TLS . Colonies were then picked , their plasmid content extracted , and subjected to DNA sequence analysis at the region of the lesion , to determine any sequence changes that had arisen during TLS . This model assay system proved to be very effective in monitoring TLS events , and shares many of the features of chromosomal TLS , including dependence on specific DNA polymerases and regulatory elements of TLS [6] , [15] , [34]–[36] . Using a gapped plasmid carrying a site-specific TT CPD in the ssDNA region we assayed TLS in PcnaK164R/K164R MEFs . As can be seen in Figure 3A and Table S1 , TLS was reduced 4 . 4-fold in the mutant PcnaK164R/K164R cells compared to Pcna-proficient MEFs . DNA sequence analysis revealed that 98% of the TLS events in both the mutant and wild-type MEFs were accurate , leading to the insertion of AA opposite the TT CPD , and consistent with the activity of polη ( Figure 3B and Table S2 ) . We used the same assay for two additional lesions: a TT 6-4 PP , representing the second most abundant UV lesion , and an intra-strand GG adduct formed by the drug cisplatin ( GG-cisPt ) . TLS across cisPt-GG occurs primarily via a two-polymerase reaction with polη performing insertion opposite the lesion , and polζ performing the extension past the lesion , whereas efficient bypass of TT 6-4 PP does not require polη , but does requires polζ [6] . As can be seen in Figure 3A and Table S1 , TLS across a TT 6-4 PP was reduced 3 . 3-fold in PCNA-Ub deficient MEFs compared to wild-type Pcna+/+ MEFs . DNA sequence analysis revealed that TLS across the TT 6-4 PP in PcnaK164R/K164R cells was more accurate than in Pcna+/+ cells ( 64% vs . 35% errors , P = 0 . 0001 , χ2 test; Figure 3C , Table S2 ) . Analysis of TLS across cisPt-GG revealed that TLS was reduced 2 . 6-fold in PCNA-Ub deficient MEF compared to wild-type Pcna+/+ MEFs ( Figure 3A and Table S1 ) . TLS was largely accurate in both cell types , however , error frequency was twofold higher in the PcnaK164R/K164R mutant compared to Pcna+/+ MEFs ( 25% vs 12% errors; P = 0 . 02 , χ2 test; Figure 3D , Table S2 ) . The PcnaK164R/K164R mutation renders PCNA resistant not only to monoubiquitination , but also to polyubiquitination ( Figure 1A ) and sumoylation ( although PCNA sumoylation was not yet found in mammals ) . To further analyze the involvement of PCNA modification in TLS we analyzed two additional mutant MEFs: A Rad18 knockout strain , which lacks the Rad18 E3 ubiquitin ligase that monoubiquitinates PCNA at K164 [37] , and an Shprh−/− Hltf−/− double knockout MEF [38] , lacking the two Rad5-homologs , which polyubiquitinate PCNA at K164 . As can be seen in Figure 4A and Table S3 , TLS in Rad18−/− MEFs was significantly reduced compared to Rad18+/+ MEFs for each of the three lesions . DNA sequence analysis revealed that mutagenicity of TLS in MEF lacking Rad18 was similar or lower compared to MEF with Rad18 ( Figure 4B–4D and Table S4 ) . In contrast , in cells lacking Shprh and Hltf , the extent of TLS across each of the three lesions was normal ( Figure 5A and Table S5 ) , and the mutagenic spectra were similar ( Figure 5B–5D; Table S6 ) . These results suggest that maximal TLS indeed requires ubiquitination of PCNA . The Usp1 deubiquitinating enzyme was shown to deubiquitinate monoubiquitinated PCNA ( PCNA-mUb; Figure 1 ) [39] . To examine the effect of Usp1 on TLS we assayed TLS in Usp1−/− MEFs . As can be seen in Figure 6A and Table S7 , TLS across a TT CPD was 2 . 3-fold higher in Usp1−/− MEFs compared to wild-type MEF . Similarly , TLS across a cisPt-GG adduct was 3 . 8-fold higher in Usp1−/− MEFs compared to wild-type MEFs . Interestingly , there was no effect on TLS across the TT 6-4 PP . Complementing the Usp1−/− MEFs with stably expressed wild-type Usp1 suppressed TLS across cisPt-GG back to wild-type levels , whereas expressing a Usp1 C90S mutant [40] failed to suppress TLS , indicating that the observed effects are indeed due to Usp1 ( Figure 6A and Table S7 ) . DNA sequence analysis revealed somewhat different mutagenicity , however with no statistical significance ( Figure 6B–6F; Table S8 ) . Thus , the absence or inactivation of the enzyme that deubiquitinates PCNA-mUb caused an increase in TLS in 2 out of the 3 lesions studied , in contrast to the decrease in TLS caused by the inability to ubiquitinate PCNA . These results are consistent with previous reports that mutations in a UV-irradiated plasmid transfected into mammalian cells were increased when Usp1 was reduced or absent [39] , [41] . The data presented above indicate that although PCNA-Ub is required for maximal TLS in mammalian cells , a significant level of TLS was observed in the absence of PCNA-Ub , suggesting the existence of a PCNA-Ub-independent pathway . We further probed this possibility by assaying UV sensitivity of PcnaK164R/K164R MEFs in which the expression of specific TLS proteins was knocked-down , using as an assay the ability to form colonies following UV irradiation ( Figure 7 ) . As can be seen in Figure 7A–7C , PcnaK164R/K164R MEFs were more UV sensitive than wild-type MEFs when treated with the control siRNA , consistent with the role of PCNA-Ub in TLS across UV lesions , and with the results presented in Figure 1B . Knocking down the expression of Rev3L , encoding the catalytic subunit of polζ , in wild-type MEF caused an increased UV sensitivity ( Figure 7A ) , consistent with previous results [36] , and reaching a sensitivity level similar to PcnaK164R/K164R MEFs treated with a control siRNA . When PcnaK164R/K164R MEFs were treated with a Rev3L-specific siRNA , UV sensitivity further increased ( Figure 7A ) , suggesting the existence of a PCNA-Ub-independent polζ-dependent pathway of TLS . Similar results were obtained when the expression of polη ( Figure 7B ) , or of Rev1 , an important regulatory protein and a dCMP transferase ( Figure 7C ) , were each knocked down in PcnaK164R/K164R MEFs . Taken together these results suggest the existence of PCNA-Ub-independent pathways of TLS , which are Polη , Rev1 and/or Polζ dependent .
The debate about the role of PCNA-Ub in polη-promoted TLS in mammalian cells prompted us to address this issue using several mutant mouse cell lines , and several assays . The latter included ( 1 ) a TLS assay based on gapped plasmids carrying defined and site-specific lesions; ( 2 ) immuno-staining of Rpa foci following UV irradiation of cells at the G1/S boundary of the cell cycle , which assays ssDNA gaps caused by UV lesions; and ( 3 ) UV sensitivity as manifested by the ability of irradiated cells to form colonies . Overall we studied three types of lesions , two of which are formed by UV radiation and one by the drug cisplatin , representing three different TLS sub-pathways [6] , [7] . The effects of the knockout mutations in each of the mutants analyzed , PcnaK164R/K164R , Rad18−/− , Shprh−/−Hltf−/− , and Usp1−/− , can be attributed to more than one pathway . Thus , the PcnaK164R/K164R mutant is deficient not only in monoubiquitination , but also in polyubiquitination and sumoylation [42]; The Rad18−/− , which is deficient in monoubiquitination of PCNA , is known to be deficient in the ubiquitination of other proteins as well ( e . g . , [43] ) , and similarly the other mutants may affect several activities . However , the similar effects on TLS of the PcnaK164R/K164R and Rad18−/− cells , suggest that ubiquitination rather than sumoylation is involved . Noteworthy , no alternative PCNA ubiquitination site was observed in mouse PCNA [29] . What about the discrimination between monoubiquitination and polyubiquitination of PCNA ? The Hltf and Shprh proteins are E3 ligases which polyubiquitinate PCNA . They were also reported to be involved in the regulation of monoubiquitination of TLS under very high UV doses [44] . The normal TLS observed in the Shprh−/−Hltf−/− cells suggests that PCNA-mUb rather than PCNA-polyUb is involved . However , it was recently reported that PCNA polyubiquitination is reduced , but not completely eliminated in Shprh−/−Hltf−/− MEFs , suggesting that an additional E3 ligase acts on PCNA [38] . Thus , an involvement of PCNA-polyUb in TLS cannot be ruled out based on these experiments alone . However , given the normal TLS in Shprh−/−Hltf−/− MEFs , and the biochemical data on the binding of TLS polymerases to PCNA-mUb [13] , [22] , [23] , [45] , it does seem that the dependence on ubiquitination is primarily due to the activity of PCNA-mUb . The chromosomal significance of these finding is indicated by the faster accumulation of Rpa foci in PcnaK164R/K164R cells UV irradiated at the G1/S boundary of the cell cycle , and their slower clearance compared to Pcna+/+ cells . Rpa strongly binds sites of ssDNA , providing a convenient tool for assaying ssDNA gaps caused by UV lesions during replication . Recent data suggest that at least in MEFs , TLS is the major pathway for repair of replication gaps caused by UV lesions [31] . Moreover , we have recently found that the disappearance of UV-induced Rpa foci is strongly reduced in cells in which the expression of polζ was knocked down , indicating involvement of TLS [33] . Thus , the inhibition of the clearance of post-UV Rpa foci in PcnaK164R/K164R cells is consistent with the decreased TLS across the TT CPD and TT 6-4 PP lesions observed in the gapped plasmid system , providing further support to the importance of PCNA-Ub for maximal TLS . The debate on the role of PCNA-Ub in mammalian TLS involved primarily the activity of polη in bypassing UV lesions , where a series of papers presented conflicting results [13] , [22]–[24] , [27] . Those studies were based on mutating , or even entirely deleting , the polη ubiquitin-binding domain . Our study directly addressed ubiquitinated PCNA , and using functional TLS assays showed that TLS across TT CPD was impaired in the absence of PCNA ubiquitination , indicating that PCNA-Ub is required for the maximal bypass activity of polη . Two studies reported that PCNA-mUb was not required for polη-promoted TLS across TT CPD in a cell-free TLS assay [46] , [47] ( but was required to bypass an N-2-acetylaminofluorene-guanine adduct; [47] ) . These cell-free systems may not have faithfully mimicked the in vivo requirements for polη-promoted TLS across TT CPD due to the inherent ability of purified polη to bypass a TT CPD unassisted by any other protein [9] , [10] . Adding our new data to the previously published data , we conclude that in a sense both sides of the controversy on the role of PCNA-Ub in TLS were right: On one hand we clearly show that TLS across three different DNA lesions , TT CPD , TT 6-4 PP and cisPt-GG , requires PCNA-Ub for maximal activity , but on the other hand TLS across each of the three lesions occurs also in the absence of PCNA-Ub , albeit at reduced extent and altered mutagenic specificity . A key issue in TLS is the mechanism that ensures the recruitment of TLS polymerases to their cognate lesions , such that the entire TLS system operates without causing excessive mutations . This mechanism is regulated by the tumor suppressor p53 , exerting its effect , at least in part , via the PCNA-binding function of the p21 protein , whose expression it regulates [2] , [15] . Is PCNA-Ub an important regulator of this process of TLS fidelity control ? TLS in this study was analyzed with lesions that span a broad range of bypass fidelity; From highly accurate TLS ( TT CPD , <1% errors ) , via mostly accurate TLS ( cisPt-GG , about 10% errors ) , up to mostly mutagenic TLS ( TT 6-4 PP , about 65% errors ) . Some variations in mutagenic spectra among the wild-type MEFs were likely caused by differences in the genetic background of the MEFs , and by changes that might have occurred during immortalization . The absence of PCNA-Ub changed the fidelity of TLS across both the ‘accurate’ cisPt-GG lesion , as well as the mutagenic TT 6-4 PP lesion , but in different directions . Thus , while TLS across cisPt-GG became more mutagenic in the absence of PCNA-Ub ( Figure 3D and Table S2 ) , it became more accurate for the bypass of TT 6-4 PP ( Figure 3C and Table S2 ) . It is easy to envisage that conditions that decrease the efficiency of TLS will also cause lower fidelity , like in the case of cisPt-GG , because the TLS machine operates under sub-optimal conditions . However , the observation that the absence of PCNA-Ub the lower TLS across TT 6-4 PP was associated with a higher accuracy is somewhat surprising . It suggests that maximal TLS for some lesions cannot be achieved without compromising fidelity . Thus , higher TLS does not necessarily mean higher fidelity , and for some lesions the advantage of more efficient TLS outweighs the cost of decreased fidelity . In the case of TT CPD TLS was very accurate in both cells types , arguing that the lack of PCNA-Ub did not cause a major change in the fidelity of TLS across this type of lesion . In summary , PCNA-Ub affects not only the efficiency of TLS , but also its mutagenicity . Our data show that a significant fraction of TLS in mammalian cells occurs in the absence of PCNA ubiquitination . This situation is different from the TLS in S . cerevisiae , where PCNA ubiquitination is essential for TLS [12] , [48] . The situation in chicken DT40 cells is more complex . It was proposed that Rev1 and PCNA-Ub function in distinct mechanisms that control TLS , and that PCNA-Ub is essential for filling postreplication gaps but not for fork progression , whereas Rev1-dependent TLS is important at stalled forks , but does not play a central role in gap filling [19] . Analysis of TLS across a site-specific TT 6-4 PP adduct in a plasmid showed normal activity in PcnaK164R mutant DT40 cells [20] . Thus , in DT40 cells , there is evidence for PCNA-Ub independent TLS . The situation in mammalian cells appears to be different , with a less distinct division of function between PCNA-Ub and Rev1 at postreplication gaps and stalled forks , respectively . Thus , unlike in DT40 cells , in mammalian cells both Rev1-dependent TLS [31] , and PCNA-Ub ( as described above ) are important for filling in of postreplication gaps . The fact that each of polη , polζ , and Rev1 contribute to UV survival of cells carrying the PcnaK164R mutation , as shown above , provides strong evidence for the participation of these polymerases in PCNA-Ub independent TLS reactions . This is in contrast to a previous study with XPV human cells , in which the expression of PCNA was reduced using siRNA , and supplementing the PCNAK164R mutant from a plasmid did not increase UV sensitivity [49] . The lack of effect in that study might have been caused by background levels of endogenous PCNA . How does TLS operate in the absence of PCNA-Ub ? A possible explanation can be proposed by considering the interactions that stabilize the TLS machinery acting on a damaged template . The composition of the TLS machinery is not fully understood , and neither is the composition of TLS complexes . However , based on the current knowledge we propose a model that includes a TLS complex with a minimal number of 3 proteins , namely the TLS DNA polymerase , PCNA and the Rev1 protein , acting as a scaffold ( Figure 8 ) . Depending on the type of DNA damage , other proteins are likely to be involved . Such a complex involves 7 known stabilizing interactions ( reviewed in [50] ) , which include the interactions of: ( 1 ) The TLS polymerase with the DNA . ( 2 ) The TLS polymerase ( via the PIP domain ) with PCNA . ( 3 ) The TLS polymerase with the ubiquitin at position K164 in PCNA . ( 4 ) The TLS polymerase with Rev1 . ( 5 ) Rev1 with ssDNA . ( 6 ) Rev1 with PCNA . ( 7 ) Rev1 with the ubiquitin at position K164 in another monomer of PCNA ( Figure 8 ) . In cells with the PcnaK164R mutation , two of these interactions are lost – of the ubiquitin with the TLS polymerase and with Rev1 ( Figure 8 lower panel ) . However , 5 stabilizing interactions are left , of which 3 directly involve the TLS polymerase: with PCNA , with the DNA , and with Rev1 . Thus , TLS DNA polymerases are recruited to damaged sites in DNA also in the absence of PCNA ubiquitination , and the TLS machinery is stable enough to perform lesion bypass , although at reduced efficiency . In conclusion , ubiquitinated PCNA is required for maximal TLS across a variety of lesions , supporting the model of recruitment of TLS polymerases to the damaged DNA via interaction of their ubiquitin-binding domain to PCNA-mUb [22] . Yet , TLS polymerases can be recruited to damaged DNA in the absence of PCNA-Ub and perform TLS , and although secondary in efficiency under normal conditions , they do contribute to the protection of cells against DNA damage .
The immortalized MEFs used in this study were prepared from mice with the following genotypes: Pcna+/+ and PcnaK164R/K164R [29]; Rad18+/+ and Rad18−/− [37]; Hltf+/+Shprh+/+ and Hltf−/−Shprh−/− [38]; Usp1+/+ , Usp1−/− , Usp1−/− complemented with wild-type Usp1 , and Usp1−/− complemented with the inactive Usp1C90S mutant [40] . The immortalized MEFs were cultured in DMEM supplemented with 2 mM L-glutamine , 100 units/ml of penicillin , 100 µg/ml of streptomycin , and 10% FBS . The primary MEFs were cultured in DMEM supplemented with 2 mM L-glutamine , 100 units/ml of penicillin , 100 µg/ml of streptomycin , non-essential amino acids ( Biological industries ) , 2-Mercaptoethanol 50 µM , and 15% FBS . The immortalized MEFs were incubated at 37°C in a 5% CO2 atmosphere . The primary MEFs were incubated at 37°C in a 5% CO2 and 4% O2 atmosphere . Separation of cells at G1/S phase of the cell cycle was performed by the elutriation method with the following modifications . The elutriation system consisted of a J6 Beckman elutriation centrifuge with a JE-5 . 0 rotor equipped with a single standard 5 ml elutriation chamber ( Beckman Coulter , Inc . , Fullerton , CA , USA ) , and a masterflex microprocessor pump drive , model 7524-05 ( Cole Parmer ) . The elutriation medium was DMEM supplemented with 1% FBS , maintained at room temperature . The speed and temperature of the rotor were set constant at 3000 rpm and 25°C . Approximately 3×108 Pcna+/+ or PcnaK164R/K164R MEFs were harvested from cultures at ∼80% confluence , centrifuged , and suspended in 10 ml of DMEM ( room temprature ) supplemented with 1% FBS . Cell suspensions were introduced to the elutriation chamber at a flow rate of 50 ml/min . After 15 minutes the flow rate was increased by 10 ml/min and three 50 ml fractions were collected at this flow rate . The flow rate was then gradually increased to 160 ml/min in 10 ml/min increments . Three 50 ml fractions were collected after each subsequent increase of the flow rate . The G1/S fraction ( analyzed by FACS ) was taken for the UV-induced Rpa foci assay . For Rpa immunostaining [33] , Pcna+/+ and PcnaK164R/K164R MEFs were fractionated by centrifugal elutriation , and cells in the G1/S boundary were seeded on 13 mm glass cover slips coated with 0 . 01% poly-L-lysine . After 2 h when the cells attached to the slides , the medium was removed and the cells were UV-C irradiated at 254 nm using a low-pressure mercury lamp ( TUV 15w G15T8 , Philips ) at doses of 8 J/m2 . The dose rate was measured using an UVX Radiometer ( UVP ) equipped with a 254-nm detector . At various time points after irradiation the cells were washed three times with PBS , pre-extracted with 25 mM HEPES pH 7 . 4 , 50 mM NaCl , 3 mM MgCl2 , 300 mM sucrose , 1% Triton X-100 for 5 minutes on ice with gentle shaking , and washed for three more times with PBS . The slides were then fixed in 4% paraformaldehyde for 15 minutes at room temperature and washed three times in PBS . Blocking was done in PBS supplemented with 5% normal goat serum for 30 minutes on ice . The cells were incubated for 4 hours on ice with anti-Rpa32 antibodies ( AbCam , cat . No . ab2175 ) that were diluted 1∶200 in blocking solution . After incubation the slides were washed three times in PBS and incubated with a secondary antibody –goat anti mouse Alexa Fluor 488 ( green ) diluted 1∶1000 , and with DAPI diluted 1∶1000 ( both in blocking solution ) for 45 minutes on ice . The slides were then washed three times in PBS and mounted on microscope slides using Aqua poly/Mount . Images were captured with a DeltaVision system ( Applied Precision ) equipped with an Olympus IX71 microscope . Optical images were acquired using CCD camera ( Photometrics , Coolsnap HQ ) and a 60×/1 . 42 objective ( Olympus ) . For each cell line at each time point at least 100 cells were counted and the percentage of cells exhibiting Rpa foci was determined . ( Figure S1 ) The assay was performed as previously described [34] , and the gapped plasmids with site-specific lesions used in this assay were prepared as previously described as follows: TT CPD and TT 6-4 PP [35]; cisPt-GG [6] . Briefly , cells were co-transfected with a DNA mixture containing 100 ng of a gapped-lesion plasmid ( GP-TT-CPD , or GP-TT-6-4 PP , or GP-cisPt-GG; kanR ) , 100 ng of a control gapped plasmid without a lesion ( GP20; cmR ) , and 5 µg of the carrier plasmid pUC18 , using jetPEI/DNA complexes for the immortalized MEFs or the Lipofectamine 2000 for the primary MEFs . After allowing time for gap filling and lesion bypass , the plasmids were extracted from the cells using alkaline lysis conditions , and used to transform an E . coli RecA reporter strain . The percentage of plasmid repair , of which most occurs by TLS , was calculated by dividing the number of transformants obtained from the gap-lesion plasmid ( kanR colonies ) by the number of transformants obtained from the control gapped-plasmid ( cmR colonies ) . A small fraction of gap-lesion plasmids can be repaired by non-TLS events , which involve formation of a DSB followed by DSB repair . These are observed as plasmid isolates with large deletions or insertions . To obtain precise TLS extents , the plasmid repair extents were multiplied by the fraction of TLS events out of all plasmid repair events , based on the DNA sequence analysis of the plasmids . To determine the DNA sequence changes that have occurred during plasmid repair , sequence analysis was carried using the TempliPhi DNA Sequencing Template Amplification Kit and the BigDye Terminator v1 . 1 Cycle Sequencing Kit . Reactions were analyzed by capillary electrophoresis on an ABI 3130XL Genetic Analyzer from Applied Biosystems . Two methods were used: depletion of ATP as a measure for viability , and colony forming ability after UV irradiation . For the viability ATP assay Pcna+/+ and PcnaK164R/K164R MEFs were seeded in 96-well plates . At 24 h after the seeding , cells were washed twice with Hanks' buffer , and irradiated in Hanks' buffer with UV-C at 254 nm using a low-pressure mercury lamp ( TUV 15w G15T8 , Philips ) . UV dose rate was measured using an UVX Radiometer ( UVP ) equipped with a 254-nm detector . After irradiation , Hanks' buffer was removed and the cells were incubated in a fresh growing medium for additional 48 h . Viability was determined using the CellTiter-Glo Luminescent Cell Viability Assay ( Promega ) . Luminescence was measured using an Infinite® M200 Luminometer ( Tecan ) . Throughout the entire experiment , none of the samples reached cell confluency . For the colony forming ability assay Pcna+/+and PcnaK164R/K164R immortalized MEFs were transfected with siRNA against TLS polymerases as described below , and incubated for 48 h . Cells were then trypsinized , counted , and plated in 10-cm Petri dishes . After incubation of 12 h , cells were UV irradiated as described above , and incubated in fresh medium for 10–12 days . Colonies were fixed and stained with 1% methylene blue ( Sigma ) . Colony forming ability was calculated by dividing the number of colonies in UV-irradiated plates by the number of colonies in unirradiated plates with pre-transfected with the same siRNA . The expression of specific DNA polymerase genes was knocked-down in Pcna+/+ and PcnaK164R/K164R MEFs by transfection with 50 nM of siRNA pools specific for PolH , Rev3L or Rev1 . The siRNAs used were from Dharmacon as follows: mRev3L SMARTpool ( M-04219 ) , mPolH ON-TARGETplus SMARTpool ( LU-063800 ) , mRev1 SMARTpool ( M-041898 ) , siGENOME non-targeting siRNA #5 ( D-001210 ) , ON-TARGETplus nontargeting Pool ( D-001810 ) . Transfection was carried out using HiPerFect ( Qiagen ) , according to the manufacturer recommendations . The effectiveness of knocking down the expression of TLS polymerases was measured by RT-PCR using total RNA that was extracted from the cells 48 h after transfection with siRNA , using the Perfect-Pure RNA cultured cells kit ( 5 PRIME ) . A hundred ng of total RNA was used for cDNA synthesis and RT-PCR by Maxime RT-PCR PreMix kit ( iNtRON BIOTECHNOLOGY ) according to the manufacturer recommendations . The following primers were used for the RT-PCRs: 5′-GTGGTACGAGTCTTCGG-3′ and 5′-TCTTGTGACTCGGGCTG-3′ for mREV3L , 5′-GAAGCCCGAGCATTTGGTG-3′ and 5′-GCCTCTCCTCAAGTTCCAG-3′ for mPOLH , 5′-AGAACGGAGAATGATGGC-3′ and 5′-GGCCCAGGATCCTCAGGTTTGCACACAGG-3′ for mRev1 , 5′-ACCACAGTCCATGCCATCAC-3′ and 5′-TCCACCACCCTGTTGCTGTA-3′ for mGAPDH . The results of knocking-down the expression of PolH , Rev3L and Rev1 are shown in Figure S2 .
|
DNA damage can block replication and lead to mutations , genomic instability , and cancer . In cases when the removal of DNA damage and restoration of the original sequence prior to replication is impossible , cells utilize DNA damage tolerance mechanisms , which help replication to bypass the lesions . A major universal tolerance mechanism is translesion DNA synthesis ( TLS ) , in which specialized low-fidelity DNA polymerases elongate the DNA across the lesion . This is a double-edged sword because the price of completing replication is an increased risk of point mutations opposite the lesion . Thus , TLS regulation is critical for preventing an escalation in mutation rates . A key element in TLS regulation is the attachment of a small protein called ubiquitin to the PCNA protein , a sliding DNA clamp that tethers the DNA polymerases to DNA , which functions to recruit the TLS DNA polymerase to the damaged site in DNA . While in yeast this modification of PCNA is crucial for TLS , there is a debate about its importance in mammals . Here we show that in mammalian cells the modification of PCNA by ubiquitin is important , but there exist secondary yet significant TLS mechanisms that operate in its absence and have an altered mutational outcome .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology"
] |
2011
|
PCNA Ubiquitination Is Important, But Not Essential for Translesion DNA Synthesis in Mammalian Cells
|
The genus Paracoccidioides consists of thermodymorphic fungi responsible for Paracoccidioidomycosis ( PCM ) , a systemic mycosis that has been registered to affect ~10 million people in Latin America . Biogeographical data subdivided the genus Paracoccidioides in five divergent subgroups , which have been recently classified as different species . Genomic sequencing of five Paracoccidioides isolates , representing each of these subgroups/species provided an important framework for the development of post-genomic studies with these fungi . However , functional annotations of these genomes have not been submitted to manual curation and , as a result , ~60–90% of the Paracoccidioides protein-coding genes ( depending on isolate/annotation ) are currently described as responsible for hypothetical proteins , without any further functional/structural description . The present work reviews the functional assignment of Paracoccidioides genes , reducing the number of hypothetical proteins to ~25–28% . These results were compiled in a relational database called ParaDB , dedicated to the main representatives of Paracoccidioides spp . ParaDB can be accessed through a friendly graphical interface , which offers search tools based on keywords or protein/DNA sequences . All data contained in ParaDB can be partially or completely downloaded through spreadsheet , multi-fasta and GFF3-formatted files , which can be subsequently used in a variety of downstream functional analyses . Moreover , the entire ParaDB environment has been configured in a Docker service , which has been submitted to the GitHub repository , ensuring long-term data availability to researchers . This service can be downloaded and used to perform fully functional local installations of the database in alternative computing ecosystems , allowing users to conduct their data mining and analyses in a personal and stable working environment . These new annotations greatly reduce the number of genes identified solely as hypothetical proteins and are integrated into a dedicated database , providing resources to assist researchers in this field to conduct post-genomic studies with this group of human pathogenic fungi .
The genus Paracoccidioides includes a series of thermodymorphic fungi responsible for causing a neglected tropical disease known as paracoccidioidomycosis ( PCM ) , which represents one of the most prevalent systemic mycoses in Latin America [1] . In fact , approximately 10 million people have been estimated to be infected by these fungi , which are distributed in large areas of Brazil , Argentina , Colombia , Venezuela , Ecuador and Paraguay [1 , 2 , 3 , 4 , 5] . The genus Paracoccidioides was originally proposed in 1908 , containing a single species , called P . brasiliensis [6] . Subsequent studies led to the characterization of several isolates , from different geographical regions , that display significant genetic variability , as well as differences in many biological characteristics , such as adaptability to laboratory culture , virulence and the ability to induce different host responses [7] . These isolates were initially distributed into five distinct subgroups ( Pl , S1 , PS2 , PS3 and PS4 ) that have been unofficially considered cryptic P . brasiliensis species over the years [8 , 9] . In 2009 , the P1 subgroup was classified as a new species , called P . lutzii , since its isolates display deeper genetic divergence , when compared with representatives of the other subgroups , which remained classified as members of the P . brasiliensis species complex [7 , 10] . More recently , the four subgroups within the P . brasiliensis complex have also been described as different species: P . brasiliensis ( subgroup S1 ) , P . americana ( subgroup PS2 ) , P . restrepiensis ( subgroup PS3 ) and P . venezuelensis ( subgroup PS4 ) [11] . Genomic studies involving Paracoccidioides spp . started to be developed in 2003 , by large-scale sequencing/characterization of Expressed Sequence Tags ( ESTs ) obtained from the isolate Pb18 , which is the main representative of P . brasiliensis ( S1 subgroup ) [12 , 13] . Subsequently , functional studies , based on the information derived from these EST analyses , demonstrated the potential of genomic approaches to increase our knowledge regarding the genetic bases that determine virulence in these fungi , as well as to provide information that may contribute to the development of new alternatives for the control and treatment of PCM [14 , 15 , 16] . These pioneering studies motivated the development of complete genome projects , which led to the characterization of draft genomes of three Paracoccidioides isolates: Pb01 , Pb03 and Pb18 ( representatives of P . lutzii , P . americana and P . brasiliensis , respectively ) [17] . This work represented an important milestone to the genetic study of this group of fungi , providing clues that helped us to better understand the evolution of the genus Paracoccidioides , as well as a series of genomic characteristics that differentiate some of the abovementioned species/subgroups . However , the sequencing of these three isolates was performed using Sanger technology , generating assemblies with large contig numbers and presenting several regions with low quality consensus sequences , which led to the development of incomplete and inaccurate genomic annotations ( v1 ) for these fungi [17] . Later on , these same isolates were submitted to a new sequencing , using Illumina's NGS platform , in order to produce more complete and precise assemblies [18] . Moreover , a re-annotation analysis performed with such assemblies allowed recovery of a large number of genes that were missed by the original annotation ( v1 ) , and this second annotation ( v2 ) was more consistent across the three reference genomes ( Pb18 , Pb03 , and Pb01 ) . Finally , these analyses were extended to contemplate the genomes of additional isolates , representing P . restrepiensis ( isolate PbCnh ) and P . venezuelensis ( isolate Pb300 ) , providing reference genomes and annotations for isolates representing all five species/subgroups of the genus Paracoccidioides [19] . Currently , genomic data from these five Paracoccidioides isolates can be obtained from several generic databases , such as GenBank [20] and Ensembl [21] , as well as from some fungal specific databases , such as MycoCosm [22] or FungiDB [23] , but the genomic annotations provided through all these repositories are inconsistent and display an unusually large number of protein-coding genes described as responsible for hypothetical proteins . For example , GenBank and RefSeq describe ~62% of all Pb01 genes in association with hypothetical proteins and this proportion is even larger ( up to 88% ) in the genomes of Pb18 , Pb03 , Pb300 and PbCnh . A similar situation is observed in other databases , such as Ensembl and FungiDB , which provide the same annotation data found in GenBank/RefSeq . On the other hand , MycoCosm presents an alternative annotation for Pb18 , in which a smaller proportion of genes ( ~68% ) is described as associated with hypothetical proteins . However , MycoCosm does not present any information regarding other Paracoccidioides isolates ( except for Pb03 , but these data are based on outdated sequencing information , as they relate to the first version of the Pb03 genome , described by [17] ) . All these discrepancies , as well as the overall low level of functional gene categorization observed among Paracoccidioides isolates may partly derive from the fact that the abovementioned databases have not been submitted to appropriate manual curation , since they are dedicated to providing genomic information for a large number of organisms that may share little genomic similarity or phylogenetic proximity . Thus , to improve and standardize the current genomic functional annotations of the main Paracoccidioides isolates , coding sequences ( CDSs ) derived from the latest genomic assemblies obtained for Pb18 , Pb03 , Pb01 , PbCnh and Pb300 [18 , 19] were initially submitted to comparative BLAST analyses against a series of databases , including generic functional databases ( InterPro , Pfam and Swiss-Prot ) [24 , 25 , 26] and fungal-specific , manually-curated databases ( Saccharomyces Genome Database , Candida Genome Database and Aspergillus Genome Database ) [27 , 28 , 29] . Information derived from all these BLAST analyses were compiled in spreadsheets , along with specific Gene Ontology ( GO ) classifications [30 , 31] . This metadata was used to develop a manually curated consensus annotation for each of these Paracoccidioides genomes . As a result of this process , the number of genes described in association with hypothetical proteins has been reduced to ~25–28% , in all isolates . The information derived from this reannotation effort has been compiled in a publicly available database named ParaDB ( available at http://paracoccidioides . com ) [32] , aimed at centralizing up-to-date genomic annotations for the major representatives of the five species/subgroups that compose the genus Paracoccidioides . Using a friendly graphical interface , ParaDB allows users to browse and download functional information for any set of genes from any of the abovementioned Paracoccidioides genomes . The ParaDB webpage also provides search tools based on keywords or DNA/protein sequence similarity , as well as fully reannotated genome files , in multi-fasta or General Feature ( GFF3 ) formats , which may greatly assist researchers in a variety of large-scale , post-genomic studies with this important group of human pathogenic fungi . Finally , the entire ParaDB environment has been configured in a Docker service [33] , which has been submitted to both the GitHub and Open Science Framework repositories , ensuring long-term data availability to researchers . This service can be downloaded and used to perform fully functional local installations of the database in alternative computing ecosystems , allowing users to conduct their data mining and analyses in a personal and stable working environment .
Files containing annotated protein coding sequences ( CDS genomes ) of the Paracoccidioides isolates were downloaded from NCBI , using the following accession numbers: Pb18 ( RefSeq# GCF_000150735 . 1 ) , Pb03 ( GenBank# GCA_000150475 . 2 ) , Pb300 ( GenBank# GCA_001713645 . 1 ) , PbCnh ( GenBank# GCA_001713695 . 1 ) , and Pb01 ( RefSeq# GCF_000150705 . 2 ) . The CDSs from these genomes were compared against each other , in order to identify all groups of orthologous genes ( OGs ) shared by two or more of the isolates , with the aid of the software OrthoFinder [34] , using the software´s default parameters . Paralogous genes present within the same OG group were compared by multiple alignment , using Clustal Omega 1 . 2 . 4 [35] . The input parameters were set as follow: Output guide tree: false; Output distance matrix: false; Dealign input sequences: false; mBed-like clustering guide tree: true; mBed-like clustering iteration: true; Number of iterations: 0; Maximum guide tree iterations: -1; Maximum HMM iterations: -1 . The Nexus-formatted matrix generated by Clustal Omega was then used to estimate genealogical relationships with the aid of Bayesian inference , using Mr . Bayes 3 . 0 [36] . The analysis involved 1 , 000 , 000 iterations , with savings at every 100th tree , 1 , 100 , 000 generations , in four heated Monte Carlo Markov chains ( MCMCs ) , with 0 . 5 annealing temperature , 100 000 MCMC generation burn-in and a 16-category C distribution . A consensus tree was generated after burn-in , using a 50% majority rule , which allowed discrimination between orthologous and co-orthologous genes in the different OG groups . This list of orthologues was then used as a guide to ensure consistent annotation of equivalent genes throughout the five Paracoccidioides isolates , during the reannotation process ( see below ) . Several genes that transcribe non-coding RNAs ( ncRNAs ) have been identified and annotated in the genomes of Pb01 and Pb18 ( the only formal datasets available for ncRNAs in Paracoccidioides spp . ) . Thus , their sequences were downloaded from the Ensembl Fungi database ftp site [21] and orthologues for each of these ncRNA genes were mapped in the genomes of Pb03 , PbCnh and Pb300 , using Bwa-MEM , version 0 . 7 . 17 . 1 [37] , running in a local Galaxy environment [38] , using the default software parameters . The resulting BAM alignment files were converted to BED files , with the aid of BAM-to-BED Converter , version 2 . 27 . 1 [39 , 40] , also using default software parameters , which facilitated organizing and comparing the predicted ncRNAs across all Paracoccidioides spp . isolates . Finally , information regarding these ncRNAs was incorporated into GFF3 files ( see below ) , with the aid of BED-to-GFF Converter [41] , version 2 . 0 . 0 , also using default parameters . All ncRNAs ( along with their respective annotations ) received identification codes consistent with the ones currently employed to describe Gene_IDs in each Paracoccidioides genome , but containing the designation NC ( for non-coding ) as a suffix . Thus , ncRNAs mapped in the genome of Pb18 received Gene_IDs starting from PADGNC_00001 , while ncRNAs for Pb01 , Pb03 , Pb300 and PbCnh received Gene_IDs starting from PAAGNC_00001 , PABGNC_00001 , ACO22NC_00001 and GX48NC_00001 , respectively . Genes responsible for transcribing additional ncRNAs , including tRNAs and rRNA genes ( 18S , 28S and 5S rRNAs ) had been previously described in the original annotations of the Paracoccidioides spp . genomes [18] and sequences for such elements were available from the "rna_from_genomic" fasta files downloaded from GenBank/RefSeq [20] . These genes were also matched to their respective orthologues , using the same procedure described above , but we chose not to change their respective gene IDs ( i . e . : they were not labeled with the designation NC ) , in order to respect their current GenBank/RefSeq IDs . The overall process employed for reannotating the Paracoccidioides genomes is schematically shown in S1 Fig . Initially , all CDSs from Pb18 were individually submitted to comparative BLAST analyses against InterPro , Pfam and Swiss-Prot . Next , these CDSs were BLASTed against the manually-curated fungal databases SGD ( Saccharomyces Genome Database ) , CGD ( Candida Genome Database ) and AspGD ( Aspergillus Genome Database ) [27 , 28 , 29] . All BLAST analyses employed high stringency criteria , which included as cutoff , an E-value < e-10 , to identify orthologous genes containing functional descriptions among these databases . Information derived from all these BLAST analyses were compiled in a spreadsheet , along with Gene Ontology ( GO ) data regarding Pb18 genes , obtained from the Database for Annotation , Visualization and Integrated Discovery ( DAVID ) , version 6 . 8 [42] . GO data were downloaded through the GO Direct option , in order to reduce the redundancy of terms , typically observed in GO analyses . Next , the metadata regarding each CDS was independently analyzed by three expert reviewers , in order to determine a consensus annotation term ( ParaDB Annotation ) for each CDS , which should be consistent with all the information derived from DAVID and from the BLAST searches previously performed . In a second round of annotation , a fourth reviewer compared the results of the three independent analyses and determined a final ParaDB Annotation term for each CDS . Finally , all information regarding the ParaDB Annotation , obtained for each Pb18 CDS , was transferred to the orthologues present in any of the Paracoccidioides isolates , using the list of orthologous genes ( available at http://paracoccidioides . com/paracoccidioides-orthologous/ ) as a guide . Next , all CDSs present in the genome of Pb01 , which did not contain an orthologue in Pb18 , were submitted to the same analysis procedure described in the previous paragraph . The same process was successively repeated with the remaining CDSs from Pb03 , Pb300 and PbCnh , generating thorough and consistent annotations for all Paracoccidioides genomes . In a subsequent annotation step , all CDSs that remained identified as hypothetical proteins were submitted to additional BLAST analyses , using a less stringent cutoff ( E-value < e-5 ) , essentially as described above . These CDSs were incorporated into the ParaDB database ( see below ) with a flag ( E-value < e-5 ) highlighting the lower stringency criterion used to determine their respective functional annotation . The information derived from the reannotation process described above has been compiled in a relational database named ParaDB , aimed at centralizing up-to-date genomic annotations for the major representatives of the five species/subgroups that compose the genus Paracoccidioides . ParaDB is available at the URL http://paracoccidioides . com and provides users with tabular archives describing each CDS and ncRNA sequences , including their final ParaDB Annotation consensus description , along with information derived from all databases evaluated during the present study . Additional information regarding these elements in all five Paracoccidioides isolates can also be downloaded ( as nucleotide , or amino acid sequences ) through multi-FASTA files , carrying strings that show , for each element , their respective locus_tag ID , protein_product ID and consensus_description ( ParaDB Annotation ) . Updated General Feature Format ( GFF3 ) files for each genome are also available through ParaDB , to assist researchers interested in performing large-scale OMICs analyses with the Paracoccidioides spp . genomes . These GFF3 files have been built upon the original GFF3 files available in RefSeq ( Pb18 and Pb01 ) and GenBank ( Pb03 , Pb300 and PbCnh ) , by replacing the original GenBank/RefSeq annotations with the respective ParaDB Annotation , with the help of MS Excel . Information regarding the ncRNAs obtained from Ensembl were introduced in the GFF3 files using the BED-to-GFF Converter , as described above . The ParaDB environment is based on the Database Management System ( DBMS ) MySQL and was developed in Docker [33] . Management configuration of the DB was made using Rancher [43] , a robust Docker systems management tool , widely used in data center environments and other complex computing ecosystems . The Rancher cluster created to manage ParaDB resources is hosted in a cloud computing environment at CloudatCost [44] . Currently , the environment provides a total of 200 GB of disk space , in solid state drives ( SSDs ) , and 20 GB of random-access memory ( RAM ) . These resources are distributed in 10 virtual CPUs ( vCPUs ) , in a Kubernetes cluster framework [45] ( see S2 Fig for details ) . The ParaDB user interface has been developed in PHP language , using Wordpress [46] and tools to assist in keyword/BLAST searches were implemented with the help of Wordpress plugins and widgets [47] . A Docker service , carrying the Docker virtualization containers necessary to run ParaDB can be downloaded from https://cloud . docker . com/u/paradb/ , allowing users to perform fully functional local installations of ParaDB , in different computational environments , given the platform-agnostic nature of Docker systems ( see below ) .
To guarantee consistent genomic annotation of protein-coding genes among Paracoccidioides isolates , orthologous genes shared by two or more isolates were initially identified with the aid of the software OrthoFinder [34] and the Paracoccidioides spp . pan-genome derived from this analysis , containing all protein-coding sequences within the group , is shown in S3 Fig ( a complete description of this pan-genome can be found in the ParaDB website , at http://paracoccidioides . com/paracoccidioides-orthologous/ ) . Overall , the five isolates display a protein-coding pan-genome composed of 8365 groups of orthologous genes ( OG ) and share a core genetic pool that consists of 6396 OGs ( ~75% ) , reinforcing the close phylogenetic relatedness among members of this group of human pathogenic fungi . Surprisingly , very few OGs could be found in association with only one isolate . In fact , no exclusive OGs were found in the genomes of Pb3 and Pb300 , while Pb18 carries only 2 exclusive OGs of this type ( OG0000046 and OG000738 ) and PbCnh displays only one ( OG0006453 ) . Even Pb01 , which represents P . lutzii , the most distantly related species within the Paracoccidioides genus , contained only 2 exclusive OGs carrying protein-coding genes ( OG0000018 and OG0006445 ) . Not surprisingly , most of these OGs carry genes that typically display structural variations even among closely-related species , since they may perform similar biochemical functions , but interact with alternative substrates: OG0000018 and OG0000046 contain a series of Ser-Thr Protein Kinases , while OG0006445 and OG0006453 contain a series of plasma membrane ATP-binding cassette ( ABC ) transporters ( OG0007385 contain genes that remained identified as hypothetical proteins ) . Currently , it is not possible to establish whether these genes contribute any kind of adaptive/biological specificities for the different Paracoccidioides isolates . GenBank/RefSeq [20] also contained information regarding a series of non-coding RNAs from the five Paracoccidioides spp . isolates , including tRNAs and rRNA genes ( 18S , 28S and 5S rRNAs ) ( see Methods ) . All isolates displayed a similar number of tRNA genes , capable of providing all amino acids required for protein synthesis ( see OG0008368 to OG0008478 , at http://paracoccidioides . com/paracoccidioides-orthologous/ ) . A similar situation was observed with the 5S rRNA genes ( OG0008365 ) , but the 18S and 28S rRNAs were present in significantly different copy numbers across the genomes of each isolate and could not be found in the genome of isolate PbCnh , probably reflecting problems with the currently available genomic assemblies ( see OG0008366 and OG0008367 , at http://paracoccidioides . com/paracoccidioides-orthologous/ ) . A total of 32 additional ncRNA genes , responsible for transcribing small RNAs ( sRNAs ) , small nuclear RNAs ( snRNAs , including the spliceosome RNAs ) , small nucleolar RNAs ( snoRNAs ) , the Telomerase RNA Component ( tercRNA ) and the Signal Recognition Particle RNA have also been mapped in the genomes of all five isolates , using as reference , a list of ncRNAs identified in the genomes of Pb18 and Pb01 , available from the Ensembl Fungi database [21] ftp site . All Paracoccidioides spp . isolates share a common set of such elements , whose orthologues were easily identified in all genomes , with the aid of the short read Bwa-MEM aligner , as described in Methods ( see OG0008479 to OG0008510 , at http://paracoccidioides . com/paracoccidioides-orthologous/ ) . To prevent using different nomenclature while annotating corresponding genes across the Paracoccidioides genomes , the reannotation process of protein-coding genes was carried out as described in Methods . Thus , genome reannotation of Paracoccidioides spp . was initially performed with Pb18 and the annotation data obtained for all genes in this isolate were propagated to the corresponding genes present in the remaining Paracoccidioides spp . genomes , using the list of orthologous genes , described above , as a guide . Next , genes present in the genome of Pb01 , which did not contain an orthologue in Pb18 , were submitted to the same procedure and the same process was successively repeated with the remaining genes from Pb03 , Pb300 and PbCnh , generating consistent annotations for all Paracoccidioides spp . genomes . As a result , specific functions and/or structural descriptions could be assigned to 6003 out of 8390 protein coding genes mapped in the genome of Pb18 , reducing the proportion of genes described in association with hypothetical proteins to 28 , 5% ( 2386 genes ) ( Fig 1 ) . In a second reannotation step , all CDSs that remained identified as hypothetical proteins were submitted to a new BLAST analysis , using a less stringent cutoff ( E-value <e-5 ) , allowing functional identification of additional 241 CDSs in Pb18 , further reducing the proportion of hypothetical proteins in this isolate to 25 . 5% ( 2145 genes ) ( Fig 1 ) . Reannotation of the remaining Paracoccidioides spp . genomes provided similar results with the other isolates , reducing the number of CDSs associated with hypothetical proteins to ~29–25% ( with E-values < e-10 and < e-5 , respectively ) , thus increasing the proportion of genes with functional/structural identification to ~71–75% ( with E-values < e-10 and < e-5 , respectively ) in all cases , significantly improving the annotations of Paracoccidioides genomes , in comparison to the annotations currently available in any public biological data repository ( see Fig 1 ) . Results from the functional reannotations shown in Fig 1 were compiled in ParaDB , a relational database developed to centralize standardized functional genomic annotation from all five reference isolates of the genus Paracoccidioides . ParaDB ( available at http://paracoccidioides . com ) presents a simple and intuitive interface , through which such information is made available ( S4 Fig ) . Initial access to the annotation data can be made by the “Databases” button at the center of the webpage , or through a specific pull-down menu , available at the upper right corner of the main ParaDB webpage ( S4A Fig ) . The “Full Database” option directs users to specific annotation data for each of the Paracoccidioides isolates under study ( S4A Fig ) . In the “Full Database” mode ( S4B Fig ) , users have initial access to a table that displays each Paracoccidioides spp . CDS ( identified by numeric codes that correspond to their original GenBank/RefSeq IDs ) and their respective ParaDB consensus functional/structural designation . Genes responsible for transcribing ncRNAs , which were not present in the GenBank/Refseq CDS files can also be accessed from these table and carry the suffix NC ( non-coding ) in the gene IDs assigned to them during our reannotation effort ( see Methods ) . Information regarding data derived from all databases employed in the comparative analyses described above can be accessed by clicking on the ( + ) symbol , available in each of the CDS cells . Alternatively , such information can be accessed by clicking the “Columns” button on the upper right corner of the table and selecting the desired databases ( S4B Fig ) . In either case , links are provided to direct users to the orthologous genes found in the fungal-specific databases ( SGD , CGD and AspGD ) , where a variety of additional information can be found . Users may also download the entire annotation files using the “Downloads” button , available from the pull-down menu ( S4 Fig ) . The download site also allows users to access multi-FASTA files ( containing either nucleotide or amino acid sequences ) for each Paracoccidioides strains and/or their respective General Feature ( GFF3 ) format files , which may be of great assistance for a series of downstream analyses , such as the evaluation of data derived from large-scale gene expression experiments involving microarray hybridization , or RNA-seq , for example . ParaDB can be accessed and browsed directly on the web , at the URL http://paracoccidioides . com , through a variety of platforms , including personal computers of any kind , as well as mobile devices , such as cell phones and tablets , operating under either android or iOS operating systems . However , it is also possible for users to download and install a fully functional version of the database in their own personal computers , avoiding problems derived from low internet trafficking , host server instability , or communication restrictions , due to the presence of firewalls in local servers , for example . To accomplish that , ParaDB was developed in Docker [33] , allowing configuration of the entire ParaDB environment in a Docker service , which has been submitted to both GitHub ( https://github . com/paracoccidioidesdb ) and Open Science Framework ( https://osf . io/3sq97/ ) repositories , ensuring long-term data availability to researchers . This service can be downloaded and used to perform local installations of the database in alternative computing ecosystems , allowing users to conduct their data mining and analyses in a personal and stable working environment . Once installed in a local host , ParaDB will operate from two single containers , called ParaDB-Web and ParaDB-BLAST . Both containers can be consistently interchanged and deployed across different platforms , regardless of hardware and/or operating system ( OS ) specificities . This implementation is designed to ensure continued and full availability of ParaDB , independently of the original installation in our servers , allowing users to maintain mirrors of the entire database in their local environments . Additionally , the container concept allows the ParaDB infrastructure to be easily scalable , ensuring that hardware resources are provisioned whenever the computational environment reaches its limitations . Hardware requirements for a local installation of ParaDB are reduced , and a personal computer ( or notebook ) , running on Linux ( recommended ) , containing 2 GB of RAM and 5 GB of available disk space can be used as host for installing a local version of ParaDB . The installation process is extremely simple and only requires previous installations of Docker [33] and Docker Compose in the host machine [48] . Once these components are available , only two steps are required to start a ParaDB environment in the host . In the first step , a Docker-Compose file is downloaded from the GitHub servers to the host machine . In the second step , the Docker Compose is executed , downloading and installing the images/containers of the standard ParaDB modules , starting the service . When Linux is the host machine's operating system , the following commands must be run on the terminal: $ git clone https://github . com/ParacoccidioidesDB/paradb . git $ cd paradb $ docker-compose up -d During the deployment process , some ports and disk volumes will be automatically configured on the host machine . Details about the ports and volumes created are available on the ParaDB website . In a standard implementation , ParaDB will use the local host address ( IP: 0 . 0 . 0 . 0; HOST: http://localhost ) as the default address for internal links . A video demonstrating the entire ParaDB implementation process in a local Linux environment is available at the URL http://paracoccidioides . com/local-install/ . Finally , it is worth mentioning that the ParaDB infrastructure can also be used by independent researchers to develop genome annotation projects involving other closely related fungi and its source code is freely available at: https://cloud . docker . com/u/paradb/ .
Databanks dedicated to the storage of genomic data from Paracoccidioides spp . started to be designed since the pioneering EST analyses conducted with isolate Pb18 [12 , 13] and the information provided by these datasets greatly assisted the scientific community in a large number of projects , which contributed to improve our knowledge about this important group of human pathogenic fungi ( recently reviewed by [49 , 50] ) . However , these original EST databases were gradually abandoned or deactivated after the release of the first draft genomes obtained for isolates Pb18 , Pb03 and Pb01 [17] . At this time , data regarding these draft genomes were deposited in a database dedicated to members of the genus Paracoccidioides , as part of the Fungal Genome Initiative ( FGI ) , developed by the Broad Institute of Harvard University and the Massachusetts Institute of Technology . This centralized database became the major source of information for subsequent genomic work on Paracoccidioides spp . , as it contained a great deal of genomic data from these three isolates ( including gene functional annotations , chromosome locations of loci and comparative evolutionary analyses of multiple gene sets ) , as well as tools for searching and downloading genetic sequences and other additional information . However , maintenance of the FGI databases , as well as their respective web interfaces , was discontinued in 2015 and , up to this moment , no other database has been able to reproduce a centralized and efficient environment for genomic analysis on Paracoccidioides spp . Efforts were initially made to incorporate Paracoccidioides genomic data into other sites that support comparative analysis of fungal genomes , including MycoCosm and FungiDB [22 , 23] . FungiDB is a subsection of the EuPathDB family of databases , maintained by the Welcome Trust and NIH . It was designed to combine and make available a plethora of biological information , obtained from a wide variety of microbial eukaryotes . FungiDB [23] includes data from both pathogenic and non-pathogenic fungi and provides information about multiple genomes and gene records , which can be compared and downloaded with the aid of user-friendly browsers . It also integrates genomic data with comments and supporting evidence from the scientific community ( including PubMed IDs , images , phenotypic information , etc . ) and offers tools for integrating and mining diverse Omics datasets . MycoCosm ( supported by JGI/DOE ) offers a large collection of fungal genomes , along with interesting web-based tools for alternative types of genome-scale analyses [22] . However , in spite of the effective resources made available through these repositories , the genomic data regarding Paracoccidioides isolates that can be currently found in these databases are discrepant , apparently due to absence of manual curation and/or to the confusion generated by the publication of a second version of Paracoccidioides genomes ( and their respective annotations ) [18 , 19] . For example , detailed analysis of the information available from MycoCosm [22] shows genomic data only for isolates Pb18 and Pb03 . However , Pb18 data refer to version 2 of the genome [18] , while Pb03 data refer to version 1 [17] . This represents a serious problem for studies involving Pb03 , since annotations referring to genomes v1 and v2 display only a portion of common genes [18] . Additionally , the Pb18 genome has ~68% of its CDSs described solely as responsible for encoding hypothetical proteins . EuPathDB/FungiDB [23] , on the other hand , presents data for the three Paracoccidioides isolates , but only contemplate the genomic information described by [17] . Moreover , their annotations display puzzling results , since the genome of Pb01 displays ~62% of its CDSs identified as hypothetical proteins , while this proportion increases to ~87–88% in the genomes of Pb18 and Pb03 . These results represent an unexpected discrepancy , especially when confronted with the comparative analysis of orthologues described herein and shown in S2 Fig . Actually , a closer analysis shows that the data contained in FungiDB [23] appear to have been incorporated directly from NCBI ( GenBank/RefSeq ) [20] or Ensembl [21] , without receiving any manual curation to enhance their accuracy . GenBank/RefSeq and Ensembl are large databases dedicated to providing genomic information for a large number of organisms , relying mostly on automated pipelines to perform genomic annotations . However , these automated pipelines perform comparisons against a large number of independent databanks , resulting in large amounts of data , which are often too complex to be automatically summarized by computer algorithms , requiring manual evaluation by expert reviewers , in order to establish a consensus nomenclature for each analyzed sequence and ensure greater efficiency in the functional identification of the genes present in an organism [51 , 52 , 53 , 54] . Unfortunately , manual curation of genomic data is a time-consuming process and the large number of genomes currently deposited in generic databases , such as GenBank/RefSeq [20] and Ensembl [21] , causes many of them to be displayed solely as the result of automated annotation pipelines [55 , 56] . As expected , manual curation of the Paracoccidioides genomes , as shown herein , reduced the proportion of genes described in association with hypothetical proteins from ~90% , in most isolates , to < 30% , in all organisms under study . Moreover , the information regarding these newly annotated genomes have been standardized and made available through a single public database , centralizing the genomic data for the main representatives of the group . Similar improvement in genomic annotation has also been verified with the human pathogenic fungus Candida albicans , whose genome has been submitted to manual curation . In fact , the C . albicans reference genome ( RefSeq #GCF_000182965 . 3 ) displays ~39% of its genes annotated as responsible for hypothetical proteins , while a manually-curated reannotation , made available through the Candida Genome Database ( CGD ) project [28] reduced the proportion of hypothetical proteins to only ~21% [57] . Unfortunately , microorganisms responsible for neglected diseases tend to attract less attention from the scientific community and , as a result , their genomic annotations are often described with considerably lower accuracy , as manually curated reannotations are rarely performed . For example , the genome of the fungus Blastomyces dermatitidis , strain ER-3 ( etiologic agent of blastomycosis ) , available through GenBank ( accession #GCA_000003525 . 2 ) , shows ~ 50% of its genes described in association with hypothetical proteins . Similar scenarios can be verified with the genomes of Cryptococcus neoformans var . grubii H99 and Coccidioides immitis RS , responsible for cryptococcosis and coccidioidomycosis , respectively , which present 47–49% of their CDSs described in association with hypothetical proteins ( see RefSeq accessions #GCF_000149245 . 1 and GCF_000149335 . 2 ) . A natural consequence derived from such lack of accuracy is that it is often difficult to use the information derived from these genomic data in large-scale post-genomic studies , such as in silico functional/metabolic reconstructions and transcriptome/proteome analyses , which could greatly contribute to improve our knowledge regarding the general biology of these fungi , or the molecular basis of their pathogenicity mechanisms . In this sense , the work described in this manuscript provides manually curated and standardized genomic annotations for the main representatives of all five species of Paracoccidioides spp . , placing members of this genus in a unique position , when compared with many dimorphic fungi , responsible for neglected mycoses . These new annotations greatly reduce the number of genes identified solely as hypothetical proteins and are integrated into a dedicated database that provides different search/analyses tools to facilitate the development of future post-genomic studies with this important group of human pathogenic fungi . It must also be highlighted that the data available from ParaDB should provide adequate genomic coverage of protein-coding genes to support in silico metabolic analyses in Paracoccidioides spp . , since the current genomic assemblies display sizes between 29 and 32 Mb , encoding approximately 8000 to 9000 proteins . Thus , gene density in these genomes is 1 CDS/~3 . 5 kb , which is close to the values observed in well-annotated fungal genomes , such as the cases of Aspergillus spp . ( 1 CDS/~3 . 1 kb ) [29] and Candida spp . ( 1 CDS/~2 . 3 kb ) [28] . Thus , the current Paracoccidioides annotations are likely to have identified most ( if not all ) protein-coding genes present in these fungi , especially when compared with other neglected fungi , such as H . capsulatum ( 1 CDS/~4 kb ) [58] and B . dermatitidis ( 1 CDS/~5 . 7 kb ) [59] . However , information regarding the presence of non-coding elements in Paracoccidioides spp . is still scarce and such elements have only recently begun to be unraveled [60] . We expect further versions of ParaDB to incorporate more information on these elements , which can be more efficiently characterized through the analysis of transcriptome data . Finally , the work presented in this manuscript proposes a pioneering and effective alternative to ensure that the data and resources provided by ParaDB shall remain available in a continued and reproducible way , by providing users with the possibility of installing fully functional mirrors of the database in their own working environments . This was accomplished by developing the entire ParaDB environment in Docker , which allowed the creation of a ParaDB Docker service that was deposited in both GitHub https://github . com/paracoccidioidesdb [61] and Open Science Framework repositories ( https://osf . io/3sq97/ ) , two of the world´s largest web-based hosting servers for open source software . The ParaDB image can be freely downloaded and deployed in any kind of local computer , with little infrastructure requirements . The Docker project is providing a new and promising virtualization strategy that consumes a considerably low amount of disk space ( when compared to Virtual Machines ) and offers the advantage of being platform-agnostic , since it relies on the configuration of containers , which can be consistently interchanged and deployed on different computing environments , regardless the specificities of their hardware and/or operating system [33] . In recent years , this type of technology has been increasingly employed to generate bioinformatics-related software and services to a large variety of research facilities , employing the concepts of Platform as a Service ( PaaS ) and Software as a Service ( SaaS ) , as a strategy to assist in replicability and reproducibility of data analysis across laboratories [62 , 63 , 64 , 65 , 66 , 67 , 68] . In this context , ParaDB is the first initiative that tries to develop a Docker system with a biological database , carrying genomic information of pathogenic microorganisms , thus introducing the concept of Database-as-a-Service ( DBaaS ) , as a strategy to guarantee long term availability of biological data and resources .
|
The genus Paracoccidioides comprises fungi responsible for Paracoccidioidomycosis ( PCM ) , a neglected tropical disease prevalent in South America that has been shown to affect approximately 10 million people and has great medical/social impact , since available treatments are poorly effective , frequently leading to relapses , chronic infections and sequelae . Genomic information available for five reference Paracoccidioides isolates could greatly assist researchers in developing new chemotherapeutic approaches against PCM , but usefulness of such data is limited , since ~60–90% of Paracoccidioides protein-coding genes ( depending on isolate ) are described as responsible for hypothetical proteins , without any functional/structural description . Such elevated number of hypothetical proteins is unexpected and probably derives from annotations performed solely by automated computing pipelines . This problem can be minimized by manual curation , when expert reviewers determine the functional designation of each gene , after comparing results derived from several reference databases . This work describes an effort to review the functional assignment of >40 , 000 genes , annotated across the five Paracoccidioides genomes mentioned above , which reduced the number of hypothetical proteins to ~25–28% , contributing to significantly increase quality and usefulness of such genomic information . These data have been compiled in a relational database named ParaDB , constituting an important resource for researchers in the field .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"fungal",
"genetics",
"fungi",
"genome",
"analysis",
"genome",
"annotation",
"research",
"and",
"analysis",
"methods",
"paracoccidioides",
"mycology",
"genomics",
"biological",
"databases",
"comparative",
"genomics",
"fungal",
"genomics",
"eukaryota",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"computational",
"biology",
"genomic",
"databases",
"organisms"
] |
2019
|
ParaDB: A manually curated database containing genomic annotation for the human pathogenic fungi Paracoccidioides spp.
|
Previous game-theoretic studies of vaccination behavior typically have often assumed that populations are homogeneously mixed and that individuals are fully rational . In reality , there is heterogeneity in the number of contacts per individual , and individuals tend to imitate others who appear to have adopted successful strategies . Here , we use network-based mathematical models to study the effects of both imitation behavior and contact heterogeneity on vaccination coverage and disease dynamics . We integrate contact network epidemiological models with a framework for decision-making , within which individuals make their decisions either based purely on payoff maximization or by imitating the vaccination behavior of a social contact . Simulations suggest that when the cost of vaccination is high imitation behavior may decrease vaccination coverage . However , when the cost of vaccination is small relative to that of infection , imitation behavior increases vaccination coverage , but , surprisingly , also increases the magnitude of epidemics through the clustering of non-vaccinators within the network . Thus , imitation behavior may impede the eradication of infectious diseases . Calculations that ignore behavioral clustering caused by imitation may significantly underestimate the levels of vaccination coverage required to attain herd immunity .
Vaccination is the primary public health measure for preventing transmission of infectious diseases as well as reducing morbidity and mortality from infections [1] . An individual's decision-making with respect to vaccination may depend on perceived risk of infection , cost of infection , cost of vaccination , and the vaccinating behaviors of other individuals [2] , [3] , [4] . Game theory has been integrated into epidemiological models to investigate vaccination behaviors [5] , [6] , [7] , [8] . Previous game-theoretic studies on vaccination dynamics typically assume that the population is homogeneously mixed and fully rational , defined as making decisions that yield the highest personal utility based on their perceived risks . In reality , there is individual heterogeneity in the number of contacts [9]–[22] and individuals frequently imitate behaviors of their contacts [23] , [24] , particularly those who appear to have adopted successful strategies [25] , [26] . In addition , peer influence is a significant determinant of vaccine uptake in many populations [27] . An imitation vaccination model was previously developed under the assumption of a homogeneously mixed population [25] , [28] . This model predicts that imitation is likely to generate oscillations in vaccine uptake , and that the oscillations tend to be large when the perceived risk of vaccination is high [25] . Since this model assumes that the population is homogeneously mixed , it cannot capture clustering of vaccination behaviors in a social network . The clustering of vaccination opinions can exacerbate disease outbreaks by interfering with herd immunity [29] , [30] , [31] . To evaluate the effect of imitation dynamics on vaccination and disease outbreaks , we develop social network models with imitation behavior . We consider three different contact network structures , a contact network based on a prior study of contact patterns within Vancouver [9] , a relative homogeneous network with a Poisson degree distribution , and a heterogeneous scale free network ( with a power law degree distribution ) . We assume that a portion of the population adopts vaccination based on a “payoff maximization” strategy that maximizes their perceived payoff , and the remaining population imitates the vaccination choices of their neighbors . For all three networks considered , we find that imitation behavior increases the equilibrium level of vaccination coverage when vaccines are inexpensive and decreases vaccination coverage when vaccines are expensive . However , when imitation increases vaccination coverage , it simultaneously leads to connected clusters of unvaccinated individuals , which increase disease prevalence . The emergence of susceptible clusters and its detrimental epidemiological effects are most prominent when vaccination coverage is close to the herd immunity threshold .
We consider a social contact network where individuals can switch between decisions of vaccinating or not vaccinating . An individual's vaccination decision is a function of both the strategies their neighbors have adopted and the perceived benefits of vaccination . Individuals know only the vaccination opinions of immediate neighbors ( i . e . , whether they are in favor of or opposed to vaccinating ) , and update their strategies either by imitating one of their neighbors ( i . e . , following their opinion ) or by maximizing their perceived benefits . The fraction of individuals with imitation behavior is indicated by , and the remaining individuals ( ) follow a payoff maximization strategy . The population opinion configuration is denoted by , where indicates the vaccination opinion of individual i , ( 1 ) Let be the perceived payoff of an individual i with opinion , then ( 2 ) ( 3 ) where CV is the individual's cost of vaccinating , CI is the individual's cost of infection , and is perceived probability of infection . The payoffs are negative because maximizing a payoff in this context means minimizing a negative health cost/impact . Let be the relative cost of vaccination ( cost of vaccination/cost of infection ) , . Without loss of generality , we can rewrite Eqs . 2 and 3 as: ( 4 ) ( 5 ) Let be the perceived probability of contracting the disease from an infectious neighbor at a given time step and be the number of non-vaccinators in the neighborhood of i , respectively . We assumed that the perceived probability of infection depends on the number of non-vaccinator neighbors ( whose status as non-vaccinators is assumed known ) , with no correlation between degree and the number of non-vaccinator neighbors . From basic probability theory we can express as ( 6 ) where denotes the perceived probability that an unvaccinated neighbor will not become infected . A payoff maximizer i will vaccinate if and will not vaccinate if . When , an individual i will adopt the vaccinator or non-vaccinator strategy with equal probability ( 50% ) . If the entire population adopts payoff maximization strategy , the system is expected to settle on the Nash equilibrium at steady state . An imitator i randomly chooses a neighbor ( ‘role model’ ) j to imitate . Imitator i adopts j's vaccination decision according to predetermined rules: We compare vaccination dynamics across three different classes of networks: a pseudo-empirical urban network based on contact patterns within Vancouver , Canada [9] , a homogeneous random network with a Poisson degree distribution , parameterized so that the average degree is equal to that of the urban network , and a highly heterogeneous , power law network in which degrees follow a truncated power law distribution . Let denote the probability that a randomly selected individual in a network has degree k . The Poisson network is given by with mean contact number ; the power law network is given by with mean contact number of 4 . 5 ( Figure 2 ) . We calibrate epidemic parameters to ensure that infection risk in an unvaccinated population is equal across all network structures [33] , [34] . More precisely , we calibrate the value of disease transmission probability to ensure that the average final epidemic size is equal across the population structures . We chose the final size to be equal to 90% , although results were found to be qualitatively robust for a range of final sizes . For each network , the population size N was equal to 5000 . The contact networks are generated using the configuration model ( CM ) algorithm for constructing finite random networks with a specified degree sequence [35] , [36] . We generated degree sequences by choosing random deviates from these degree distributions . To investigate the effect of imitation behavior on vaccination and disease outbreaks , we assumed that the perceived transmission probability is equal to the transmission probability of the infectious disease . We perform Monte Carlo simulations on vaccinating opinion formation and disease transmission according to these four steps:
We found that imitation ( ) tends to increase vaccination coverage when the cost of vaccination ( r ) is low and to decrease vaccination coverage when r is high ( Figure 3 ) . However , the effect of imitation varies with the degree of distribution of the contact network ( Figure 3 ) . Comparing two extreme cases of ( : fully payoff maximization and : fully imitation ) , imitation dynamics ( ) can promote near-universal coverage when the cost of vaccination is very low compared to that of infection ( small values of r ) ( Figure 3 ) . This difference is particularly pronounced in the power law network . Individuals have a high incentive to vaccinate when the relative cost of vaccination ( r ) is low , and not to vaccinate when it is high . Moreover , when most of an individual's neighbors adopt a given strategy , an individual has more incentive to adopt the opposite strategy . That is , if an individual is surrounded by vaccinators , their risk of infection and resulting incentive to vaccinate will both be low; if an individual is surrounded by non-vaccinators , their risk of infection and incentive to vaccinate will be high . However , imitators have a non-zero probability of copying the vaccination strategy that is adopted by most of their neighbors , even when such a strategy may be less suboptimal for them . For low values of r , payoff maximizers have a high incentive to vaccinate , and thus imitators are likely to have vaccinators as role models; the opposite should be true under high values of r . Therefore , for low values of r , imitators may have higher vaccination coverage than payoff maximizers , whereas for high values of r , imitation may lead to fewer vaccinators than anticipated by payoff maximization strategy ( Figures 3 and 4 ) . The power law network was shown to be more sensitive to the effect of imitation behavior than Poisson and Urban networks ( Figures 3 and 4 ) . This is due to the fact that the power law network has a highly skewed degree distribution , with a small density of highly-connected individuals . Highly-connected individuals ( hubs ) have a high incentive to vaccinate , whereas individuals with few contacts have less incentive to vaccinate . By imitating their highly-connected neighbors , individuals with few contacts become more likely to vaccinate , which may substantially increase vaccination coverage ( Figures 3 and 5 ) . However , this increase of vaccination coverage overall decreases the incentive for hubs to vaccinate ( Figure 5 ) . Depending on the density of hubs and the value of the relative cost of vaccination , this decrease in the incentive of hubs to vaccinate may reduce the total vaccination coverage within the population ( Figure 3 ) . We found that imitation ( ) increases the final size of the outbreak ( i . e . , the fraction of the population infected ) for intermediate costs of vaccination ( r ) ( Figure 6 ) . Numerical investigation showed that this range of values , which varies with the contact network ( Figure 6 ) , has an upper bound that represents the value of r above which it is disadvantageous for anyone to vaccinate , resulting in a full blown epidemic , and a lower bound which represents the value of r below which the average final epidemic size was less than twice the size of the initial inoculum of 10 infected individuals . Imitation dynamics can increase the vaccination coverage relative to a population with payoff maximization strategy , when the cost of vaccination is low , but can never decrease the final epidemic size ( Figures 3 and 6 ) . As a result of behavioral clustering that emerges from imitation dynamics , the size of the epidemic does not necessarily decrease as vaccination coverage increases . That is , vaccinators tend to contact vaccinators , and non-vaccinators tend to contact non-vaccinators ( Figure 7 ) . Because herd immunity is considerably high in these pockets of vaccinators , further vaccination within these pockets reduces transmission to a lesser degree than if vaccination were increased in regions of the network with relatively low vaccination coverage . The clusters of non-vaccinators fuel transmission and increase the probability of an outbreak . This effect of imitation is most prominent when vaccination coverage is close to the herd immunity threshold ( Figures 3 and 6 ) . To investigate the sensitivity of our results to the degree to which imitators respond to payoff differences between themselves and their neighbors , we compared weak responsiveness to strong responsiveness ( Figure 2 ) . For strong responsiveness , individuals reliably copy the strategy of successful neighbors . However , if most neighbors of an imitator adopt a given strategy , then the opposite strategy becomes advantageous , and the imitator would be more likely to choose the opposite strategy . Strong responsiveness , relative to weak responsiveness , leads imitators to rarely copy unsuccessful neighbors ( Figure 2 ) . Therefore , as the degree of responsiveness increases ( α = 15 ) , vaccination coverage under pure imitation ( ) tends to converge towards the vaccination level predicted by the payoff maximization equilibrium ( Figures 3 and 8 ) . A similar convergence occurs for final epidemic size ( result not shown here ) .
Classic economic theory has not considered the reality that individuals frequently imitate others [2] , [3] , [24] , [31] , [37] . Imitation begins with simple behaviors in infancy and evolves into more complex behaviors in childhood and adulthood [31] , [37] . In the context of epidemiology , imitation behavior can influence vaccination patterns and thus the dynamics of disease outbreaks [2] , [3] , [24] . In this work , we address the impact of imitation on vaccination coverage , disease prevalence , and the herd immunity threshold . We develop a model that allows contact patterns to be heterogeneous and individuals to incorporate varying degrees of imitation into decision-making . Individuals within a social contact network can switch between the strategies of vaccinating and not vaccinating . An individual's decision regarding whether to vaccinate is affected by the strategies that their neighbors have adopted or the perceived net benefits of vaccination . Monte Carlo simulations show that imitation dynamics increase the equilibrium vaccination coverage when vaccination cost is relatively low and may decrease vaccination coverage when vaccination is costly . In both cases , imitation actually exacerbates disease transmission when vaccination is inexpensive through the social clustering of non-vaccinators . The detrimental effects of imitation are most prominent when the vaccination is close to the herd immunity threshold . Salathe and Bonhoeffer recently developed a vaccination opinion formation model to reveal that opinion clustering increases the size of an epidemic [29] . Their model assumed that opinions are determined by the proportion of neighbors that have the same opinion about vaccination , such that whenever an individual switches opinion , another individual has to switch the opinion in an opposite way in order to maintain constant vaccination coverage level [29] . Extending this previous seminal model , we consider both imitation ( opinion formation ) and payoff maximization consideration ( individuals are not just blindly imitating neighbors; they are trying to optimize a payoff function ) . Our model thereby recognizes that vaccine decision-making is not a purely imitative process , and often depends on actual health considerations . Additionally , by incorporating payoffs , we are able to analyze the impact of vaccine cost on the dynamics of vaccination . This analysis takes an initial step towards understanding the combined impacts of payoff maximization and imitative decision-making on vaccination coverage and epidemiological dynamics . The model , however , rests on several simplifying assumptions . For example , the contact networks are assumed to remain static throughout the epidemic , and to be identical for both disease and behavioral transmission . These assumptions could be relaxed by incorporating temporal changes in network structure [38] , and modeling multiple different edge types ( e . g . individual variation in susceptibility and infectivity ) [39] . The model can also be extended to allow individuals to follow mixed vaccination strategies , or by incorporating the effects of past epidemics on vaccine decision-making [40] . We find that imitation leads to clustering of susceptible individuals , which may exacerbate outbreaks of infectious diseases . For example , imitation may explain how outbreaks of measles have occurred in countries with high overall vaccination coverage [26] , [29] , [41] , [42] . Given that vaccine decisions are likely to be influenced by social contacts [29] , [32] , [41] and that such imitation can have detrimental epidemiological effects [29] , it is important that policy makers understand its causes , magnitude , and implications for disease elimination . Our findings indicate that the common assumptions of simple payoff maximization and homogeneous mixing can lead to misestimates of the level of vaccination coverage necessary to control a disease outbreak . Our model provides a general framework for investigating the effect of imitation on vaccination decision-making and disease outbreaks . The model can be applied to study the interactions between behavior , public health , and epidemic dynamics for specific infectious diseases . Data describing real world imitation behavior in vaccination decision-making will be critical to future public health applications of the model .
|
Both infectious diseases and behavioral traits can spread via social contacts . Using network-based mathematical models , our study addresses the interplay between these two processes , as disease spreads through a population and individuals copy their social contacts when making vaccination decisions . Imitation can produce clusters of non-vaccinating , susceptible individuals that facilitate relatively large outbreaks of infectious diseases despite high overall vaccination coverage . This may explain , for example , recent measles outbreaks observed in many countries with universal measles vaccination policies . Given that vaccine decisions are likely to be influenced by social contacts and that such imitation can have detrimental epidemiological effects , it is important that policy makers understand its causes , magnitude and implications for disease eradication .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"game",
"theory",
"medicine",
"public",
"health",
"and",
"epidemiology",
"mathematics",
"epidemiology",
"infectious",
"disease",
"epidemiology",
"applied",
"mathematics",
"biology",
"computational",
"biology",
"public",
"health"
] |
2012
|
The Impact of Imitation on Vaccination Behavior in Social Contact Networks
|
Hybrid incompatibilities ( HIs ) cause reproductive isolation between species and thus contribute to speciation . Several HI genes encode adaptively evolving proteins that localize to or interact with heterochromatin , suggesting that HIs may result from co-evolution with rapidly evolving heterochromatic DNA . Little is known , however , about the intraspecific function of these HI genes , the specific sequences they interact with , or the evolutionary forces that drive their divergence . The genes Hmr and Lhr genetically interact to cause hybrid lethality between Drosophila melanogaster and D . simulans , yet mutations in both genes are viable . Here , we report that Hmr and Lhr encode proteins that form a heterochromatic complex with Heterochromatin Protein 1 ( HP1a ) . Using RNA-Seq analyses we discovered that Hmr and Lhr are required to repress transcripts from satellite DNAs and many families of transposable elements ( TEs ) . By comparing Hmr and Lhr function between D . melanogaster and D . simulans we identify several satellite DNAs and TEs that are differentially regulated between the species . Hmr and Lhr mutations also cause massive overexpression of telomeric TEs and significant telomere lengthening . Hmr and Lhr therefore regulate three types of heterochromatic sequences that are responsible for the significant differences in genome size and structure between D . melanogaster and D . simulans and have high potential to cause genetic conflicts with host fitness . We further find that many TEs are overexpressed in hybrids but that those specifically mis-expressed in lethal hybrids do not closely correlate with Hmr function . Our results therefore argue that adaptive divergence of heterochromatin proteins in response to repetitive DNAs is an important underlying force driving the evolution of hybrid incompatibility genes , but that hybrid lethality likely results from novel epistatic genetic interactions that are distinct to the hybrid background .
As populations diverge , their ability to reproduce with each other diminishes . Hybrid incompatibility ( HI ) , the reduced viability and fertility of interspecific hybrids , is a major cause of reproductive isolation between nascent species and thus an important contributor to speciation . Many of the genes causing HI show evidence of adaptive evolution , typically manifest as excessive numbers of amino-acid-changing mutations compared to neutral expectations [1] , [2] . These data do not , however , imply that natural selection acts directly on HI phenotypes . Rather , the prevailing model of HI formulated by Dobzhansky and Muller ( D-M ) emphasizes that incompatibilities evolve in two distinct steps . First , two or more loci diverge independently in two nascent species . Then , if these species later interbreed , these diverged genes may interact to cause deleterious HI phenotypes . The key insight of the D-M model is that hybrid lethality and sterility evolve as byproducts of intraspecific divergence [1] . Adaptive evolution therefore does ultimately lead to HI , but if we wish to identify the evolutionary forces that drive the divergence of HI genes , then we need to understand the function of these genes within species . The mechanisms by which HI genes cause sterility or lethality are important but separate issues . In fact , it remains uncertain whether the wild type functions of HI genes are generally predictive of the deleterious phenotypes that they cause within hybrids . Pinpointing the function of HI genes and the causes of their adaptive evolution is a challenging goal . For example , the Hybrid male rescue ( Hmr ) gene causes large reductions in hybrid fitness [3] . Loss-of-function mutations in D . melanogaster , however , have only moderate effects on fertility and provide few insights into mechanistic underpinnings [4] . The nucleoporins provide an intriguing counterexample . Several have been implicated in hybrid lethality and found to evolve under adaptive evolution [5] . Mutations in nucleoporin subunits are lethal in D . melanogaster , but the genes have many pleiotropic functions and the challenge is to pinpoint which one ( s ) are driving evolutionary divergence . Here we investigate two hybrid lethality genes , Lethal hybrid rescue ( Lhr ) and Hmr , which interact to cause F1 hybrid male lethality between D . melanogaster and D . simulans [6] . Both genes show extensive divergence in their coding sequences that is consistent with positive selection [6] , [7] . For Hmr this sequence divergence appears to be required for hybrid lethality because the D . melanogaster ortholog of Hmr causes hybrid lethality but the D . simulans ortholog does not [7] . For Lhr , however , both orthologs have hybrid lethal activity , with D . simulans Lhr having greater activity due to its higher expression level in hybrids [8] . That study left open the possibility that Lhr coding sequence divergence makes some contribution to hybrid lethality . Furthermore we found that Lhr from the more diverged species D . virilis has no hybrid lethal activity , suggesting that more extensive coding sequence divergence does have substantial functional consequences [9] . These previous studies leave unanswered the fundamental question of what evolutionary force is driving adaptive sequence change , and necessitate a detailed understanding of Hmr and Lhr function within each of the hybridizing species . Loss of function alleles of Hmr and Lhr are strong suppressors of hybrid lethality , but are largely viable within D . melanogaster and D . simulans , respectively [10] , [11] . Lhr ( also known as HP3 ) protein localizes to heterochromatin [6] , [12] . Several other Drosophila HIs also involve heterochromatin or heterochromatin proteins , which is intriguing because genome size varies widely among Drosophila , largely as a consequence of variation in repetitive DNAs that make up the heterochromatin [13] , [14] . Heterochromatin may have a much wider role in incompatibility because repetitive DNA variation is the major cause of the ∼1000-fold variation in genome size among multi-cellular eukaryotes [15] . These DNAs can increase in copy number by general host processes such as unequal crossing over and duplication [16] . Alternatively , they may increase copy number by selfish properties such as transposition for TEs [17] and meiotic drive for satellite DNAs [18] . In either case , over-proliferation can be deleterious to their host species by causing genome instability , leading to the evolution of host defense mechanisms [19] . For example , one major mechanism is the piRNA pathway , where small ( 23–30 nt ) RNAs derived from TE sequences are used to silence TE activity [20] . There are also hints that the piRNA pathway may regulate satellite DNAs [21] . Interestingly , piRNA regulatory genes often show signatures of adaptive evolution among Drosophila species [22] . Genetic conflicts with selfish DNAs have been proposed as an important driver of HI [1] , [2] , [23] , but little is known about what specific sequences are interacting with HI genes . D . simulans and D . melanogaster have great potential for addressing this question because they differ substantially from each other in genome size [14] , satellite DNA content [13] , [14] , and in both the types and number of TEs that they harbor [24] . Here we report that Hmr and Lhr are required to repress transcription from both TEs and satellite DNAs . Hmr and Lhr also regulate telomeres , a third specialized type of heterochromatic sequence that serves to protect the ends of linear chromosomes [25] and is composed of rapidly evolving DNA and proteins [26]–[28] . Telomere variation can affect host fitness and genome stability , and has been proposed as another potential source of meiotic drive [27] , [29] . We used a D . simulans mutation in Lhr , comparative cytology , and interspecific complementation with Hmr transgenes to identify classes of TEs and satellites that are regulated differentially between the species . We conclude that Hmr and Lhr provide an adaptive defense against multiple classes of repetitive DNA sequences that change rapidly in evolutionary time , can reduce host fitness , and have high potential to provoke genetic conflict .
Lhr protein localizes to a subdomain of pericentric heterochromatin in early embryos [8] . To explore possible similarities with Hmr , we examined the localization of Hmr with a 3X-HA epitope-tagged Hmr transgene ( see Materials and Methods ) . mel-Hmr-HA colocalizes with HP1a and H3K9me2 at heterochromatin in nuclear cycle 14 embryos ( Figure 1A ) . We then used Immuno-FISH to determine its localization relative to specific heterochromatic satellite DNA sequences . mel-Hmr-HA does not overlap with the X-linked 359-bp satellite but colocalizes with dodeca , a GC-rich pericentromeric satellite on chromosome 3 . This pattern mimics that seen previously with Lhr [8] . Additionally , mel-Hmr-HA colocalizes with GA-rich repeats and the 2L3L satellite in embryos ( Figure 1B ) . Colocalization between mel-Hmr-HA with both dodeca and GA-rich repeats is also observed in ovarian nurse cells from Hmr3; mel-Hmr-HA females , indicating that localization is not a consequence of overexpression ( Figures S1B , C ) . Unlike Lhr [8] , mel-Hmr-HA localizes to the nucleolus in early embryos ( Figure 1C ) , suggesting that Hmr may have some functions distinct from Lhr . The largely similar localization patterns of Hmr and Lhr raise the possibility that they physically interact . We performed co-immunoprecipitation ( co-IP ) studies from embryo extracts and found that mel-Lhr-HA and mel-Hmr-FLAG co-IP ( Figure 1D ) . mel-Lhr-HA was previously shown to express at wild type levels [8] , and mel-Hmr-FLAG is expressed significantly lower than wild type levels ( Figure S2 ) , demonstrating that these results are not due to overexpression . Lhr was previously shown to bind to , co-localize with , and be dependent on HP1a for correct heterochromatic localization [6] , [9] , [12] , [30] . We therefore tested if HP1a also associates with Hmr . IPs with HP1a pulled down mel-Lhr-HA and mel-Hmr-FLAG , but the reciprocal IPs failed to pull down detectable HP1a ( Figure 1E ) . Yeast two-hybrid assays show that Hmr and Lhr from D . melanogaster interact , suggesting that the co-IP reflects a direct interaction between the proteins ( Figure 1F ) . This interaction is likely mediated via the BESS domains within Lhr and Hmr [6] , a 40 amino-acid motif found in 19 proteins in D . melanogaster that has been implicated in protein-protein interactions and homo-oligomerization [31] . We also found that the D . simulans orthologs interact , as do the heterospecific combinations; the strength of interactions varied widely but exploring the potential significance of this result will require a more quantitative assay . We next examined protein localization in mutant backgrounds to test the potential mutual dependence of Lhr and Hmr for their localization to heterochromatin . We made a D . melanogaster Lhr mutation by recombining a mini-white gene into the Lhr locus to create the LhrKO allele ( Figure S3A ) . In LhrKO , transcription from Lhr but not flanking genes is greatly reduced , and no Lhr protein is detectable ( Figure S3B , C ) . These results demonstrate that LhrKO is a strong loss of function allele , which we confirmed in hybrid rescue crosses ( see Materials and Methods ) . Lhr-HA levels are greatly reduced in Hmr3 mutant embryos but when examined at high gain a small amount of Lhr-HA is detectable in heterochromatin ( Figure 1G ) . This result suggests that Hmr is not absolutely required to localize Lhr to heterochromatin , though it remains possible that some Hmr protein is made in the Hmr3 mutant . In a reciprocal experiment , Hmr-HA localization appears normal in LhrKO ( Figure 1H ) . In combination with previous results , our data suggest that Lhr localization to heterochromatin depends on HP1a , and that Hmr stabilizes Lhr . LhrKO flies are almost fully viable ( 22 . 25% compared to the expected 25% in crosses between heterozygotes at 27°; p<0 . 05 by Chi-squared; N = 2813 total flies scored ) . However , comparison of LhrKO with a background-matched Lhr+ control ( see Materials and Methods ) showed that LhrKO females have substantially lower fertility , particularly at higher temperatures . One to five day old LhrKO females display only a fraction of the fertility of LhrKO/+ and later become sterile ( Figure 2A ) . We confirmed this in a different Lhr− background where a similar reduction in fertility occurs at later ages ( Figure 2B ) . In a separate experiment we found that the hatch rate of the eggs laid by LhrKO/LhrKO mothers is low and declines with increasing maternal age ( Table S1 ) . This LhrKO female fertility phenotype is strikingly similar to that of Hmr mutants [4] , suggesting that Hmr and Lhr may function in a common regulatory pathway . We performed an RNA-Seq comparison of ovaries from LhrKO and Lhr+ to investigate the cause of this fertility reduction and discovered a widespread increase in transposable element ( TE ) transcripts . Using two different TE mapping methods ( see Materials and Methods ) we found that transcripts from 99 families were at least 2-fold upregulated , with 38 elements being at least 10 fold upregulated ( Figure 3A; Table S2 ) . Mis-regulated TEs include elements with germline expression such as the telomeric non-LTR retrotransposons HeT-A ( 350 . 7 fold ) and TART ( 51 . 76 fold ) , the LTR retrotransposon copia ( 19 . 8 fold ) , and the DNA transposon bari-1 ( 44 . 7 fold ) . TEs expressed only in the somatic follicle cells , such as Gypsy ( 3 . 8 fold ) and Zam ( 7 fold ) were also upregulated . In addition , qRT-PCR in two different genetic backgrounds confirmed the massive increase in HeT-A transcript levels ( 185–846-fold; Figure S4 ) . These results demonstrate that the telomeric TEs are especially sensitive to Lhr regulation . We also performed RNA-Seq analysis of an Hmr mutant ( Df ( 1 ) Hmr−/Hmr3 , abbreviated below as Hmr− ) . We compared it to a heterozygous control ( Df ( 1 ) Hmr−/y w Hmr+ , abbreviated below as Hmr−/Hmr+ ) because it closely matches the genetic background of the mutant genotype , and also serves as a control for Hmr transgenic genotypes that are described below . We found that 55 different TE families are upregulated at least 2 fold in Hmr mutants , with 14 being upregulated at least 10 fold ( Figure 3B; Table S3 ) . Notably , the telomeric retrotransposons HeT-A and TART are again among the most highly upregulated . Strikingly , the TEs affected by Hmr are largely a subset of Lhr-regulated TEs , suggesting that they act together to regulate multiple TE families ( Figure 3C ) . The smaller number of mis-regulated families in Hmr− likely reflects the fact that we are comparing Hmr− mutants to heterozygotes , but Lhr mutants to wild type . Since some germline TE repressor genes also regulate somatic TE expression [32] , we performed RNA-Seq to compare TE expression between 72–76 hour-old Df ( 1 ) Hmr−/Y and Hmr+/Y D . melanogaster male larvae . This also served as a control for experiments described below to address whether TE mis-expression may be contributing to hybrid lethality . We found that 31 TEs exhibit a statistically significant ≥2 fold upregulation ( Figure 3D; Table S4 ) , but there are two striking differences compared to Hmr mutant ovaries . First , different TEs are affected , with the telomeric retrotransposons in particular not upregulated in the larvae . Second , the magnitude of TE derepression is lower in larvae . We next examined potential effects on protein-coding genes . Remarkably few genes ( 11 in Hmr−; 0 in LhrKO ) show a statistically significant misregulation in either Lhr or Hmr mutants ( FDR 0 . 05; Tables S5 , S6 ) . However , a comparison of fold change in the expression of all heterochromatic versus all euchromatic genes found that heterochromatic genes are downregulated to a greater extent for both mutants , although the effect is stronger in LhrKO ( Figure 4 ) . Lhr preferentially associates with heterochromatic genes in an embryonic cell culture line [12]; our results suggest that Lhr and Hmr have a small positive effect on expression of some heterochromatic genes . Drosophilidae have lost the telomerase-based mechanism of telomere elongation and instead use the regulated transposition of the HeT-A , TART and TAHRE retrotransposons [33] . Strikingly , these were among the 3 most strongly affected TEs in LhrKO and Hmr− ovaries ( Figure 3 ) . We therefore investigated in more detail the localization of Lhr and Hmr proteins to the telomere [6] . Cytological markers on polytene chromosomes have been used to describe three distinct regions in the telomere , with HP1a localizing exclusively to the “cap” , a proteinacous structure at the most distal end of telomeres [25] , [28] . mel-Lhr-HA and mel-Hmr-HA overlap with HP1a , showing that Lhr and Hmr localize to the cap but not to more proximal regions ( Figure 5A , B ) . Localization is not due to the doubling of the dosage of these proteins in the transgenic lines because it also occurs in the Hmr3; Hmr-HA/Hmr-HA and LhrKO/+; Lhr-HA/+ genotypes ( Figure S5 ) . The localization of Lhr and Hmr to the cap , the primacy of the cap in the regulation of telomeric length , and the increase in the transcript levels of telomeric retro-transposons in Lhr and Hmr mutants led us to ask if these mutations cause long telomeres . We quantitated HeT-A DNA copy number by qPCR in LhrKO flies maintained at 27°C separately from its matched wild-type control strain for ∼40 generations . We found that HeT-A copy number increased approximately 6 fold in LhrKO ( Figure 5C ) . We also examined HeT-A DNA copy number in an Hmr3 mutant stock , and found ∼4–16 fold higher abundance than in the Hmr+ stocks y w and Canton-S ( Figure 5D ) . Hmr and Lhr both localize to pericentric heterochromatin , which is largely composed of TEs and satellite DNAs . The potential effects of heterochromatin proteins on the levels of transcripts from satellites have not been widely explored . We therefore used our RNA-Seq data to examine transcript levels from 143 repeats in a repeat-sequence database ( see Materials and Methods ) . Transcripts from most repeats are found at low abundance in Lhr+ with only 17 producing more than 10 reads ( Table S7 ) . Four different satellite classes are significantly higher in LhrKO versus Lhr+ ovaries , including three that collectively make up more than 8% of the D . melanogaster genome [13]: AAGAC , AACAC , and the GA-rich satellites ( Figure 6a ) . The GAGAA satellite showed the strongest effect , with an approximately 30-fold increase . These results raise the question of whether transcriptional regulation of specific satellite DNAs reflects a direct association with Lhr . Lhr was not previously tested for association with either GA-rich satellites , which are found on all chromosomes in D . melanogaster [34] , or with the AACAC satellite found on chromosomes 2 and Y [35] . We found that Lhr-HA colocalizes extensively with the GA-rich and AACAC satellites in the nurse cell nuclei of early stage egg chambers ( Figure 6B , S1A ) . In our Hmr RNA-Seq data the number of reads mapping to each repeat family was generally very small , but 3 satellite families are significantly derepressed by at least 4 fold in Hmr− ( Figure 6C; Table S8 ) , including GAGAA , which has a 19 fold increase in expression . This finding is consistent with the localization of mel-Hmr-HA to GA-rich satellites above ( Figure 1B ) . Additionally , the satellite Z37541 , which binds nuclear lamins , is upregulated 5 fold in Hmr− [36] . Although Lhr-HA localizes to the dodeca satellite [8]; we detected very few reads in either our Lhr+ or LhrKO samples; likewise we did not find upregulation of dodeca in our Hmr RNA-Seq data . We conclude that Hmr and Lhr proteins are required to regulate transcript levels of a subset of satellites to which they localize . The wide spectrum of TEs derepressed in Lhr and Hmr mutants is similar to mutations in piRNA regulatory genes such as Ago3 and aub that post-transcriptionally regulate TEs via small-RNA-mediated silencing [37] , [38] . We therefore investigated a range of phenotypes that are associated with defects in the piRNA pathway . Ago3 and aub mutants disrupt Vasa localization to the peri-nuclear small-RNA processing center , the nuage , and exhibit drastic reductions in the piRNA fraction ( 23–30 nt ) [38] , [39] . We found , however , that Vasa localizes normally in LhrKO ( Figure 7A ) . We then sequenced the small RNA pool in LhrKO and found that the piRNA level is broadly comparable to Lhr+ with only a minor reduction in longer piRNAs ( Figure 7B ) . This pattern contrasts with mutants such as aub and spn-E that show a severe loss of piRNAs [39] . We looked more closely for TE-specific defects and found that piRNAs mapping to most individual TE families are comparable between Lhr+ and LhrKO ( Figure 7C; Table S9 ) . We also examined “ping-pong” processing , which produces piRNAs from opposing strands with a characteristic 10 nucleotide overlap [38] , [39] . Ping-pong scores are generally higher in Lhr+ ( Figure 7D; Table S10 ) but several points argue against there being a significant defect in ping-pong or piRNA processing in LhrKO . First , the magnitude of the difference between genotypes is low , with the ping-pong score being > = 2-fold higher in Lhr+ for only 26/140 TEs . Furthermore , half of these 26 have ping-pong scores <0 . 10 in Lhr+ ( Table S10 ) , suggesting that those TE families are not significantly processed by ping-pong in wild type flies . Second , these differences in ping-pong scores between Lhr+ and LhrKO are much milder compared to mutations in genes such as spn-E [39] . Third , many of the TEs showing differences in ping-pong scores are not strongly depressed in LhrKO . Conversely , many TEs that are strongly derepressed in LhrKO , including HeT-A , have ping-pong scores that are comparable to wild-type . Fourth , some TEs with elevated mRNA levels also show increased ping-pong signatures , probably because of increased processing through a functional ping-pong pathway . We suggest therefore that the moderate trend towards reduced ping-pong scores in LhrKO does not reflect a failure in the ping-pong cycle . Instead , it may result from a skew in the ratio of sense∶antisense piRNAs , because LhrKO flies have high levels of TE transcripts that can be processed into sense piRNAs . An analogous argument has been made for mutations in the Drosophila Gtsf1/asterix gene , which derepress TEs and give an altered ratio of sense and antisense piRNAs but appear to do so downstream of piRNA biogenesis [40] . We searched further for possible defects in piRNA production by examining piRNAs that map to 122 primary-piRNA-generating heterochromatic clusters [41] . piRNAs originating from most of the major clusters are not significantly affected in LhrKO but 16 and 11 of the 122 clusters are at least two-fold higher or lower , respectively , in LhrKO ( Figure 7E; Table S11 ) . Some of the most strongly affected clusters are associated with telomeres . Cluster 3 consists entirely of telomeric retrotransposons and is upregulated 4 . 3 fold in LhrKO . Sub-telomeric cluster 11 shows a complete loss of unique piRNAs , while clusters 33 and 4 are 2 . 6 and 2 . 9 fold downregulated , respectively . These 3 clusters consist mainly of HETRP telomere-associated ( TAS ) repeats and are therefore not expected to contribute to TE repression; their misregulation instead suggests that Lhr is required for regulating chromatin states at telomeres . The siRNA pathway has also been implicated in repressing TEs in the ovary [42]–[44] . We found that siRNAs mapping to the vast majority of TE families , including those mapping to HeT-A , are not significantly different between LhrKO and Lhr+ , suggesting that Lhr is not generally required for siRNA biogenesis ( Figure 7F; Table S12 ) . Taken together , our results indicate that defects in small RNA synthesis are not the cause of TE derepression in LhrKO . An intriguing possibility is that Lhr is a piRNA-dependent effector of TE silencing . We propose that the dynamic sequence turnover of repetitive DNAs is the selective pressure driving the adaptive sequence divergence of Lhr and Hmr . This hypothesis implies that the localization and/or function of the Lhr protein have changed between species , due to co-evolution with species-specific repetitive DNAs . The Lhr1 allele in D . simulans [10] presents a rare opportunity to compare the function of a rapidly evolving heterochromatin protein between sibling species . We performed RNA-Seq from ovaries of Lhr1 females and a matched Lhr+ control ( see Materials and Methods ) . We found essentially no Lhr transcript reads in the Lhr1 mutant strain ( Table S13 ) , strongly suggesting that this allele is null . D . simulans has many of the same satellites as D . melanogaster but they are generally of lower abundance [13] . We therefore first examined satellite DNA expression in the Lhr1 and Lhr+ ( control ) RNA-Seq data . Unlike in D . melanogaster LhrKO , we found few satellite reads in either genotype and no significant differences between them . We conclude that Lhr has a unique role in D . melanogaster to repress satellite DNA transcription . The AACAC satellite that Lhr co-localizes with in D . melanogaster ( Figure 6B ) is absent in D . simulans [35] . The GAGAA satellite is also drastically different in D . simulans , being eight-fold less abundant and found only on the sex chromosomes [13] , [35] . To determine if this interspecific difference in satellite content reflects divergent localization of Lhr orthologs , we examined D . simulans ovaries expressing a previously characterized sim-Lhr-HA transgene [8] . While Lhr-HA is juxtaposed to dodeca in both species , as previously described [8] , the strongest foci in D . simulans do not overlap with GAGAA ( Figure 8A ) . These results demonstrate that Lhr has evolved distinct localization patterns to at least two satellites between D . melanogaster and D . simulans . We next examined TE expression and discovered a broad spectrum of TEs derepressed in D . simulans Lhr1 , with 80 TEs showing a greater than two-fold up-regulation ( Figure 8B; Table S14 ) . Upregulated TEs again include the telomeric transposable elements HeT-A , TART , and TAHRE , other germline elements such as Nomad , and somatic TEs such as Zam and Gypsy 5 . 53 transposable elements were commonly mis-regulated in both D . melanogaster and D . simulans , showing that the function of Lhr in repressing TEs is broadly conserved between species ( Figure 8C ) . However , the fold increases of most individual TE families are lower than seen in D . melanogaster LhrKO . For example , HeT-A is 352 fold upregulated in LhrKO but only 23 . 8 fold upregulated in Lhr1 . We further discovered that Lhr loss in D . simulans does not significantly affect the expression of heterochromatic genes ( Figure 8D , Table S13 ) , in contrast with our similar analysis of LhrKO in D . melanogaster ( Figure 4A ) . This result suggests that pericentric genes in D . melanogaster are more sensitive to changes in heterochromatin state than in D . simulans . Overall , our results demonstrate that Lhr function correlates with the increased repeat content and larger amount of heterochromatin found in D . melanogaster . To examine the functional consequences of Hmr divergence , we took an alternative approach of transforming sim-Hmr transgenes into D . melanogaster . We found that sim-Hmr-HA , like mel-Hmr-HA , localizes to heterochromatin in D . melanogaster ( Figure 9A ) . To examine potential differences in TE and satellite regulation , we used parallel mel-Hmr-FLAG and sim-Hmr-FLAG transgenes , crossed them into an Hmr− background ( Df ( 1 ) Hmr−/Hmr3 ) , and performed RNA-Seq on ovarian mRNA . Our expectation was that divergence of Hmr between the orthologs might manifest as the failure of sim-Hmr-FLAG to complement the derepression of TEs in Hmr− . As a control for the function of the transgenes , we compared the heterozygous wild type Hmr−/Hmr+ to Hmr−; ø{mel-Hmr-FLAG}/+ , as each genotype has one wild type copy of Hmr+ . The majority of the upregulated TEs in Hmr− ( Figure 3B ) are suppressed by the mel-Hmr-FLAG transgene; however , 9 out of 182 families ranged from 2 to 9 times more highly expressed in Hmr−; ø{mel-Hmr-FLAG}/+ than Hmr−/Hmr+ ( Figure 9B ) . This result suggests that mel-Hmr-FLAG does not fully complement the Hmr mutant phenotype , which may reflect its decreased expression compared to a wild type allele ( Figure S2 ) , though it is also possible that some differences may result from TE polymorphisms that remain between the strains . qRT-PCR also demonstrated that sim-Hmr-FLAG expresses in D . melanogaster at ∼3× the level of mel-Hmr-FLAG ( Figure S2 ) , a difference previously seen with Lhr transgenes [8] . Because Hmr is a negative regulator of TE expression , we suggest that this expression difference will not bias against our goal of identifying TEs that are not fully repressed by sim-Hmr-FLAG . We did not find any difference in satellite DNA expression; however , we found 11 TE families that are differentially expressed between the transgenic genotypes ( Figure 9C ) . Five are more highly expressed in Hmr−; ø{mel-Hmr-FLAG}/+ with fold changes ranging from 2–3 , of which 3 are incompletely repressed by mel-Hmr-FLAG in the control cross described above ( Transpac , Tirant , and Batumi ) . The differential expression of these 5 families likely reflects the inability of mel-Hmr-FLAG to fully complement Hmr− and the higher expression level of sim-Hmr-FLAG . More intriguing are 6 TE families that are 2–6× more highly expressed in Hmr−; ø{sim-Hmr-FLAG}/+ than in Hmr−; ø{mel-Hmr-FLAG}/+ , implying that sim-Hmr-FLAG is unable to fully complement the derepression of these elements . BS and Doc6 ( also known as Juan ) elements are present at a mean frequency of about 0 . 1 in a population of Portuguese D . melanogaster [45] and have low pairwise identity in the reference genome [46] , suggesting that they are likely active . The mean population frequencies of 4 of the other families ( BS3 , Circe , Helena , and FW2 ) are near 1 , suggesting that these TEs are fixed and therefore currently inactive in D . melanogaster . Helena , though , appears to have been active more recently within D . simulans [47] . We suggest that BS , Doc6 and Helena are candidates for future investigation of co-evolution with Hmr in either D . melanogaster or D . simulans . In light of our discovery that Lhr and Hmr are required for TE repression within D . melanogaster and D . simulans , we investigated TE activity in lethal ( Hmr+ ) hybrid male larvae . Because most TEs have different expression levels between D . melanogaster and D . simulans , we defined mis-regulated TEs as being at least two-fold higher than both parental species , as done in a previous analysis [48] . We found that 42 LTR and non-LTR elements are significantly upregulated in lethal ( Hmr+ ) hybrid male larvae with 2 others being downregulated ( Figure 10A; Table S15 ) . We next examined whether TE misregulation correlates with hybrid lethality by comparing the lethal Hmr+ hybrid males to viable Hmr− hybrid males ( Figure 10B , Table S16 ) . The expression of 29 TEs is significantly lower in Hmr− hybrids . Because Hmr functions as a repressor of TEs in D . melanogaster male larvae ( Figure 3C ) , these differences may reflect a general difference between lethal and viable hybrids rather than the presence or absence of Hmr activity . In fact , only 4 of the 29 TEs downregulated in Hmr− hybrid male larvae are upregulated in Hmr− D . melanogaster male larvae ( Table S4 ) . In addition , we found modest increases ( 2–4 fold ) in the activity of 5 TE families in living hybrids . None of these are significantly upregulated in Hmr− D . melanogaster male larvae ( Table S4 ) . They include TAHRE and may reflect higher levels of cell proliferation in viable hybrids . Taken together our results suggest that TE overexpression is unlikely to be causing hybrid lethality .
We and others previously reported that Lhr ( also known as HP3 ) interacts with HP1a [6] , [9] , [12] , [30] . Here we report that Hmr also interacts with Lhr , and both are present in a complex together with HP1a . Consistent with this interaction , many of the roles we report here for Lhr and Hmr have been described for HP1a , including localizing to heterochromatin , regulating TE and pericentric gene expression , and controlling telomere length [49]–[51] . However , unlike mutations in Su ( var ) 205 which enodes HP1a [52] , mutations in Hmr and Lhr are viable . Furthermore , Hmr and Lhr do not localize to the 359 bp satellite which forms a substantial fraction of X-linked pericentric heterochromatin Figure 1; [ref . 8] . These findings suggest that Hmr and Lhr are not ubiquitous heterochromatin proteins , leaving open the intriguing question of what guides their localization specificity . The interaction of Hmr and Lhr with HP1a has recently been independently reported [53; AA Alekseyenko and M . Kuroda , personal communication] . Thomae et al . [53] also report other findings similar to ours here including repressive effects of Hmr and Lhr on TEs in somatic tissues and their localization to telomeres . Several conclusions are similar between the two studies and with previously published conclusions . Thomae et al . [53] observe upregulation of TEs in hybrids but conclude that they are unlikely to be the direct cause of hybrid lethality , a conclusion we reach below using different methods . Their conclusion that hybrids are highly sensitive to Hmr dosage is in concordance with previous studies , such as the previous observation that a ∼9 . 7 kb Hmr+ transgene causes dosage-dependent lethality to hybrid females [3] . This conclusion also fits well with the discovery that hybrids are highly sensitive to Lhr dosage [8] . One area of possible discrepancy is the viability effects and cellular phenotypes associated with Hmr and Lhr mutants versus RNAi knockdown . Thomae et al report a high rate of mitotic defects in Lhr RNAi knockdown tissue culture cells , yet we found that LhrKO flies are almost fully viable ( see Results ) , as are Lhr RNAi knockdown animals [53] . We also have not observed the lethality or morphological defects in Hmr mutants that are reported for Hmr RNAi cells and animals [53] . For example , Aruna et al . [4] found reduced longevity but no effect on viability up to eclosion of flies carrying the Df ( 1 ) Hmr− allele , a deletion of the 5′ end of Hmr . Further work is necessary to determine if these discrepancies reflect phenotypes associated with the use of RNA interference or differences between assaying whole animals versus tissue-culture cells , such as the aneuploid state of cultured cell lines [54] . Several HIs involve heterochromatin proteins or heterochromatic sequences , leading to the suggestion that genetic conflicts between selfish DNAs and host fitness are an important force that is driving the evolution of HI [1] , [2] , [23] , [55] . TE and satellite abundance varies widely among species and is a major contributor to genome-size variation . The evolutionary causes of this variation have been widely debated for many years [56] . When considering genetic conflict theories , it is important to first exclude alternative evolutionary causes of repetitive DNA variation . One explanation is neutrality , with repeat variation governed by mutational processes , in particular the balance between insertions and deletions [57] . Insertion/deletion models are particularly appropriate for inactive and degenerate TEs , and perhaps also for certain classes of satellites that are no longer homogenized by concerted evolution [58] . Selectionist models fit better for active repeats , and must be invoked if the adaptive evolution of heterochromatin proteins is proposed to reflect co-evolution with repetitive DNA . One model is that some repeats are co-opted for host functions . Drosophila's telomeric retrotransposons are a relevant example that is discussed below . We also consider three , non-mutually exclusive selective costs associated with repetitive DNA when discussing the evolution of Hmr and Lhr One potential cost arises from the overall load of repetitive DNAs , including increased genome size and instability . A second is direct genetic conflict . We define genetic conflict here to refer to fitness costs imposed by selfish DNAs that have evolved specific mechanisms to increase their transmission [59] . Such conflicts could be caused by highly active individual repeats , for example during hybrid dysgenesis caused by introduction of a TE family into naive strains [60] . Finally , genetic conflicts can have indirect costs , such as pleiotropic fertility defects caused by repeat expansions involved in meiotic drive [61] . TEs define selfish DNA [56] . They infect most genomes , can self-mobilize and increase their copy number , and destabilize genomes via spontaneous mutations , ectopic recombination , and deleterious increases in genome size [62] , [63] . Adaptive evolution of TE-defense genes can therefore be readily interpreted as the host species responding to the fitness cost of TEs [19] . Like Hmr and Lhr , many piRNA pathway genes are also evolving under positive selection [22] . This raises the possibility that Lhr and Hmr are co-evolving with the piRNA pathway proteins . However , the lack of major perturbations in the piRNA pool in LhrKO suggests that Lhr and Hmr function downstream or independently of piRNA biogenesis . Piwi , guided by piRNA , has been proposed to recruit repressive heterochromatin components including HP1a and histone methyl transferases to transposable elements [51] , [64] . One possibility is that Lhr and Hmr function downstream of HP1a to repress TEs via RNA degradation machinery such as the nuclear exosome [65] . We note that Ago3 is moderately down-regulated in both LhrKO ( 3 . 4 fold ) and Hmr− ( ∼2 fold ) ( Tables S5 , S6 ) , likely because the gene is peri-centromeric . Two results demonstrate that this modest reduction in Ago3 cannot explain the broad effects on TEs in Hmr and Lhr mutants . First , Ago3 expression is unaffected in D . simulans Lhr1 , which also shows widespread TE derepression . Second , Ago3 mutants have major disturbances to their piRNA pool [38] , which we did not observe in LhrKO ( Figure 7 ) . While TE repression is typically viewed in terms of genetic conflicts , the relationship between Lhr , Hmr and the telomeric TEs resembles symbiosis . These TEs have been domesticated by Drosophila species for tens of millions of years to serve a vital host function , and thus are not considered selfish DNA [33] , [66] . The telomeric TEs were among the most strongly derepressed in Hmr and Lhr mutants , in some cases more than 100 fold . We also observed increases in HeT-A DNA copy number in Hmr and Lhr stocks . Increased telomeric TE expression does not necessarily increase HeT-A DNA copy number and cause longer telomeres , suggesting that multiple factors control telomere length [67] . If so , then Lhr and Hmr must control multiple processes at the telomere . This is supported by the localization of both proteins to the telomere cap , a protective structure that prevents telomere fusions [28] . The strong reduction in LhrKO of piRNAs from three TAS-repeat containing sub-telomeric piRNA clusters is particularly intriguing . piRNA production from clusters is dependent on them maintaining a heterochromatic state [68] , which could explain why Lhr is required for TAS piRNA expression while it acts as a repressor in most other circumstances . We discovered several striking examples that suggest species-specific co-evolution of Hmr and Lhr with satellite DNAs . We found that D . melanogaster Hmr and Lhr proteins localize to and repress transcripts from GA-rich satellites . GA-rich satellites are ∼8 fold less abundant in D . simulans [13] but are cytologically detectable; nevertheless we find that sim-Lhr does not localize to them . GA-rich satellites also have low abundance in the outgroup species D . erecta [13] , implying that the differential abundance with D . simulans reflects an increase in D . melanogaster . Similarly we discovered that mel-Lhr-HA localizes to AACAC in D . melanogaster , a repeat that is absent in D . simulans [69] . Furthermore , we detected moderate up-regulation of several other satellite transcripts only in D . melanogaster . Our results suggest that Lhr and Hmr may have evolved in D . melanogaster to mitigate the deleterious consequences of satellite expansion , which can include ectopic recombination , increased genome size , and destabilized chromosome segregation [16] , [70] . Satellite transcripts have been reported from various tissues in wild type D . melanogaster [71] , [72] but little is known about their production . They could be products of either non-specific transcription or read-through from adjacent TEs . Increased levels of satellite transcripts are observed in D . melanogaster spn-E mutants , suggesting that RNA interference or piRNA pathways control satellite transcript levels [21] . We find that at a broad scale , Lhr and Hmr from both D . melanogaster and D . simulans regulate heterochromatic repetitive DNAs but very few genes . This finding is consistent with previous analyses demonstrating that some functions of these genes are conserved between species [4] , [7]–[9] . But many of the repeats regulated by Lhr and Hmr are rapidly evolving , raising the question of whether specific repetitive DNAs are directly driving the adaptive evolution of the Lhr and Hmr coding sequences between species . A simple prediction is that D . simulans orthologs should fail to fully repress such repeats when placed into D . melanogaster , a prediction that we tested for Hmr . The BS non-LTR retrotransposon is significantly derepressed in D . melanogaster Hmr− and LhrKO , and in D . simulans Lhr1 mutants . Interestingly , BS appears to be transpositionally active in D . melanogaster but inactive in D . simulans [73] . One interpretation is that BS was active in the common ancestor and regulated by Hmr and Lhr . The genes would continue to co-evolve with BS in D . melanogaster , making the sim-Hmr ortholog less effective at repressing BS elements in D . melanogaster . In this scenario Hmr and Lhr are engaged in a recurrent genetic conflict with BS elements that leads to their sequence divergence . Consistent with this prediction we found significantly higher expression in Hmr−; ø{sim-Hmr-FLAG}/+ compared to Hmr−; ø{mel-Hmr-FLAG}/+ . Copia shows a different pattern , with ∼20-fold up-regulation in LhrKO but only ∼2-fold in Lhr1 ( and only when mapping to the consensus-sequence database ) , as well as significant derepression in Hmr− . Copia expression level can be high in D . melanogaster but is variable among populations . In contrast , copia elements in D . simulans typically contain deletions in regulatory elements required for expression , and transcripts are undetectable by Northern blot analysis [74] . These results suggest that Hmr and Lhr could be D . melanogaster host factors that defend against a TE that is currently active within the species . However , copia was fully repressed in Hmr−; ø{sim-Hmr-FLAG}/+ , demonstrating that adaptive divergence of Hmr by itself does not affect copia regulation . Overall , we found surprisingly few cases of overexpression associated with Hmr divergence , including no effects on satellite DNAs ( Figure 9 ) . We also note that most of the TEs identified other than BS and Doc6 are likely transpositionally inactive in D . melanogaster [45] , which makes it more challenging to fit a scenario of direct and recurrent evolution between Hmr and specific TEs . We suggest several possible interpretations of these results . One is that Hmr and Lhr adaptive divergence is in fact driven largely or solely by BS and/or Doc6 , a hypothesis that will require understanding the mechanism by which Hmr and Lhr affect expression of these TEs . Second is that Hmr and Lhr may be co-evolving with other genes , and that multiple diverged genes need to be replaced simultaneously in order to detect their effects on other TEs and satellite DNAs . Third is that more sensitive assays are needed , for example monitoring TE transposition rates over multiple generations . A fourth possibility is an alternative to genetic conflict scenarios that arises from population-genetic models . These models suggest that the fitness costs of individual TE families are likely extremely weak under most circumstances . The adaptive evolution of repressor proteins may therefore reflect the cumulative load of repeats within a genome [22] . This alternative view could be applicable to Hmr and Lhr since they repress a large number of TEs and satellites . Finally , Hmr and Lhr may have additional unidentified phenotypes that are also the targets of adaptive evolution . D . simulans has a smaller genome with ∼4-fold less satellite DNA [13] , [14] and significantly fewer TEs [24] , [75] compared to D . melanogaster . This large difference in repeat content between D . melanogaster and D . simulans may have wider consequences . We found reduced expression from pericentric heterochromatin genes in Hmr and Lhr mutants in D . melanogaster . This reduction may reflect the fact that pericentric genes have evolved to use heterochromatin proteins such as Lhr and Hmr to maintain gene expression in a repeat-rich environment [76] . Pericentric genes in species with fewer repeats would presumably not require these proteins . Consistent with this model , we found that Lhr loss in D . simulans has a negligible impact on pericentric gene expression . This finding suggests that Lhr and Hmr have an adaptive role in blocking effects on gene expression arising from increasing repetitive DNA copy number . If each genome is uniquely adapted to its repetitive DNA content , then the shock of hybridization may lead to misregulation of TEs and satellites . TEs are activated in various animal and plant hybrids but the consequences , if any , for hybrid fitness are largely unclear [77] . We found substantial TE misregulation in hybrid male larvae ( Figure 10A ) . Since these hybrids are agametic [78] , this TE expression comes from somatic tissues . The fitness cost of this upregulation is unclear as somatic TE overexpression is not necessarily lethal within D . melanogaster [79] , [80] . Comparison of lethal Hmr+ and viable Hmr− hybrid males demonstrates that lethal hybrids have more TE expression ( Figure 10B ) than the viable hybrids , which in turn have more TE expression than either of its parents . However , this TE misregulation seems unconnected with Hmr as the TEs differentially expressed between Hmr+ and Hmr− hybrid male larvae are largely distinct from those between Hmr+ and Hmr− D . melanogaster male larvae . Further , while Hmr− causes rampant TE over-expression within D . melanogaster , it is associated with reduced TE levels in hybrids . These observations argue that the TE derepression in hybrids is unrelated to the pure species function of Hmr . This finding is consistent with previous genetic studies that demonstrate that the wild type Hmr+ allele causes hybrid lethality and thus behaves as a gain-of-function allele in hybrids [81] , [82] . More generally it underscores the unique nature of the hybrid genetic background [1] . Somatic TE overexpression may result from breakdown in the siRNA or piRNA pathways due to incompatibilities among multiple rapidly evolving TE regulators . One clear example is known where a species-specific difference in a satellite DNA causes incompatibility between Drosophila species [83] . But the toll caused by heterochromatic differences may more commonly be indirect , as heterochromatin proteins diverge in response to changes in heterochromatic DNA repeats . Recent work suggests that hybrid female sterility may be caused by incompatibilities among rapidly evolving piRNA proteins rather than by species-specific differences in TEs [48] . We suggest that the role of Hmr and Lhr in regulating the activity of three highly dynamic classes of heterochromatin has led to their recurrent adaptive evolution , and secondarily , to their involvement in interspecific hybrid lethality .
We used the pW25 donor vector and ends-out homologous recombination method to make an Lhr mutant allele [84] . The donor vector was designed to recombine a w+ marker into Lhr and simultaneously remove 26 bp of the coding region . iProof ( Biorad ) was used to PCR amplify two genomic fragments from y; cn bw sp ( D . melanogaster ) genomic DNA . The 3768 bp Lhr upstream fragment , including 128 bp of the coding region of Lhr , was amplified with primers LUF-Fwd: 5′- ttggcgcgccAACAGGGTCGGCTGTCACATTT and LUF-Rev: 5′-ttggcgcgccGCGAGCATCTCCATGAGCAG ( Tm = 63°C ) and cloned into the AscI site of pW25 using the underlined sequences . The 3935 bp Lhr downstream fragment that includes 806 bp of the Lhr coding region was amplified with primers LDF-Fwd: 5′-AAGCGGCCGCAGGTGGAGCCCAAAATGGACG and LDF-Rev: 5′- AAGCGGCCGCCACACATTGCGAATGCA G AAA ( Tm = 65°C ) and cloned into the NotI site using the underlined sequences . Restriction digestion was used to pick a clone in which the 2 inserts and the mini-white gene were in the same orientation . The construct was injected into a strain of w1118 ( Genetic Services ) and a transgenic line , P{w+ , Lhr-KO}5-1 , with a lethal insertion on the X chromosome was obtained . P{w+ , Lhr-KO}5-1/FM6 females were crossed to y w; P{ry+ , hs-flpase} , P{v+ hs-I-Sce}/TM6 , Ubx males . Two to three day-old larvae were heat shocked and P{w+ , Lhr-KO}5-1/y w P{ry+ , hs-flpase} , P{v+ hs-I-Sce}/+ female progeny were crossed to w1118 males . Rare w+ sons were screened for homologous recombination events by PCR . Primer pairs Lhr-f1 5′- TTCGCACGTTGTGTTCAAGTAA-3′ , /Lhr-r1 5′-GTAGCTTTCTCTTGGCGCTCTT-3′ and Lhr-f2 5′- AACGTGCTCGTAGCTTTGGT-3′/ , Lhr-r2 5′-TCGCGAAAATACTTCCGTCT-3′ ( Tm = 58°C ) produce no amplicons in the presence of the white insertion . Attempts to remove the w+ marker by Cre recombination were unsuccessful and the w+-disrupted Lhr locus was designated as LhrKO . To test the genetic effects of this mutation , we took advantage of a recent observation that a deficiency chromosome which deletes D . melanogaster Lhr can weakly rescue D . melanogaster-D . mauritiana hybrid males to the pharate adult stage [8] . When we crossed LhrKO homozygous females to D . mauritiana males at 18° , we obtained 10 . 6% rescue of live males ( 17 males and 161 females ) . The stronger rescue observed here may be due to the fact that the mothers of the cross were homozygous for the LhrKO allele , since Lhr likely has strong maternal expression based on its high protein abundance in early embryos [8] . A D . melanogaster Hmr-FLAG transgene was made by inserting a 3× FLAG tag sequence [85] immediately upstream of the stop codon of Hmr using fusion PCR into plasmid p72 , which is a pCaSpeR2 vector containing a ∼9 . 7 kb fragment of the Hmr region [3] . Two Hmr fragments ( L-arm and R-arm ) were amplified from p72 with iProof polymerase by using primer pairs 739/738 and 736/740 , respectively . The primers 738 and 736 contain sequence encoding the FLAG tag and partially overlap to allow fusion in the subsequent stage . The primers 739 and 740 were combined with L-arm and R-arm products to produce a fused partial fragment of Hmr containing the 3× FLAG sequence . This fragment was cloned into the pCR-BluntII-Topo vector ( Invitrogen ) and sequenced completely between the AvrII and KpnI restriction sites . The AvrII/KpnI fragment was then cloned into the corresponding sites of the p72 plasmid . A 300 bp fragment containing the attB site was then PCR amplified from plasmid pTA-attB ( gift from Dr . Michele Calos ) using primers 502 and 503 and cloned into the NotI site . This fragment was digested with NotI ( on the ends of 502 and 503 ) , gel purified , and inserted into the NotI site of the plasmid containing Hmr-FLAG . We refer to this transgene as mel-Hmr-FLAG . A D . melanogaster Hmr-HA transgene was made by inserting a 3XHA epitope tag between codons 466 and 467 of Hmr . Primers 215/1246 and 1247/495 were used to amplify 573 and 316 bp fragments , respectively . Primers 1246 and 1247 overlap and encode the HA tag . Fusion PCR containing these 2 products and primers 215/495 was performed . The PCR product was cloned into pCR-Blunt II-TOPO , and the insert was checked by sequencing . The insert was then cloned using SpeI and BsiWI back into a modified p72 containing an attB site inserted into the NotI site . The orientation and presence of the HA tag were checked by double digests and PCR . We refer to this transgene as mel-Hmr-HA . A D . simulans Hmr-FLAG transgene was made by inserting the 3× FLAG tag sequence upstream of the stop codon in p89 , a pBluescript II KS ( + ) plasmid containing the D . simulans Hmr insert that was used for the p92 transformation construct in [7] . Primers 751/753 and 750/752 were used to amplify 1 . 3 kb and 1 . 8 kb fragments of the insert , respectively , which were then joined by fusion PCR using primers 750/751 . The fusion PCR product was cloned into pCR-Blunt II-TOPO and confirmed by sequencing . The insert was designed to have an HpaI site near one end and a NotI site near the other . The NotI site was destroyed during cloning; however , the pCR-Blunt II-TOPO vector contains a NotI site within 40 bp of the destroyed sequence . The insert was then cloned back into p89 using HpaI and NotI . The orientation of the insert , as well as the addition of the FLAG tag , was checked by double digest with ClaI and HpaI . The D . simulans Hmr-FLAG insert was then removed as a SacII fragment . Klenow enzyme was used to fill-in the ends to allow cloning into the StuI site of pCaSpeR2 containing an attB site inserted at its NotI site . We refer to this transgene as sim-Hmr-FLAG . The D . simulans Hmr-HA transgene was made from plasmid p89 by inserting the HA tag at the region orthologous to mel-Hmr-HA [7] . Primers 135/1365 and 1247/1364 were used to amplify 861 bp and 827 bp fragments , respectively , from the p89 template , and were fused together using primers 1364/135 . The fusion PCR product was then cloned into pCR-Blunt II-TOPO and the entire insert was checked by sequencing . The insert was then cloned back into p89 using SpeI and BlpI . Blunt end ligation , used for sim-Hmr-FLAG above , proved inefficient for transferring the insert into the transformation vector . Therefore an XbaI site was added to the 3′ end of Hmr-HA by amplifying the entire insert using primers 1402/1403 . The PCR product was then gel purified and cloned back into pCR-Blunt II-TOPO . The polylinker contains an XbaI site 5′ to the insert , allowing us to clone the entire insert into the XbaI site of pCaSpeR2 containing an attB site inserted at its NotI site . We refer to this transgene as sim-Hmr-HA . Oligonucleotides for Hmr transgenes ( all written 5′-3′ ) . 739: AGCCAAATTGCCGACAGTAGCCAAG; 738: ATCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCAGGCGGTGGCGGATTGACCTTG; 736: GACGGTGATTATAAAGATCATGACATCGATTACAAGGATGACGATGACAAGTAGCTCTCGAAACTTTTGGCACACGTAG; 740: TTGTACTGCCATTAGGTATAGCTAACCATCC; 502: AAACCCGCGGCCGCATGCCCGCCGTGACCGTC; 503: AAACCCGCGGCCGCGATGTAGGTCACGGTCTCG; 152: TCTTCTTAGACTGCGGGTTG; 215: CAGCGCATGCGCGGCACCGTAT; 1246: ATAGTCCGGGACGTCATAGGGATAGCCCGCATAGTCAGGAACATCGTATGGGTACATTGCACTGTTGGTCATGCTCGT; 1247: TCCCTATGACGTCCCGGACTATGCAGGATCCTATCCATATGACGTTCCAGATTAC;GCTAGCACTGCCACAAGCATTGG; 495: GACACGCCCGTTCCCATAGT; 751: ACAGCGATTTGCGCAAGCCG; 753: TCGATGTCATGATCTTTATAATCACCGTCATGGTCTTTGTAGTCAGGCGGTGGCGGATTTGCCTTCTTGGCGTATTTAGA; 750: GTGAATTGTAATACGACTCACTATAGGGCG; 752: GACGGTGATTATAAAGATCATGACATCGATTACAAGGATGACGATGACAAGTAGCTCTCGAATCATTGGCACACG; 135: GAGGAGGACCCCACCTATAACTAC; 1365: ATAGTCCGGGACGTCATAGGGATAGCCCGCATAGTCAGGAACATCGTATGGGTATGCACTGTTAGAAATGCTTGTGCTG; 1364: GCTGGCAATTTGGACTTTGT; 1402: GCGGGCGGTCATTATTAA; 1403: TATCTAGAGCGGCCGCGAGCTCTAATA . φC31-mediated transgenesis was performed by Genetic Services using the P{CaryP}attP2 integration site at cytological position 68A4 [86] . Site specificity of integration was checked by PCR assays described in references [8] , [87] . D . melanogaster transformants were recognized by their w+-eye color and were crossed to a y w strain . Wild type activity of the Hmr-HA and Hmr-FLAG transgenes was tested for complementation of an Hmr rescue mutation in hybrids as done previously for Hmr+ transgenes [3] , [7] . Here we crossed Df ( 1 ) Hmr− , y w v/FM6; ø{mel-Hmr-HA}/+ females to D . simulans w501 males . We recovered 193 w501/Y; +/+ hybrid males but only 1 w501/Y; ø{mel-Hmr-HA}/+ hybrid male , demonstrating that the transgene is Hmr+ . Likewise , we crossed Df ( 1 ) Hmr− , y w v; ø{mel-Hmr-FLAG}/+ females to D . simulans v males , and recovered 451 v females , 258 w males and only 3 w+ males . LhrKO was outcrossed to w1118 for six generations . Sibling crosses were then used to generate a homozygous w1118; LhrKO/LhrKO ( abbreviated as LhrKO ) , a heterozygous LhrKO/+ , and a wildtype w1118; Lhr+/Lhr+ line ( abbreviated as Lhr+ ) . All experiments with Lhr in this paper use these matched mutant and sibling controls unless otherwise specified . The D . simulans Lhr1 allele is caused by an insertion in the 5′ UTR and appears to make no transcript by RT-PCR [6] . Lhr1 was outcrossed to the inbred wild-type line w501 for 3 generations to generate the stock w501; Lhr1 ( abbreviated as Lhr1 ) and w501 , Lhr+ ( abbreviated as Lhr+ ) . Lhr-HA transgenes were described previously [8] . y w F10 was created by single-pair matings between siblings for 10 generations . We refer to the P{EPgy2}Hmr3 allele that is marked with y+ and w+ described in [4] as Hmr3 . Df ( 1 ) Hmr− , y w v , abbreviated as Df ( 1 ) Hmr− , is described in [88] . In order to match backgrounds for the Hmr RNA-Seq experiments , the Hmr3 stock and the transgenic lines ( mel-Hmr-FLAG and sim-Hmr-FLAG ) were outcrossed to y w F10 for 6 generations and then made homozygous . Individual 1–2 day old virgin LhrKO and LhrKO/+ sibling females , obtained from crosses of LhrKO/+ at 27°C , were crossed to two w1118 males . Flies were transferred to a fresh vial every 5 days for 15 days . Vials in which either the female or both males were missing or dead were not scored or transferred . To create the heteroallelic siblings LhrKO/Df ( 2R ) BSC44 and LhrKO/SM6a , LhrKO/LhrKO were crossed to the Lhr− deletion stock Df ( 2R ) BSC44/SM6a [6] . The fertility assay was carried out as above except vials were flipped every 4–5 days . LhrKO/+ or LhrKO/LhrKO females were crossed to w1118 males at 27°C . Egg lays were carried out on grape juice/agar plates for 3 hour periods at either 2–3 days , 5–6 days or 10–11 days after eclosion of the female parents . The plates were maintained at 27°C and monitored over the next 24–36 hours for hatched eggs . y w Hmr3; +/+ females were crossed to y w; ø{mel-Hmr-FLAG}/ø{mel-Hmr-FLAG} males . F1 males were crossed to Df ( 1 ) Hmr−/FM6; +/+ females to generate both y w Hmr3/Df ( 1 ) Hmr−; ø{mel-Hmr-FLAG}/+ and y w Hmr3/Df ( 1 ) Hmr−; +/+ . Similarly , y w Hmr3; +/+ females were crossed to y w; ø{sim-Hmr-FLAG}/ø{sim-Hmr-FLAG} males . F1 males were crossed to Df ( 1 ) Hmr−/FM6; +/+ females to generate y w Hmr3/Df ( 1 ) Hmr−; ø{sim-Hmr-FLAG}/+ . Lastly , y w; +/+ females were crossed to y w; ø{mel-Hmr-FLAG}/ø{mel-Hmr-FLAG} males . F1 males were crossed to Df ( 1 ) Hmr−/FM6; +/+ females to generate the heterozygous wildtype control , y w/Df ( 1 ) Hmr−; +/+ . These crosses were done at 27°C and in triplicate to generate 3 biological replicates . The Df ( 1 ) Hmr− , y w v/FM7i , P{w+mC = ActGFP}JMR stock ( abbreviated as Df ( 1 ) Hmr−/FM7i , GFP ) was described previously [88] . A stock with the matching Hmr+ genotype , y w v/FM7i , P{w+mC = ActGFP}JMR ( abbreviated as Hmr+/FM7i , GFP ) was created by crossing y w v/Y males with Df ( 1 ) Hmr−/FM7i , GFP females . FM7i , GFP/Y males from this Hmr+ stock were then crossed to Df ( 1 ) Hmr−/FM7i , GFP females for 10 generations in order to make the autosomal backgrounds comparable between the two stocks . To generate hybrids , Df ( 1 ) Hmr−/FM7i , GFP or Hmr+/FM7i , GFP were crossed to v/Y D . simulans males . For each cross , 6 replicates were made each containing 25 0–12 hour-old virgin females and 50 4–6 day-old virgin males . Hybrid larval sons not carrying the balancer were selected by their y− mouth hook and GFP− body phenotypes . Additionally , some crosses were allowed to develop to ensure that only Df ( 1 ) Hmr− crosses produced hybrid sons . To generate D . melanogaster samples , 3 replicates of 10 Df ( 1 ) Hmr−/FM7i , GFP or Hmr+/FM7i , GFP virgin females were crossed to 15 FM7i , GFP/Y males . Larval sons not carrying the balancer were selected by y− and GFP− phenotypes . To generate D . simulans samples , 3 replicates of 10 y w D . simulans virgin females were crossed to 15 v/Y D . simulans males . Larval sons were selected by y− . 50 mg of 1–17 hr embryo collections were dounced 30 times with a tight pestle in 500 ul buffer A1 ( 15 mM HEPES , pH = 7 . 5; 15 mM NaCl; 60 mM 1M KCl; 4 mM MgCl2; 0 . 5% TritonX-100; 0 . 5 mM DDT ) and then centrifuged for 5 minutes at 4°C . The pellet was washed with 500 µl buffer A1 and centrifuged . This process was repeated another two times . The pellet was lysed by douncing in 200 µl SDS lysis buffer ( 500 µl 10% SDS , 200 µl 1M Tris , pH = 8 . 0 , 40 µl 0 . 5M EDTA , 100 µl 100× protease inhibitor , 10 µl 0 . 5M EGTA , 50 µl 100 mM PMSF , 9 . 1 ml water ) . The lysate was allowed to rotate at 4°C for 20 minutes and then centrifuged . The supernatant was removed , quantitated using the Bradford assay and was run on an SDS-PAGE gel . An Lhr cDNA was cloned into pDEST17 ( Invitrogen ) . The expressed protein from E . coli was purified using Ni-Ag beads under denaturing conditions ( 8M urea ) , dialyzed down to 2M urea and injected into rabbits ( Cocalico ) . The antisera was then purified by coupling purified His-Lhr to CnBr-activated Sepharose beads in the presence of 1% Triton-X and removing urea by dialysis . Antisera was eluted in 0 . 2 M glycine , pH 2 . 8 and then neutralized with 1M Tris , pH 8 . 5 . The antibody failed to detect Lhr in immunofluorescent experiments but was used for Western blots in Figure S3 at 1∶4000 in 5% milk-TBST and HRP conjugated anti-rabbit secondary antibody at 1∶2000 dilution . HA-tagged Lhr was detected with 1∶1000 dilution of rat anti-HA ( Roche , 3F10 ) and HP1a was detected with a 1∶700 dilution of mouse monoclonal supernatant ( C1A9 , DSHB ) . 0∼16 hour-old embryos were collected , dechorionated and snap frozen in liquid nitrogen . Embryos were then resuspended to 10× embryo volume of Buffer A ( 10 mM Tris-Cl pH 8 . 0 , 300 mM sucrose , 3 mM CaCl2 , 2 mM Mg acetate2 , 0 . 1% Triton X-100 , 0 . 5 mM DTT , 0 . 5 mM PMSF ) and homogenized with a dounce homogenizer . The homogenized lysate was centrifuged at 700 g for 10 minutes at 4° to pellet the nuclei . The supernatant was removed , the pelleted washed once in Buffer A , the nuclei centrifuged again and then resuspended in 1× embryo volume of Buffer MN ( 15 mM Tris-Cl pH 7 . 4 , 250 mM sucrose , 60 mM KCl , 1 . 0 mM CaCl2 , 0 . 5 mM DTT , 1× protease inhibitor cocktail ) . The nuclear lysate was sonicated briefly , micrococcal nuclease added to a concentration of 500 units/ml , and the chromatin digested for 1 hour at 4° with gentle agitation . EDTA and Triton X-100 were then added to a concentration of 5 mM and 0 . 1% respectively , to inactivate nuclease activity and solubilize the proteins , followed by incubation at 4° for 1 hour . After a second brief sonication , the digest was centrifuged at 12 , 000 g for 10 min at 4° and the supernatant was collected . 50 µl of the chromatin digest was diluted in IP Wash Buffer ( 50 mM Tris-Cl pH 7 . 4 , 100 mM NaCl , 0 . 1% Triton X-100 ) with 1× protease inhibitor cocktail to a final volume of 125 µl per co-immunoprecipitation mixture . 15 µl of protein G-conjugated magnetic beads and 2–5 µl of antibody were added followed by incubation for 4 hours at 4° with gentle agitation . The beads were washed 3 times in IP Wash Buffer . The immunoprecipitated proteins were then eluted by boiling the beads in 1× Laemmli sample buffer for 5 minutes and analyzed by immunoblotting . RNA extraction , cDNA synthesis and qRT-PCR assays were performed as in reference [8] , using 2–5 µg of RNA . qRT-PCR experiments included three technical replicates of three separate biological replicates . Primers included: Lhr-f1 5′caccATGAGTACCGACAGCGCCGAGGAA , Lhr-r1 5′ ACACTTGGTTTTCGGCACATC CGC , Lhr-f2 5′ GTAGCTTTCTCTTGGCGCTCTT , Lhr-r2 5′ GTAAGTGAACTGAAGCTGC GTTGG , EDTP-F 5′GCTGGCAGGTGG TTACCGACA , EDTP-R 5′CGTGGCCAGGTTCA TGGATGA , Bap55-F 5′ CCGAGAGTC TCTTTGACAATGCA , and Bap55-R 5′GCCTCTT CGTACTCCTGCGA . Hmr-f1 5′ TAAGTTCGCCTTCCGCACATACC and Hmr-r1 5′ GACCAGAAACCTGAGTTGCTCCA . HeT-A and RpL32 ( also known as Rp49 ) transcript levels were measured with primers from reference [89] . The Invitrogen DNEasy kit was used to make genomic DNA from LhrKO and Lhr+ female carcasses that were free of ovarian tissue . Primers Het-s2 and Het-as2 amplify from the coding sequence of HeT-A [90] . HeT-A copy number was normalized to RpL32 ( also known as Rp49 ) copy number using primers from reference [89] . For samples from ovaries , flies were kept at 27°C for several generations prior to and during the experiment . Freshly eclosed females were collected and aged 2–3 days and then transferred to fresh food with yeast paste for another 2–3 days . RNA was extracted , from ovaries dissected in chilled 1× PBS , using Trizol . Ovarian mRNA-Seq libraries were constructed at the Epigenomics Core Facility at Weill Cornell Medical College using the poly ( A ) enrichment method . Libraries were sequenced using the Illumina HiSeq2000 platform to produce 50 bp single reads which were then trimmed for quality and filtered to remove rRNA reads . One biological sample each from LhrKO and Lhr+ was duplexed and run in a single lane . 51 , 193 , 832 filtered reads were obtained for Lhr+ and 41 , 688 , 028 reads for LhrKO . Three biological replicates each of D . simulans w501 and Lhr1 ovarian mRNA libraries were run on a single lane and the number of filtered reads ranged from 36 , 472 , 726 to 43 , 449 , 879 . For experiments with Hmr , two biological replicates were included for each genotype and all 8 samples were multiplexed in a single lane . The number of filtered reads for each sample ranged from 23 , 863 , 381 to 27 , 490 , 644 . For larval samples , around 30 larvae were collected for each genotype and flash frozen in liquid N2 . RNA was extracted from 2 biological replicates of each genotype using Trizol . Larval RNAseq libraries were generated and bar-coded using the TruSeq kit , and run in one lane of an Illumina HiSeq 2000 100 bp yielding 13 , 707 , 247 to 20 , 373 , 267 filtered reads per sample , except for one library which produced only 7 , 840 , 004 reads . Reads mapping to either rRNA or repetitive DNA were filtered out using Bowtie [91] and the filtered reads were mapped to the unmasked D . melanogaster genome using Tophat [92] . The BAM file outputs were used by Cuffdiff with the -b option [93] . All * . fasta and * . gtf files were based on the release 5 . 68 of the D . melanogaster genome from ENSEMBL . To find differentially expressed genes in D . simulans , we aligned reads to the D . melanogaster genome with Tophat , allowing two mis-matches . While this approach could potentially reduce mapping ability for diverged genes , it allowed us to take advantage of the better assembly and annotation of the D . melanogaster genome . To maximize the TEs considered in our analyses , we mapped reads to two different databases using Bowtie . First , reads were uniquely mapped to a database consisting of all the annotated TE insertions in the D . melanogaster and D . simulans genomes [48]; we refer to this as the individual-insertion database . While this database likely represents most TE families present in our stocks , some TEs may either be absent from the assembled genome or be represented by copies that are sufficiently diverged such that they impact our ability to correctly assess transcript levels . These elements include the telomeric element TAHRE , which has only a few insertions in the genome and is known to be absent from the reference genome since only two telomeres are included in the assembly [94] . Therefore we also mapped reads , allowing for either 0 mismatches when aligning reads from D . melanogaster or 3 mismatches when aligning reads from D . simulans or hybrids , to a database consisting of the consensus sequences of the annotated TEs and repeats found in Repbase as well as de novo predicted TEs generated by piler-DF using the 12 Drosophila genomes [48]; we refer to this as the consensus-sequence database . Only reads that mapped uniquely within the same family were included in the subsequent analyses of differential expression . Mismatches allowed for each alignment are mentioned in figure legends . Statistical significance of differential expression among TEs was calculated with F . E . T . in the DEG-seq package [95] . To analyze reads mapping to satellite DNAs , we built a database using a curated file from the Berkeley Drosophila Genome Project ( http://www . fruitfly . org/sequence/sequence_db/na_re . dros ) which itself was constructed from GenBank sequences . This file includes some mis-annotated TEs and non-satellite sequences . We counted reads that mapped to these repeats without any mismatches and calculated statistical significance of differential expression among satellites with F . E . T . in the DEG-seq package . Libraries were prepared as described but no oxidation was carried out [38] . Briefly , total RNA was extracted from 5–6 day old LhrKO and Lhr+ ovaries using the mirVANA kit ( Invitrogen ) . Total RNA was size fractionated on a 15% Urea-PAGE gel to enrich for 18–29 nt small RNA , excised and eluted and then subjected to 2S rRNA depletion . This small RNA was ligated to a 3′ RNA adapter , gel purified , and then ligated to a 5′ DNA adapter . The adapter-ligated small RNAs were reverse transcribed and PCR amplified . The amplified PCR products were gel purified , quantified and sequenced in two lanes of a HISeq 2000 machine . Only reads with a 3′ adapter were kept , which was then removed using a custom script [48] . These reads were binned by size as either miRNA/siRNA ( 17–22 nt ) or piRNA ( 23–30 nt ) . rRNA , tRNA and snoRNA sequences were filtered from these reads and the remaining reads were further filtered to keep only those reads that mapped to either the unmasked genome , or the satellite DNA database described above , or Repbase consensus sequences [96] . These filtered reads included 89 , 953 , 149 piRNA reads and 40 , 859 , 119 siRNA reads in LhrKO , and 120 , 143 , 855 piRNA reads and 36 , 388 , 192 siRNA reads in Lhr+ . piRNA reads were mapped uniquely to all D . melanogaster sequences from Repbase using Bowtie , allowing for one mismatch . Ping-Pong scores were calculated using reads mapped with up to 1 mismatch , as described in reference [48] . For mapping to piRNA clusters , we built an index using sequences extracted from the Release 5 DM3 genome on the UCSC genome database and GenBank with coordinates of individual piRNA clusters obtained from reference [41] . piRNA reads were uniquely mapped to piRNA clusters with zero mismatches and significance for differential expression was calculated using F . E . T implemented in DEG-seq . siRNA reads were mapped uniquely to all D . melanogaster sequences from Repbase with Bowtie , without allowing for any mismatches . Immunofluorescence and FISH were performed on embryos and ovaries as described in references [4] , [83] . Polytene chromosomes were dissected in 0 . 7% NaCl , squashed , and fixed in 1 . 8% PFA , 45% acetic acid for 17 minutes . They were then washed in 1% Triton X in PBS for 10 minutes , then washed in 5% milk in PBS for 1 hour , incubated with primary antibody overnight at 4°C , washed in 5% milk in PBS for 10 minutes , incubated with secondary antibody for 1 hour at room temperature , and then washed for 10 minutes in buffer A ( 0 . 15M NaCl , 0 . 2% NP40 substitute , 0 . 2%Tween 20 ) followed by 10 minutes in buffer B ( 0 . 20M NaCl , 0 . 2% NP40 substitute , 0 . 2%Tween 20 ) . Rat anti-HA antibody ( Roche , 3F10 ) was used at 1∶100 , rat anti-Vasa ( DSHB ) was used at 1∶25 , Fibrillarin ( Abcam , Ab5281 ) was used at 1∶100 , anti-HP1a antibody ( C1A9 , DSHB ) was used at 1∶100 . Alexa fluorophore-conjugated secondary antibodies were used to detect the primary antibody . Fluorescently labeled probes against GA-rich satellites , AACAC , 2L3L , 359 bp and dodeca were obtained from Sigma with sequences described in references [8] , [83] , [97] . Imaging was carried out using a Zeiss 710 confocal microscope at Cornell University's Microscopy and Imaging Facility . A full-length coding-sequence plasmid of D . melanogaster Hmr was made by correcting 3 frame-shift errors in the RE54143 cDNA [3] . Two errors in exon 5 were replaced by ligating in a ∼1 . 6 kb XbaI-HindIII fragment from the LD22117 cDNA , followed by replacement of a 2172 bp NdeI-ZraI fragment from the p83 genomic clone [3] . The coding sequence was then PCRd out and cloned into pENTR/D-TOPO . The D . simulans Hmr CDS was PCRd out of cDNA and cloned into pENTR/D-TOPO . The Lhr plasmids and yeast two-hybrid destination vectors and assays are described in reference [6] . Illumina sequence data from this study are available from the NCBI website under BioProject number PRJNA236022 .
|
Sister species capable of mating often produce hybrids that are sterile or die during development . This reproductive isolation is caused by incompatibilities between the two sister species' genomes . Some hybrid incompatibilities involve genes that encode rapidly evolving proteins that localize to heterochromatin . Heterochromatin is largely made up of highly repetitive transposable elements and satellite DNAs . It has been hypothesized that rapid changes in heterochromatic DNA drives the changes in these HI genes and thus the evolution of reproductive isolation . In support of this model , we show that two rapidly evolving HI proteins , Lhr and Hmr , which reproductively isolate the fruit fly sister species D . melanogaster and D . simulans , repress transposable elements and satellite DNAs . These proteins also help regulate the length of the atypical Drosophila telomeres , which are themselves made of domesticated transposable elements . Our data suggest that these proteins are part of the adaptive machinery that allows the host to respond to changes and increases in heterochromatin and to maintain the activity of genes located within or adjacent to heterochromatin .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"biology",
"evolutionary",
"biology",
"genomic",
"evolution",
"evolutionary",
"genetics"
] |
2014
|
The Hmr and Lhr Hybrid Incompatibility Genes Suppress a Broad Range of Heterochromatic Repeats
|
Functional connectivity concerns the correlated activity between neuronal populations in spatially segregated regions of the brain , which may be studied using functional magnetic resonance imaging ( fMRI ) . This coupled activity is conveniently expressed using covariance , but this measure fails to distinguish between direct and indirect effects . A popular alternative that addresses this issue is partial correlation , which regresses out the signal of potentially confounding variables , resulting in a measure that reveals only direct connections . Importantly , provided the data are normally distributed , if two variables are conditionally independent given all other variables , their respective partial correlation is zero . In this paper , we propose a probabilistic generative model that allows us to estimate functional connectivity in terms of both partial correlations and a graph representing conditional independencies . Simulation results show that this methodology is able to outperform the graphical LASSO , which is the de facto standard for estimating partial correlations . Furthermore , we apply the model to estimate functional connectivity for twenty subjects using resting-state fMRI data . Results show that our model provides a richer representation of functional connectivity as compared to considering partial correlations alone . Finally , we demonstrate how our approach can be extended in several ways , for instance to achieve data fusion by informing the conditional independence graph with data from probabilistic tractography . As our Bayesian formulation of functional connectivity provides access to the posterior distribution instead of only to point estimates , we are able to quantify the uncertainty associated with our results . This reveals that while we are able to infer a clear backbone of connectivity in our empirical results , the data are not accurately described by simply looking at the mode of the distribution over connectivity . The implication of this is that deterministic alternatives may misjudge connectivity results by drawing conclusions from noisy and limited data .
In the early days of neuroscience much attention was devoted to identifying the functional specialization of different brain areas [1] . More recently , this focus has shifted towards revealing how these areas are organized into networks and how these networks , rather than their individual constituents , are related to cognition [2–4] and neurological or psychological pathology [5–7] . The increasing interest in neuronal connectivity sprouted its own subdiscipline known as connectomics [8–10] . Within connectomics , one distinguishes between structural connectivity and functional connectivity . Structural connectivity is concerned with the anatomical white-matter fiber bundles that connect remote regions of the brain . It may be estimated in vivo by diffusion weighted MRI ( dMRI ) , which measures the fractional anisotropy of the diffusion of water molecules [11] . Functional connectivity in turn expresses the ( degree of ) dependency between the neuronal activity of separate brain regions [6 , 12] and is typically measured non-invasively via either functional MRI , electro- or magnetoencephalography ( fMRI , EEG and MEG , respectively ) [13] . Several measures to quantify ( the degree of ) functional coupling exist [14 , 15] , of which the most prevalent is covariance . When the activity signal is normalized to have zero mean and unit variance , covariance coincides with Pearson correlation . As the correlation matrix is easy to compute , it has become the de facto standard in operationalizing functional connectivity . It does however have an important drawback: it is unable to differentiate between direct and indirect effects . For example , if regions A and B are correlated , and similarly B and C show correlation , then correlation between A and C is induced [16 , 17] . This poses a problem for functional connectomics , as it introduces type 1 errors . The problem may be remedied to some extent by using partial correlations instead . Its interpretation is similar to Pearson correlation , but it captures only direct effects as the influence from other regions is partialled out . In practical terms , the matrix of partial correlations may be obtained by taking the inverse of the covariance matrix , known as the precision matrix , and rescaling this . Assuming the data are normally distributed , both the precision matrix and the partial correlation matrix capture the conditional independence structure of the considered variables , i . e . when two regions are conditionally independent given all other regions , their precision and partial correlation are zero . Ideally , the partial correlation matrix would correctly reflect the functional connectivity that generated the observed data . If this matrix is sparse , the corresponding conditional independence graph provides an intuitive representation of the interaction between different regions . In practice however , the obtained partial correlation matrices are not sparse , which makes the estimated connectivity more difficult to interpret . In addition , if the number of samples is small and the number of regions large , there is no unique inverse of the covariance matrix and consequently no unique matrix of partial correlations . Even when these conditions are met , the maximum likelihood solution is often ill-behaved , in which case the solution must be regularized [18] . A popular approximation of the precision matrix is acquired via the graphical LASSO ( Least Absolute Shrinkage and Selection Operator ) , which regularizes the elements of the precision matrix using the ℓ1-norm [14 , 16 , 19] . This approach shrinks the partial correlations towards zero so as to create sparse solutions , which are easier to interpret . Although the graphical LASSO was found to be one of the must accurate methods in identifying connectivity in a comparative study [14] , it introduces a bias that underestimates functional connectivity , thus creating type 2 errors [20] . In addition , both the original maximum likelihood solution as well as the LASSO estimate provide point estimates that do not quantify the reliability of their outcome . In earlier work , we have proposed a Bayesian alternative to the graphical LASSO that uses the G-Wishart distribution to restrict the partial correlation estimates to a previously defined conditional independence graph . We showed that structural connectivity provides an elegant candidate for this graph , and that this approach was able to outperform the graphical LASSO on simulated data [20] . Importantly however , we assumed that the conditional independence graph was available a priori . In the current contribution we take this line of reasoning a critical step forwards and learn both functional connectivity as well as its conditional independence structure simultaneously . Apart from estimating the degree to which two regions have correlated activity , we can now also express the probability of these regions being conditionally independent . As we will show , this results in a more effective approach to regularization than the graphical LASSO , while retaining the additional benefits of the Bayesian framework . At the foundation of this contribution lies a probabilistic generative model that describes how a particular independence structure generates partial correlations that in turn generate observable data . Using a neurologically plausible simulation with several different conditions , as described by Smith et al . [14] , we show that in many cases our Gaussian graphical model approach is favorable to both the maximum likelihood alternative and graphical LASSO regularized solutions . Subsequently , we apply the model to estimate functional connectivity between bilateral accumbens , amygdala , caudate , hippocampus , pallidum , putamen and thalamus using their blood-oxygenation level dependent ( BOLD ) signal time courses , measured using resting-state fMRI . Finally , we demonstrate how the advantages of a Bayesian approach can be put to practice by showing two extensions to our connectivity model . First , we show how the problem of data fusion for connectivity studies [21 , 22] may be tackled by simply providing multiple likelihood terms; one for each imaging modality . This is demonstrated empirically by combining the fMRI time series with dMRI probabilistic tractography results . Second , we describe how further background knowledge on putative connections may be used to both constrain and inform functional connectivity .
From a methodological perspective , elucidating functional connectivity is often rephrased as a covariance selection problem . This boils down to finding a sparse partial correlation matrix associated with the time series ( activity ) of a set of variables ( brain regions ) , a problem known as covariance selection . Here , the problem is approached using a Gaussian graphical model ( GGM ) , where we assume that the data X = ( x1 , … , xn ) T consist of n independent draws from a p-dimensional multivariate Gaussian distribution N ( 0 , K - 1 ) , with zero mean and precision ( inverse covariance ) matrix K . Here , K ∈ P p , with P p the space of positive definite p × p matrices . The likelihood of K is given by P ( X ∣ K ) = ∏ i = 1 n N ( x i ∣ 0 , K - 1 ) ∝ | K | n / 2 exp [ - 1 2 ⟨ K , Σ ⟩ ] , ( 1 ) where Σ = XT X and ⟨⋅ , ⋅⟩ the trace inner product operator . The assumption of Gaussianity is justified empirically , as BOLD data has been shown to follow a Gaussian distribution [23] . The precision matrix has the important property that zero elements correspond to conditional independencies , provided the data are normally distributed . In other words , Eq ( 1 ) specifies a Gaussian Markov random field with respect to a graph G = ( V , E ) , with V = {1 , … , p} and E ⊂ V × V , in which the absence of a connection indicates conditional independence , i . e . ( i , j ) ∉ E → kij = 0 [24 , 25] . In order to estimate the precision matrix K of a zero-mean multivariate Gaussian density from data X one may maximize the log-likelihood which gives the maximum likelihood estimate ( MLE ) : K ^ = arg max K ∈ P p ( log | K | - ⟨ Σ K ⟩ ) ( 2 ) where the maximization is constrained to precision matrices in the family of p × p positive definite matrices P p . If Σ is positive-definite , there exists a unique solution to Eq ( 2 ) in the form of Σ−1 . However , if the number of samples is small compared to the number of variables , the solution does not exist , and even if n > p , the maximum likelihood estimate is often ill-behaved and requires regularization [18] . A frequently used method of regularization is called the graphical LASSO [26] , which penalizes the magnitude of the elements of K . The LASSO approach gives the following MLE: K ^ = arg max K ∈ P p [ log | K | - ⟨ Σ K ⟩ - λ ∥ K ∥ 1 ] , ( 3 ) in which the shrinkage parameter λ determines the amount of penalization that is applied . Several studies have applied the graphical LASSO in order to estimate functional connectivity [14 , 16 , 19] . Alternative regularization schemes are available [27] , such as ridge regression or elastic net [28] , but we will not consider these methods in detail here . Rather , we emphasize that each of these regularization approaches provides only a point estimate , instead of a posterior distribution over K . This makes it impossible to quantify the uncertainty associated with the estimate , which can lead to incorrect conclusions about functional connectivity in light of finite data . Moreover , it has been shown that the graphical LASSO is not guaranteed to find the true graph even in the limit of infinite data [29] . In addition , solutions obtained through regularization tend to underestimate functional connectivity [20] . Recently , extensions of the ( graphical ) LASSO approach have been proposed that allow for statistical inference . For example , [30] introduce a significance test that can be applied to LASSO estimates while [31 , 32] describe a desparsified LASSO that attempts to de-bias the results using a projection onto the residual space . However , these approaches make assumptions on the sparsity of K , which may not be warranted . Alternatively , a Bayesian approach can be applied to the covariance selection problem , which dispenses with these assumptions . It requires that we specify a prior distribution on K . As we hope to identify conditional independencies between the considered variables , a convenient prior distribution arises in the form of the G-Wishart distribution [33]: P ( K ∣ G , δ , D ) = W G ( δ , D ) = | K | ( δ - 2 ) / 2 Z G ( δ , D ) exp [ - 1 2 ⟨ K , D ⟩ ] 1 K ∈ P G , ( 4 ) in which P G is the space of positive definite p × p matrices that have zero elements wherever ( i , j ) ∉ G , δ is the degrees of freedom parameter , D is the prior scaling matrix and 1 x evaluates to 1 if and only if x holds and to 0 otherwise . The G-Wishart distribution is conjugate to the multivariate Gaussian likelihood in Eq ( 1 ) , so that P ( K ∣ G , δ , D , X ) = W G ( δ + n , D + Σ ) = | K | ( n + δ - 2 ) / 2 Z G ( δ + n , D + Σ ) exp [ - 1 2 ⟨ K , D + Σ ⟩ ] . ( 5 ) Note that the Wishart distribution is a special case of the G-Wishart distribution , with which it coincides if G is a fully connected graph . It should be pointed out that in the limit of n → ∞ , any prior will be fully dominated by the data . In theory , even when the true precision matrix K contains very small elements , the probability of a corresponding edge will go to 1 in the limit of an infinite amount of data . The interesting question is what happens if the magnitude of these elements scales as a function of n , e . g . , as 1/n . Where asymptotic analyses have been successfully applied to better understand the behavior of regularization approaches such as the graphical LASSO [34 , 35] , such analyses of Bayesian procedures are complex and may lead to counterintuitive results [36] . For the G-Wishart prior in particular , similar analyses have , to the best of our knowledge , not yet been pursued . The preliminaries described above allow us to specify the distribution that is central to this work , i . e . the joint posterior over both the conditional independence graph and the precision matrix ( an illustration of the graphical model is provided in Fig 1A ) : P ( G , K ∣ X ) ∝ P ( X ∣ K ) P ( K ∣ G ) P ( G ) . ( 6 ) Note that the necessary hyperparameters are typically omitted for clarity . In practice , functional connectivity is more intuitively understood in terms of partial correlations than as elements of the precision matrix . The partial correlation matrix R may be obtained from the precision matrix by applying the transformation r i j = { 1 if i = j , - k i j k i i k j j otherwise . ( 7 ) By transforming each element of K in Eq ( 6 ) , the distribution P ( G , R ∣ X ) is constructed . When discussing our experimental results , we will focus on partial correlations rather than precision values , unless explicitly stated otherwise . Note that the relation between the dependency structure G and the precision matrix K , as discussed above , also holds between G and the partial correlations R . That is , absence of a connection in ( i , j ) ∈ G implies rij = 0 . The Bayesian generative model must be completed by specifying a prior distribution to draw G from . Here , we assume that a priori all edges are marginally independent and each have probability θ . That is , we have P ( G ∣ Θ ) = ∏ i < j θ i j g i j ( 1 - θ i j ) 1 - g i j , ( 8 ) with gij ∈ {0 , 1} , gij = 1 ↔ ( i , j ) ∈ G and Θ = ( θij ) i < j . Initially we use θij = 0 . 5 ∀i , j to indicate that we have no a priori preference for either a dependence or an independence . The impact of different values for θij on the posterior estimates is discussed in S3 Text , where it is shown that the prior is to a large extent dominated by the likelihood . One of the benefits of the Bayesian framework is that extensions to the generative model are straightforward to implement . In this section we use the distribution given in Eq ( 6 ) to provide two illustrations of such extensions for analyzing connectivity . To analyze the performance of the Gaussian graphical model approach to functional connectivity , we compare our results to those presented in [14] . Here , realistic BOLD time series are generated according to the dynamic causal modeling ( DCM ) fMRI forward model [42] , that makes use of the nonlinear balloon model [43] , based on a known constructed network as its starting point . In total , 28 simulations with different parameters such as number of nodes , number of generated samples , sampling frequency and noise levels were constructed . For each simulation , 50 different time series are generated , simulating different ‘subjects’ ( throughout we will refer to these pseudo-subjects as ‘runs’ , to avoid confusion with the empirical data later on ) . The networks in the simulations were composed of 5 , 10 , 15 or 50 nodes and for each node between 50 and 10 000 samples were generated . For 15 of the 28 simulations , additional characteristics were introduced , such as shared input between a number of nodes , or mixing in timeseries between nodes ( mimicking the effect of bad ROI definition ) [14] . For the full description of the approach as well as the additional simulation parameters , we refer to the original description in [14] as well as the corresponding web page where the simulation may be downloaded ( http://www . fmrib . ox . ac . uk/analysis/netsim/ ) . In the simulation study , it was shown that using partial correlation ( both maximum likelihood as well as LASSO regularized point estimates ) resulted in the best ( undirected ) reconstructions of the ground truth . As these methods performed best , and are closely related to our approach , we use these to compare our results with . The evaluation procedure is as follows: For each run of each of the 28 different simulations , the time series X of that run are used to compute P ( G , R ∣ X ) . In addition , for each run the maximum likelihood estimate ( MLE ) is computed , as well as the graphical LASSO regularized point estimate using the same regularization as in [14] ( i . e . λ ∈ {5 , 100} ) . The quality of the reconstruction of the ground truth is quantified in three ways . Let R* be the ground truth functional connectivity that we are trying to recover and let T be a matrix that has 1 in its elements whenever the corresponding edge is present in the ground truth network , and 0 otherwise ( ignoring directionality ) . Then Γ = ∣R* − R∣ gives the reconstruction error ( where R is either a sample from P ( G , R ) , or a point estimate ) . The total reconstruction error is η ( Γ ) = 2 p ( p - 1 ) ∑ i < j γ i j , the true positive error is η tp ( Γ ) = 1 N tp ∑ i < j γ i j δ t i j ≠ 0 , where Ntp is the number of nonzero elements in the ground truth R* , i . e . the number of true present connections , and finally the true negative error is given by η tn ( Γ ) = 1 N tn ∑ i < j γ i j δ t i j = 0 , where Ntn is the number of zero elements in the ground truth R* , i . e . the number of true absent connections . The indicator function δx evaluates to 1 if and only if its argument x holds true , and to 0 otherwise . In [14] , a null distribution is computed for each of the different methods , by randomly permuting the node labels in the different runs ( to remove any influence between the different nodes ) , which is subsequently used to derive a z-score for an error measure similar to η . However , in the case of Bayesian functional connectivity , a distribution characterizing the uncertainty of the results is already available in the form of P ( G , R ) . By applying η to each of the samples of this distribution , we obtain P ( η ) . The standardized scores of a point estimate R relative to the BGGM distribution may be computed as z ( R ) = ( η ( R* , R ) − μ ) /σ , in which μ and σ are the mean and standard deviation of the distribution , respectively . The procedure is illustrated in Fig 2 . The Bayesian formulation of the model allows us to describe and compare the shapes of the different posterior distributions . We compute the entropy of the posterior distributions as H = - ∑ G [ P ( G , K ∣ X ) log 2P ( G , K ∣ X ) ] , ( 13 ) to indicate the diversity of models that have been encountered in the Markov chains . In addition , the posterior probability of the maximum a posteriori sample is derived , i . e . P ( G ^ , K ∣ X ) = max G P ( G , K ∣ X ) , ( 14 ) to quantify how much of the posterior distribution is dominated by its mode . The Markov chain Monte Carlo ( MCMC ) scheme as described in S1 Text was used to approximate the posterior distributions of interest for each subject using either the simulated BOLD signal time series , the BOLD time series data for the fourteen subcortical regions ( see Eq ( 6 ) ) , the combination of time series data and tractography output for the subcortical regions ( see Eq ( 11 ) ) or finally the BOLD time series data in combination with the informed prior . Throughout , a vague prior on the precision is used: P ( K ∣ G ) = W G ( 3 , I p ) , cf . [44] . The parameters of the probabilistic streamline model are set to ( α , β ) = ( 1 , 0 . 5 ) , which expresses that high streamline counts are most likely associated with a structural connection , while still allowing for tractography noise [40] . Once convergence was established , the approximated distributions were uniformly thinned to T = 1 000 samples , to make subsequent analyses more manageable and to have an equal number of samples for all different settings . Details of convergence monitoring and computation speed are provided in S2 Text .
Fig 3 shows the ( smoothed ) histograms of z-scores aggregated over the 50 runs per simulation , for the graphical LASSO approach with λ = 100 ( the results for λ = 5 and the MLE are almost identical; the MLE results are shown in S1 Fig ) . In this figure , distributions of errors with high z-scores have substantially larger errors than the errors from the BGGM approach , while distributions with low z-scores have smaller errors . The significance threshold at p < 0 . 01 is indicated by the red dotted lines . The first row of Fig 3 shows the total scores ( both true positives and true negatives ) for each simulation , while the second and the third row split this score into the contributions for true positive connections and true negative connections , respectively . These results indicate that in terms of true positives , the LASSO approach typically has an equal to slightly better performance than our Bayesian alternative . However , the BGGM approach identifies true negatives at least as well as G-LASSO , and in several cases significantly outperforms it . On the whole , the proposed method is up to par with the graphical LASSO ( for λ ∈ {5 , 100} ) and the MLE , while at times outperforming them greatly . To obtain insight in the behavior that creates these results , we take a closer look at some of the simulation results . As an example , Fig 4A shows the ground truth network and the reconstruction by the graphical LASSO , as well as the expectation ( i . e . posterior mean of the samples ) using the BGGM approach . In addition , the figure shows for three different connections the estimated partial correlation in detail . The first , between nodes 1 and 5 , is present in the ground truth network . Our approximation is ( correctly ) confident that this node pair is not independent , and assigns a posterior partial correlation distribution close to the ground truth . The graphical LASSO estimate is slightly closer to the ground truth than the mode of the distribution . For the second node pair , between nodes 3 and 5 , a connection should be absent , but because of the limited number of data samples the signals of these nodes have become correlated . This time , the BGGM approach shows a bimodal distribution . The first mode is centered close to the graphical LASSO estimate , but the second mode is at zero , as there is non-negligible evidence for this pair of nodes being disconnected . This means that on the whole ( i . e . the entire distribution ) , the BGGM approach correctly estimates this connection strength lower than the graphical LASSO . A similar observation can be made for the third node pair , between nodes 1 and 4 , of which the BGGM estimate is fairly certain about their independence . Because of this , most of the partial correlation mass is at zero , rather than at the value indicated by the graphical LASSO estimate . These results beg the question: what if we regularize the graphical LASSO even more ? Although Smith et al . report no further improvement after λ = 100 [14] , it is possible that more regularization brings the graphical LASSO estimate closer to the BGGM results . In Fig 4B , the same visualization is provided , but this time for λ = 10 000 . This time , we see that indeed the graphical LASSO estimate is closer to the BGGM expectation than before . In particular for the connection between nodes 1 and 4 , the graphical LASSO now correctly estimates the absence of this connection . However , for the connection between nodes 3 and 5 , the results hardly change , which means that the BGGM estimate is closer to the ground truth still , as , conditioned on the absent connection , the estimated partial correlation is zero . Finally , for the true positive connection between nodes 1 and 5 , we see that the strong regularization causes the graphical LASSO to underestimate the connection , which will only become worse when we increment λ even further . These results may similarly be interpreted in terms of the ( in ) dependence graph . For weak regularization , the graphical lasso suggests false positives due to limited data . For more regularization , the same dependency structure is recovered as using ( the mean of ) the BGGM approach ( see for example Fig 4B ) . Regularizing even stronger introduces false negatives . Note that these results follow from the results of the recovered partial correlation structures and are therefore not explicitly presented here . In addition , we applied the extended BIC over the ‘graphical LASSO path’ ( i . e . we applied the EBIC penalty to the graphical lasso estimates over a logarithmic range of λ , with the maximum penalty corresponding to the empty graph , as used in [54] ) to a number of simulations . However , this analysis did not result in a λ will results significantly different than those already presented here , and has been omitted here . The pattern of simulations in which the BGGM outperforms the graphical LASSO is not random . In [14] , each of the simulations is based on a network consisting of 5 nodes , except for simulations 2 , 3 , 4 , 6 , 11 , 12 and 17 , which consist of networks of 10 , 15 , 50 , 10 , 10 and 10 nodes , respectively . Precisely these simulations benefit the most from the BGGM approach , as can be seen in Fig 3 . As for these simulations the ratio N/p is smallest , it is here that the most improvement can be obtained from regularization , e . g . by the graphical LASSO [14] . As we have shown above , the BGGM provides further improvement still , because this approach conditions on conditional independencies . We further analyzed the effect of sample size on recovery of the ground truth by taking the simulation with the most available samples ( simulation 7 in [14] ) and attempting to recover the ground truth using increasingly smaller subsets of the samples . We compared the BGGM results with the graphical LASSO with λ ∈ {5 , 100 , 1 000 , 10 000} . The outcome of this experiment is shown in Fig 5 , once again split into the total error , error in recovery of true positives and error in recovery of true negatives . The results indicate that for small sample size , the BGGM approach already outperforms the graphical LASSO in total error , although the differences become more pronounced as more samples are considered . Extremely strong regularization ( i . e . λ = 10 000 ) does result in better estimation of absent connections ( by simply forcing almost all connections to zero ) , but this comes at the cost of excluding connections that should be present . For weak regularization ( i . e . λ = 5 ) , small sample size appears to be somewhat beneficial in recovery of true positive connections , as here the performance of the graphical LASSO is similar to our approach . However , this effect diminishes as more samples are acquired ( inducing more spurious connections ) . In terms of true negatives , weak regularization is clearly outperformed by the BGGM approach . In addition , we analyzed the effect of small sample sizes on the estimates . We used simulation 3 ( with p = 15 ) and repeated the procedure as before , but this time the number of samples was varied n ∈ {5 , 10 , … , 45 , 50} , so that situations of n < p were included . The results of this experiment are shown in Fig 6 . They show that , unsurprisingly , weak regularization ( i . e . λ = 5 ) is insufficient to recover the ground truth when few samples are available . Strong shrinkage ( i . e . λ = 10 000 ) results in a low recovery error , but this comes at the expense of significantly underestimating true positive connections . In general , the BGGM approach performs approximately equal to the graphical LASSO for small to moderate regularization , given this limited sample size scenario . Below we discuss the connectivity estimates we obtained on the empirical data , for the original BGGM model , the data fusion variant and the effect of incorporating background information .
Functional connectivity may be quantified using different metrics . The most obvious approach is to use Pearson correlation , but this metric is sensitive to polysynaptic influences . An alternative that does not suffer from this drawback is partial correlation , which was further advocated for its ability to retrieve true connections and its capacity to deal with noise [14] . Partial correlation between two variables may be interpreted as Pearson correlation conditioned on all other variables . In practice , partial correlation can be computed by applying a simple transformation to the precision matrix of a multivariate Gaussian distribution . The precision matrix and , consequently , the matrix of partial correlations , has the interesting property that conditional independence between variables , given all other variables , appears as a zero value in the corresponding matrix element [20 , 55] , which may conveniently be collected in a conditional independence graph . Typically , this graph is mostly ignored , while the precision or partial correlation matrix is considered the quantity of interest . In this paper , we have provided a Bayesian generative model for functional connectivity in which the conditional independence graph plays a central role , as it is assumed to generate the precision matrix and thus functional connectivity . As opposed to regularized maximum likelihood estimates for the precision matrix , our approach characterizes the full posterior distribution of both conditional ( in ) dependencies and partial correlations . In addition to this model , we described a number of model variants that address specific issues with , and conceptual extensions to , connectivity . We subjected our approach to the simulations that were presented in [14] , and compared its performance to the maximum likelihood estimate as well as to the graphical LASSO . The latter of these two has been shown to be the most successful in recovering connectivity in these simulations [14] . The results of the simulation are encouraging . Although we observe that for true positive connections , our approach occasionally underestimates connections , it more than compensates for this in correctly estimating true negatives ( i . e . the sparsity structure of the network ) . When true positives and true negatives are both taken into account , corrected for their respective numbers of occurrence , we find that our approach performs at least as well as the graphical LASSO , and significantly better for simulations with small sample size compared to the number of nodes in the network . A closer look at these results shows that when estimating partial correlations , conditioning on the presence or absence of a connection provides a considerable advantage over shrinkage . In particular for connections with a moderate probability of independence our method yields a bimodal distribution of partial correlations , differentiating between the conditionally dependent and independent node pairs . In addition to our simulation results , we used our approach to approximate the posterior distribution of functional connectivity between subcortical areas for twenty participants . This allowed us to identify a connectivity backbone that consists of strong connections and partial correlations . At the same time , we see that a number of connections are strongly dependent , but foster only weak partial correlations . This emphasizes that a richer picture of connectivity is obtained by looking at both the structure of conditional independence , as well as the strength of these connections in terms of partial correlation . Partial-correlation based methods are susceptible to common input effects that may induce spurious connections if they are not accounted for , for example when variables ( i . e . brain regions ) are missing [56 , 57] from the analysis . If instead the full neural system is observed , it is straightforward that direct functional connections presuppose anatomical connections between the corresponding regions . This allows us to combine the generative model for functional connectivity with a similar model for structural connectivity [39] using probabilistic tractography obtained from diffusion weighted MRI . Conceptually , this results in a data fusion model in which an underlying model of anatomy drives both the observations for functional interactions , as well as for estimates of structural fibres . Compared to alternatives that , for example , weigh a regularization parameter by the strength of structural connectivity [22 , 38 , 58–61] , our approach is based on a generative model in which data fusion is made possible by the use of different likelihood terms . Furthermore , in our model both sources of data affect both types of connectivity; structural connectivity regularizes functional connectivity and simultaneously functional dependencies influence the probability of structural connections . On empirical data the data fusion approach leads to sparser connectivity , in particular between hemispheres . However , some connections are conditionally dependent to such a degree that the model infers a connection regardless of the lack of support by the tractography data . This is helpful in estimating structural connectivity , as it is well known that structural connectivity based on diffusion weighted imaging suffers from a large number of false negatives [62] . In addition , data fusion lowers the variance for many of the partial correlations , indicating that combining both imaging modalities leads to more robust estimates [49 , 62–64] . However , for a number of connections the data for functional and structural connectivity appear to contradict each other , which actually results in increased variance . Note that our data fusion approach has similarities to linked ICA [65] , which also uses a Bayesian generative model to integrate different data modalities . However , whereas linked ICA assumes that each data modality may be decomposed into a number of ( shared ) components , our model assumes that anatomical connectivity is the variable that is shared across modalities . Our final model variant uses an informative prior which encodes the assumption that between-hemisphere connections are restricted to those between functionally homologous regions ( cf . for example [66] ) . This is only one of many prior distributions that , depending on the research question and available background information , may be used to inform the connectivity estimates . As expected , the prior removes the negative partial correlations that are visible for contralateral connections in the other model variants . Indirectly , the prior also affects the partial correlations within hemispheres , as they become slightly lower in magnitude across the board . These results touch upon an unresolved issue in connectomics concerning the interpretation of negative ( partial ) correlations . It has been suggested that a substantial number of negative partial correlations are due to global signal regression and are therefore artifactual in nature rather than biological [67–69] . On the other hand , it has been shown that even without global signal regression , negative connections exist and these may even have biological meaning [70] . Although it is outside the scope of this paper to resolve this matter , we have shown that an informed prior may be used to encode such assumptions or correct for biases . As our approach is Bayesian it directly allows for statistical inference , so that the uncertainty associated with our estimates may explicitly be quantified . In terms of a binary graph that indicates conditional ( in ) dependence , this expresses itself by providing an expectation of a connection rather than a point estimate . For partial correlations , the approach provides the supported distribution instead of a single value . These posterior distribution shapes reveal that none of the model variants are dominated by their mode . In particular for the original model the distributions are very broad and contain many unique models . Although a number of connections is consistently present , the conditional independence graphs vary substantially across subjects . In contrast , the data fusion approach and the informed prior result in distributions that are more tightly centered around the maximum a posteriori connectivity , yet even here there remains substantial support for alternative models . This has important implications for connectomics studies . These are typically aimed at obtaining a point estimate ( which can often be interpreted as the mode of an implicit posterior distribution ) , so a substantial number of connections with significant support from the data will be excluded and spurious connections will be suggested . The widths of the posterior distributions strongly advocate a Bayesian approach , or at the very least point-estimated connectivity results should be treated with great care , e . g . by applying a bootstrapping procedure [71] . The main limitation of our study is one of scale . Bayesian inference has the drawback of being computationally demanding in approximating the posterior distributions , and although state-of-the-art machinery has been applied to make this process efficient , it remains impossible to apply the same methods to a large number of variables . Applying the models to large-scale data sets requires either more efficient implementations , e . g . by using GPU programming , or additional efficiency gains in the field of Gaussian graphical models . Finally , a fundamental assumption in Gaussian graphical model estimation is that the functional data are normally distributed . Should this assumption fail , it may prove difficult to interpret the estimated connectivity . However , as discussed by [23] , BOLD time series do tend to be mostly Gaussian . The most pressing issue for future work is , as mentioned above , improving the methodology to handle a larger number of variables . However , a number of interesting research questions may be addressed even with a limited number of regions . For example , a model may be constructed that defines the BOLD time series to be generated by a mixture of partial correlation matrices , instead of a single one . By applying appropriate constraints , such as that consecutive datapoints are likely to be generated by the same connectivity matrix , this setup can be applied to differentiate experimental conditions based on their connectivity distributions [72] . Similarly , subjects may be assigned to either patients or healthy controls by defining a shared conditional independence graph for either group . The data fusion approach may be extended to incorporate any number of imaging modalities , provided that a forward model can be constructed that shares at least one variable with the other modalities . For example , structural connectivity may inform functional connectivity estimated from MEG instead of or in addition to fMRI data [59] . Additional information may also be incorporated into the prior . This may be explicit evidence for ( the absence of ) a connection , e . g . tracer studies that reveal the presence of a fiber bundle can make particular connections more likely or , conversely , knowledge about white-matter lesions may preclude connections . Alternatively one could construct a prior in which the probability of a connection is a function of the distance between the corresponding end points . In conclusion , the proposed Bayesian approach to functional connectivity has demonstrated that connectivity may be meaningfully divided into structure and strength . Several model variants have been discussed , each with their own characteristics . Application of the models has shown convincingly that multiple unique structures are possible given the same data . This illustrates the advantages of a Bayesian approach to connectivity , and provides a word of caution for traditional ( regularized ) maximum likelihood estimators .
|
Significant neuroscientific effort is devoted to elucidating functional connectivity between spatially segregated brain regions . This requires that we are able to quantify the degree of dependence between the signals of different areas . Yet how this must be accomplished—using which measures , each with their own limitations and interpretations—is far from a trivial task . One frequently advocated metric for functional connectivity is partial correlation , which is related to conditional independence: if two regions are independent , conditioned on all other regions , then their partial correlation is zero , assuming Gaussian data . Here , we use a probabilistic generative model to describe the relationship between functional connectivity and conditional independence . We apply this Bayesian approach to reveal functional connectivity between subcortical areas , and in addition we propose different variants of the generative model for connectivity . In the first , we address how a Bayesian formulation of connectivity allows for integration with other imaging modalities , resulting in data fusion . Secondly , we show how prior constraints can be incorporated in our estimates of connectivity .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Bayesian Estimation of Conditional Independence Graphs Improves Functional Connectivity Estimates
|
Fibrillin-1 is a ubiquitous extracellular matrix molecule that sequesters latent growth factor complexes . A role for fibrillin-1 in specifying tissue microenvironments has not been elucidated , even though the concept that fibrillin-1 provides extracellular control of growth factor signaling is currently appreciated . Mutations in FBN1 are mainly responsible for the Marfan syndrome ( MFS ) , recognized by its pleiotropic clinical features including tall stature and arachnodactyly , aortic dilatation and dissection , and ectopia lentis . Each of the many different mutations in FBN1 known to cause MFS must lead to similar clinical features through common mechanisms , proceeding principally through the activation of TGFβ signaling . Here we show that a novel FBN1 mutation in a family with Weill-Marchesani syndrome ( WMS ) causes thick skin , short stature , and brachydactyly when replicated in mice . WMS mice confirm that this mutation does not cause MFS . The mutation deletes three domains in fibrillin-1 , abolishing a binding site utilized by ADAMTSLIKE-2 , -3 , -6 , and papilin . Our results place these ADAMTSLIKE proteins in a molecular pathway involving fibrillin-1 and ADAMTS-10 . Investigations of microfibril ultrastructure in WMS humans and mice demonstrate that modulation of the fibrillin microfibril scaffold can influence local tissue microenvironments and link fibrillin-1 function to skin homeostasis and the regulation of dermal collagen production . Hence , pathogenetic mechanisms caused by dysregulated WMS microenvironments diverge from Marfan pathogenetic mechanisms , which lead to broad activation of TGFβ signaling in multiple tissues . We conclude that local tissue-specific microenvironments , affected in WMS , are maintained by a fibrillin-1 microfibril scaffold , modulated by ADAMTSLIKE proteins in concert with ADAMTS enzymes .
Mutations in fibrillin-1 cause the pleiotropic features of the Marfan syndrome ( MFS , OMIM#154700 ) . MFS is recognized by its unique combination of skeletal , cardiovascular , and ocular features ( long bone overgrowth , aortic root dilatation and dissection , and ectopia lentis ) . More than a thousand different mutations in FBN1 , the gene for fibrillin-1 , are known to cause MFS , suggesting that the same general pathogenetic mechanisms are initiated by each of these distinct mutations . In contrast , Weill-Marchesani syndrome ( WMS , OMIM #608328 ) is a rare disorder described as “opposite” to MFS [1] . WMS , one of several types of acromelic chondrodysplasias , is characterized by short stature , brachydactyly , thick skin , and ectopia lentis . Previous studies reported that the autosomal dominant form of WMS is caused by mutations in FBN1 [2] , [3] , while mutations in ADAMTS10 were shown to cause recessive WMS [4] , [5] . Since the clinical features of WMS and MFS may sometimes overlap [6] , it is not certain how rare mutations in FBN1 can bring about WMS instead of MFS . Additional investigations are required in order to clearly establish the role of fibrillin-1 in causing WMS . A role for fibrillin-1 in skin fibrosis was first suggested when a mutation in Fbn1 was identified in the tight-skin ( tsk ) mouse [7] . More recently , mutations in FBN1 were found in Stiff Skin Syndrome ( SSKS , OMIM #184900 ) , a rare disorder characterized by hard , thick skin and joint contractures [8] . Both the tsk and SSKS phenotypes are caused by heterozygous mutations . However , the tsk mutation is a large in-frame gene duplication , while SSKS mutations are missense mutations confined to exon 37 . The molecular mechanisms by which fibrillin-1 regulates skin fibrosis are obscure . Why the tsk and SSKS mutations do not cause MFS is also obscure . Fibrillin-containing microfibrils are small diameter fibrils that are usually found in bundles or in association with elastic fibers . Individual fibrillin microfibrils are long and can be extended in vivo when tissues are under tension [9] . In humans and mice with MFS , fibrillin microfibril bundles were fragmented in the skin [10] , [11] . In contrast , fibrillin microfibrils in human scleroderma skin were disorganized , when labeled and examined by immunoelectron microscopy [12] . The latter observation was extended by ultrastructural studies of fibrillin microfibrils in SSKS [8] . In SSKS , fibrillin microfibrils were found in large aggregates within which individual microfibrils appeared to be short [8] . These electron microscopic observations suggest that structural abnormalities in fibrillin microfibrils may underlie the differences between MFS and SSKS disease pathologies . Another possibility is that mutations causing SSKS or WMS perturb growth factor signaling , since fibrillin-1 targets and sequesters the large latent Transforming Growth Factor β ( TGFβ ) complex [13] , [14] as well as multiple Bone Morphogenetic Proteins ( BMPs ) and Growth and Differentiation Factor-5 ( GDF-5 ) [15]–[17] . In MFS , abnormal activation of TGFβ signaling contributes to phenotypes in the lung [18] and aorta [19] , but TGFβ signaling may not be abnormally activated in the skin . In SSKS or WMS skin , overproduction of collagen may be predicted to be due to abnormal activation of TGFβ . But , it is unclear why abnormal activation of TGFβ signaling would be limited to the skin in SSKS or WMS and alternatively to the lung and aorta in MFS . Different mechanisms may be involved in the activation of TGFβ signaling in these disorders and/or important unknown factors may limit the effects of mutations in fibrillin-1 to specific tissues . Here we identify a novel mutation in fibrillin-1 in a family with WMS . In order to reveal the complex mechanisms by which fibrillin-1 differentially regulates connective tissues , we replicated this mutation causing human WMS in the mouse . We show that this mutation does not cause MFS , since the WMS mouse survives normally and does not display major features of MFS , even in homozygosity . Instead , WMS mutant mice develop skin fibrosis associated with distinctive ultrastructural abnormalities in fibrillin microfibrils . In addition , WMS mice demonstrate retardation of long bone growth . Therefore , WMS mice recapitulate cardinal features of human WMS . To elucidate molecular mechanisms of WMS , we provide evidence that the WMS mutation abolishes the binding site in fibrillin-1 for a novel family of proteins , the ADAMTSLIKE ( ADAMTSL ) proteins . We also connect ADAMTSL proteins with ADAMTS-10 and propose that these proteins form a complex with fibrillin-1 . These findings implicate ADAMTSL proteins , together with ADAMTS-10 , in the regulation of fibrillin microfibril structure . Insights gained from these studies are relevant to MFS and to the expanding genetic diseases that constitute the Marfan-related disorders . Furthermore , our results point out the importance of fibrillin-1 in the local regulation of tissue-specific microenvironments .
A family with autosomal dominant WMS was previously described , and linkage analysis identified FBN1 on chromosome 15q21 . 1 as the disease locus [2] . This family includes affected individuals in three generations ( Figure 1a ) . Affected individuals exhibited characteristic features of WMS including microspherophakia , ectopia lentis , glaucoma , brachydactyly , short stature , and thickening of the skin , as previously documented for individuals 5016 , 4084 , and 5010 [2] . We further examined individuals 5010 ( 45 years of age ) , 5011 ( 15 years of age ) , and 6013 ( 10 years of age ) and unaffected family members . Early removal of ocular lenses and short stature were common . Mild brachydactyly of the toe and some limitation of small joints were found in 6013 . 5011 showed brachydactyly , more pronounced limitation of large and small joints , and thickened skin . 5010 showed severe limitation of small and large joints with pain and loss of dexterity and thick forearm skin without striae . In addition , all three individuals showed increased truncal and axial muscle bulk . No history or clinical evidence of valvular cardiac or aortic disease was found in this family . Southern blotting of genomic DNA and PCR followed by DNA sequencing revealed a heterozygous 7895 nt genomic deletion in FBN1 ( Figure S1a and Figure 1b ) with boundaries in introns 8 and 11 . PCR results from unaffected and affected family members demonstrated that the mutation segregated with the disease ( Figure 1b ) . Transcripts from the mutant allele lacked exons 9–11 ( Figure 1b ) , predicting in-frame translation of fibrillin-1 molecules in which the first 8-cysteine domain , the proline-rich region , and EGF-like domain 4 are missing ( Figure 1c ) . No similar mutations have been reported in the FBN1 mutation database , where only 23 of 1 , 013 mutations were deletions or insertions [20] . By using ribonuclease protection assay , we were able to show that the wildtype and the mutant FBN1 allele were equally expressed ( Figure S1b ) . Using a quantitative sandwich ELISA , we found that wildtype and mutant fibrillin-1 proteins were equally secreted by affected WMS fibroblasts ( Figure 1d ) . Immunofluorescence of skin from an affected individual ( 5011 ) showed fibrillin fibrils that appeared normal and not fragmented ( Figure 1e ) , unlike the fragmented fibrils observed in MFS skin [10] . In contrast to the numerous different mutations in FBN1 known to cause MFS , there is only one report of an FBN1 mutation in a family with autosomal dominant WMS [3] . There are also reports of individuals with WMS that overlap with MFS [6] . In order to test whether the three-domain deletion in fibrillin-1 found in our family with WMS causes WMS and not MFS , we replicated the mutation in a mouse model ( WMΔ ) using a gene targeting strategy ( Figure 2a ) . WMΔ heterozygous ( WMΔ/+ ) and homozygous ( WMΔ/WMΔ ) mice breed well and are viable . Both heterozygous and homozygous mutant mice live longer than 1 . 5 years with no signs of aortic disease typical of MFS . Aortic root morphology in heterozygous and homozygous mutants is normal , even at 10 months of age ( Figure 2b and Figure S2 ) , in contrast to heterozygous and homozygous mutant mouse models of MFS [11] , [19] , [21]–[23] . In addition , with the exception of the mutant mgR/mgR [21] , which is hypomorphic for normal Fbn1 and dies during early adulthood , homozygous mutant mouse models of MFS die in the early postnatal period [11] , [22] , [23] . By these two major criteria for MFS in mice—aortic disease and early death of homozygotes—WMΔ mice do not model MFS . Brachydactyly and short stature are features of WMS , while arachnodactyly and tall stature are characteristic of MFS . Therefore , long bones in the WMΔ mutant mice were measured . Growth of long bones appeared to be normal in the first two weeks of postnatal life but was reduced by 3–4 weeks of age in homozygous mice ( WMΔ/WMΔ ) . Measurements of the long bones at 1 month of age in the WMΔ/WMΔ mice using μCT showed a statistically significant ( P<0 . 05 ) reduction in lengths of the radius , ulna , and tibia of 6–10% ( Figure 2c ) compared to age and gender matched wildtype controls . At 3–4 weeks of age , length measurements of metacarpals and proximal and distal phalanges in fore- and hindpaws were also reduced between 2–23% in the WMΔ/+ and WMΔ/WMΔ mice ( Figure 2d ) . These findings are consistent with the WMS phenotype . However , by 5 months of age , these differences in length were normalized . Gross examination suggested a thickened , less elastic skin in WMΔ mutant mice ( Figure 3a ) . Histology of skin biopsies from WMΔ mice showed excessive collagen deposition in the dermis starting at 1 month of age . Hematoxylin and eosin or Masson's Trichrome stains ( Figure 3b , 3c ) revealed a widened dermal layer with decreased hypodermal fat , and thicker , more densely packed collagen fibers in mutants compared to wildtype mice . qPCR analyses demonstrated upregulated expression of collagen genes in the skin from mutant mice ( Figure 3d ) . Skin thickness , as determined by histological stains and detected by touch by 7 months of age , persisted through old age in the mutant mice . Electron microscopy after immunogold labeling with anti-fibrillin-1 antibodies showed alterations in fibrillin microfibril ultrastructure in WMΔ/+ and WMΔ/WMΔ skin: large bundles of microfibrils as well as microfibrils around elastin cores showed reduced periodicity of immunogold labeling in the mutants ( Figure 4a , 4b ) . In addition , large accumulations of microfibrils were prominent in WMΔ/+ and WMΔ/WMΔ skin ( Figure 4a ) , and elastic fibers appeared moth-eaten compared with wildtype littermates ( Figure 4b ) . The disorganized appearance of microfibrils is better visualized in the three-dimensional aligned tilt series of immunolabeled microfibrils from WMΔ/WMΔ and wildtype skin samples supplied Videos S1 and S2 . While some areas of the skin appeared normal , electron microscopic examination of skin from an 18 year old individual with WMS ( unrelated to the WMS family described above ) revealed unusually large abnormal aggregates of microfibrils ( visible at low magnification in Figure 4c , left panel , arrows ) . Elastic fibers also appeared moth-eaten ( data not shown ) . Similar to observations of WMΔ mutant mouse skin ( Figure 4a ) , immunogold labeling with antibodies specific for fibrillin-1 demonstrated both irregular labeling of the microfibril aggregates and periodic labeling of apparently normal microfibrils ( Figure 4c , middle panel ) . Based on these immunolocalization results with antibodies specific for fibrillin-1 , we conclude that the small and large aggregates are composed of abnormal bundles of fibrillin-1 microfibrils . Large microfibril aggregates similar to those in human WMS ( Figure 4c , left and middle panels ) were found in skin from older ( 11–20 month old ) WMΔ/WMΔ mice ( Figure 4c , right panel ) . These results provide evidence for a common pathogenetic mechanism for fibrosis in human WMS and in this mouse model of WMS . We previously showed that ADAMTSL-6 interacts with the N-terminal half of fibrillin-1 with high affinity ( KD = 80 nM ) [24] . In order to test whether other ADAMTSL family members also bind to fibrillin-1 and whether the binding site utilized by ADAMTSL-6 is perturbed by the WMS three-domain deletion in fibrillin-1 , we generated recombinant fibrillin-1 polypeptides , rF84 and rF84WMΔ ( Figure S3 ) and rF90 and rF90WMΔ ( Figure 1c ) , as well as recombinant human ADAMTSL-1 , -2 , -3 , and mouse papilin polypeptides ( Figure S3 ) . Surface plasmon resonance ( SPR ) technology was employed to measure interactions between fibrillin-1 and ADAMTSL polypeptides . Similar to ADAMTSL-6 [24] , ADAMTSL-2 , -3 , and papilin polypeptides interacted with the N-terminal half of fibrillin-1 , while ADAMTSL-1 did not . Binding to the C-terminal half of fibrillin-1 was negative for all ADAMTSL proteins tested ( data not shown ) . SPR sensorgrams are shown for ADAMTSL-2 binding to fibrillin-1 and for ADAMTSL-3 binding to fibrillin-1 ( Figure 5a ) . ADAMTSL-2 , -3 , and -6 and papilin polypeptides did not bind to recombinant fibrillin-1 polypeptides with the WMS three-domain deletion ( Figure 5a and Table S1 ) . Binding constants for all of these interactions were calculated from the SPR data ( Table S1 ) . We conclude that the fibrillin-1 domains consisting of the first 8-cysteine domain , the proline-rich region , and the 4th generic EGF-like domain contain the ADAMTSL binding site ( s ) . Because recessive WMS is caused by mutations in ADAMTS10 [4] , [5] , FBN1 and ADAMTS10 share a genetic pathway . Therefore , we hypothesized that fibrillin-1 , ADAMTS-10 , and some ADAMTSL proteins form protein complexes . In a pull-down assay with Ni-NTA as a resin , we showed that full-length , His6-tagged ADAMTS-10 in media of stably transfected EBNA 293 cells bound to the N-terminal half of fibrillin-1 ( Figure 5b ) . From SPR interaction studies , a KD of 450 nM was calculated for the binding of the N-terminal fibrillin-1 polypeptide to the C-terminal end of ADAMTS-10 ( data not shown ) . The C-terminal recombinant ADAMTS-10 polypeptide used in the SPR studies represents the noncatalytic region of ADAMTS-10 , a region composed primarily of Tsp1 repeats ( Figure S3 ) . SPR also showed that the C-terminal end of ADAMTS-10 interacted with the C-terminal end of ADAMTSL-3 with high binding affinity ( KD = 2 nM ) ( Figure 5c ) . However , neither ADAMTSL-2 nor -1 bound to ADAMTS-10 , indicating that ADAMTS enzymes may partner only with specific ADAMTSL proteins . Taken all together , these results suggest that direct interactions between fibrillin-1 , ADAMTS-10 , and specific ADAMTSL proteins are involved in the pathogenesis of WMS . Based on in vitro studies , we hypothesized that the mutant WMS fibrillin-1 cannot interact properly in vivo with certain members of the ADAMTSL family of proteins . To test if localization of ADAMTSL proteins is altered in WMΔ mutant mice , we stained skin with antibodies specific for ADAMTSL-6 [24] . Results showed a reduction in ADAMTSL-6 immunofluorescence in skin from WMΔ/+ and WMΔ/WMΔ mutant mice compared to wildtype littermate skin ( Figure 6a ) . Fibrillin-1 immunofluorescence was equal in pattern and abundance in WMΔ mutant and wildtype mice ( data not shown ) . Since antibodies specific for ADAMTSL-2 and -3 are not yet available , we were unable to determine whether these proteins also colocalize with fibrillin-1 in skin and whether these are also reduced in WMΔ mutant mice . ADAMTSL-6 and ADAMTS-10 promote fibrillin-1 fibril formation [24] , [25] . Therefore , we examined human and mouse fibroblasts for defects in fibrillin-1 fibril formation . The Marfan cell culture assay [10] was used . In this assay , control fibroblasts assemble a matrix containing abundant immunofluorescent fibrillin-1 fibrils , while MFS fibroblasts assemble only few or no fibrils [10] . Unlike MFS fibroblasts , fibroblasts from a member ( 5010 ) of the WMS family described here showed abundant immunofluorescent fibrillin-1 fibrils ( Figure 6b ) . However , these fibrils appeared to be much thinner and less bundled than control ( CRL2418 ) fibrillin-1 fibrils . Fibroblasts from wildtype , heterozygous and homozygous WMΔ littermates also showed abundant immunofluorescent fibrillin-1 fibrils ( Figure 6c ) . Immunofluorescence staining of WMΔ/+ and WMΔ/WMΔ cultures suggested thicker bundles of fibrillin-1 fibrils than those in wildtype cultures ( Figure 6c ) . These results underscore the conclusion that the WMS mutation in fibrillin-1 works mechanistically differently than other FBN1 mutations that cause MFS . In addition , results suggest that the mutant WMS fibrillin-1 causes defects in fibrillin-1 fibril aggregation or bundling , but observed differences between the human WMS and the mouse WMΔ cultures cannot currently be explained . Another potential mechanism contributing to pathogenesis of WMS is abnormal activation of TGFβ signaling . Our findings of upregulated collagen genes in the skin of WMΔ mice and increased Trichrome staining of WMΔ dermis ( Figure 3 ) suggested increased TGFβ signaling . However , Western blotting for pSmad 2/3 showed no differences between wildtype and WMΔ skin at multiple time points ( data not shown ) , and qPCR quantitation of TGFβ-responsive genes such as Ctgf , Pai1 , and Postn also demonstrated no differences at multiple time points ( data not shown ) . When total ( Figure 7a ) and active TGFβ ( Figure 7b ) were measured in the medium of human cultured fibroblasts , no significant difference was found between control and WMS fibroblasts . Skin samples from heterozygous and homozygous WMΔ mice of different ages were examined for the presence of myofibroblasts . Staining with antibodies specific for α-smooth muscle actin did not reveal increased numbers of myofibroblasts in mutant WMΔ mice ( Figure 7c ) . We also tested interactions between Latent TGFβ Binding Proteins ( LTBPs ) and ADAMTSL proteins , since an interaction between ADAMTSL-2 and the middle region of LTBP-1 was previously identified [26] . SPR binding studies ( summarized in Table S2 ) showed no interaction between ADAMTSL-2 and the middle region of LTBP-1 ( rL1-M [13] ) . However , binding between ADAMTSL-2 and the C-terminal domains of LTBP-1 present in rL1-K [13] was detected . Binding between ADAMTSL-3 and the C-terminal domains of LTBP-1 was also detected , but neither ADAMTSL-2 nor -3 interacted with LTBP-4 .
WMS is considered to be clinically homogenous , even though the genetic basis of WMS is heterogeneous [27] . Both recessive and dominant forms of WMS present equally with myopia , glaucoma , cataract , short stature , brachydactyly , thick skin , and muscular build . There may be significant differences in incidence of microspherophakia , ectopia lentis and joint limitations between the recessive and dominant forms [27] , but these features are also common to both . Currently , there is a single report of a mutation in FBN1 in a family with dominant WMS [3] . Results presented here identify a novel mutation in FBN1 in a family with dominant WMS , which was previously linked to FBN1 [2] . The mutation is predicted to result in fibrillin-1 molecules that lack the first 8-cysteine domain , the proline-rich region , and the adjacent EGF-like domain . Replication of this mutation in mouse Fbn1 clearly demonstrated that the mutation reproduces at least one cardinal feature of WMS—thick skin—and does not cause the clinical equivalent of MFS . Reduced long bone growth was also found in WMΔ mice , consistent with the human WMS traits of short stature and brachydactyly . However , by 5 months of age , long bone growth was normalized , perhaps reflecting differences between humans , in whom growth plate closure occurs at skeletal maturity , and rodents , whose growth plates fail to close [28] , and who grow for a longer period of time than humans using cellular processes which are not active in adult humans [29] . In addition , hypermuscularity , a feature of human WMS , is also found in WMΔ mutant mice ( data not shown ) . Biochemical studies comparing wildtype and WMS mutant fibrillin-1 revealed that the WMS mutation abolished a specific binding site in fibrillin-1 for ADAMTSL-2 , -3 , -6 and papilin . Further biochemical studies suggested that specific ADAMTSL proteins may interact with ADAMTS-10 and that ADAMTS-10 binds to fibrillin-1 . Our results are consistent with previous studies of papilin , the first of the ADAMTSL proteins to be described , and procollagen N-proteinase ( now called ADAMTS-2 ) , which indicated that the “papilin cassette” ( homologous to the noncatalytic regions of ADAMTS enzymes ) may interact with ADAMTS metalloproteinases [30] . In addition , binding between ADAMTS-10 and fibrillin-1 was recently demonstrated [25] . Therefore , we propose that a ternary complex of ADAMTSL , ADAMTS-10 , and fibrillin-1 may be formed . Such a ternary complex is depicted in Figure 8 , showing how ADAMTSL-3 and ADAMTS-10 may bind as a complex to fibrillin-1 in microfibrils . Mutations in ADAMTSL2 cause geleophysic dysplasia [26] . Mutations in ADAMTS17 cause a Weill-Marchesani-like syndrome [31] , and mutations in ADAMTSL4 cause autosomal recessive isolated ectopia lentis [32] . Both geleophysic dysplasia and WMS are acromelic dysplasias sharing features of short stature , brachydactyly , thick skin , limited joint mobility , and hypermuscularity . Ectopia lentis is a common feature of MFS and WMS . All together , these related genetic disorders suggest that specific ADAMTSL ( at least ADAMTSL-2 and -4 ) and ADAMTS ( ADAMTS-10 and -17 ) proteins modulate fibrillin-1 function in the skeleton , skin , joints , muscle , and eye . Our biochemical data also implicate ADAMTSL-3 and -6 in these pathways . Whether all members of the ADAMTSL/ADAMTS family perform similar roles in the modulation of fibrillin-1 function is unknown . However , if similar functions are performed , differences in temporal and spatial regulation of the expression of these genes could account for tissue-specific variation in these related disorders . An in-frame deletion of 24 nucleotides was found in FBN1 to cause autosomal dominant WMS [3] . This mutation ( 5074_5097del ) deletes 8 amino acid residues ( R1692 – Y1699 ) from the fifth 8-cysteine domain ( also called TB5 ) in fibrillin-1 . When fibrillin-1 is modeled within microfibrils [33] , the fifth 8-cysteine domain in one molecule of fibrillin-1 is close to the ADAMTSL binding site in an adjacent fibrillin-1 molecule ( Figure 8 ) . Recently , 16 novel heterozygous mutations in FBN1 causing geleophysic dysplasia or acromicric dysplasia were also identified in the fifth 8-cysteine domain [34] . WMS , geleophysic dysplasia , and acromicric dysplasia are members of the acromelic group of dysplasias with similar as well as distinctive clinical features . In our model , the clustering of fibrillin-1 domains associated with acromelic dysplasias and with binding sites for ADAMTSL proteins involved in acromelic dysplasias may specify a new microenvironment controlling thick skin and musculoskeletal growth . It is interesting that , when fibrillin-1 is modeled within microfibrils [33] , the single RGD-containing domain in fibrillin-1 is close to the ADAMTSL binding site in fibrillin-1 ( Figure 8 ) . Integrin binding to RGD sites is known to perform important roles in matrix assembly [35] and to be critically dependent on the surrounding sequences , which can silence RGD function [36] . SSKS is caused by missense mutations in FBN1 exon 37 , which encodes the domain containing the RGD site [8] . Therefore , it can be speculated that SSKS mutations in FBN1 lead to diminished integrin activity . Because dermal fibrosis and the abnormal ultrastructural appearance of fibrillin microfibrils were similar in both SSKS [8] and WMS ( Figure 4 ) , it seems likely that integrin interactions with fibrillin-1 are perturbed in both SSKS and WMS . Furthermore , the proximity of the RGD-containing domain to the ADAMTSL binding site in fibrillin-1 suggests that integrins may cooperate with ADAMTSL proteins and ADAMTS-10 in modulating fibrillin microfibril [24] , [25] aggregation and organization . However , the molecular mechanisms of this cooperation remain unknown . Abnormal TGFβ signaling may play a role in these disorders , since fibrillin-1 microfibrils target and sequester the large latent TGFβ complex [13] . We found upregulated collagen gene expression and increased Trichrome staining in the skin of WMΔ mutant mice , results that are consistent with activated TGFβ signaling . In addition , molecular interactions of LTBP-1 with both ADAMTSL-2 and ADAMTSL-3 were determined , suggesting that loss of the ADAMTSL binding site in WMΔ mutant mice might render the large latent TGFβ complex more susceptible to activation . However , if activation of TGFβ signaling underlies the dermal fibrosis in WMΔ mutant mice , this activation of signaling did not manifest detectable differences in other conventional TGFβ signaling readouts ( e . g . , increased α-smooth muscle actin positive cells ) . It has been recently speculated that mechanical forces may be required to activate the latent TGFβ complex [37] . Therefore , it is possible that local changes in the fibrillin microfibril matrix could influence force-dependent activation of latent TGFβ , perhaps leading to a local increase in signaling . Activation of TGFβ signaling has been shown for geleophysic dysplasia [26] , [34] , acromicric dysplasia [34] , and for MFS [18] , [19] . We propose that , in consort with the different tissue-specific manifestations of disease in WMS and MFS , activation of TGFβ signaling in these diseases may be limited ( WMS ) or more global ( MFS ) in scope . In MFS , the broad activation of TGFβ signaling in multiple tissues matches the pleiotropic features of the disease and the requirement for general pathogenetic mechanisms initiated by multiple different disease-causing mutations . In the case of WMS , we propose that fibrosis limited to the skin is due to the dysregulated interaction between abnormally organized microfibrils and the large latent TGFβ complex within dermal microenvironments . Our data suggest that direct interactions between ADAMTSL proteins , fibrillin-1 , and LTBP-1 ( Figure 8 ) may be dysregulated in WMS , leading to concomitant structural and signaling abnormalities within local spaces . However , the discrepancy in the data measuring TGFβ activity between our WMS fibroblast cultures and geleophysic and acromicric dysplasia fibroblast cultures [26] , [34] is not yet understood . Although further investigations are required in order to determine the roles of ADAMTSL/ADAMTS-10 complexes , integrins , and LTBP-1 in the fine regulation of TGFβ signaling in WMS skin fibrosis , we conclude that these molecular pathways work locally—in microenvironments—to control skin fibrosis . While the importance of the microenvironment is appreciated in development and cancer [38] , this is to our knowledge the first evidence for microenvironmental regulation by fibrillin-1 . In summary , our results suggest an improved concept of the architectural and regulatory functions of fibrillin-1 . Previously , the microfibrils of elastic , distensible tissues were thought to function mechanically only as a limiting component for a cross-linked , isotropic elastin matrix . Subsequently , the attachment of LTBPs and BMPs demonstrated that fibrillin microfibrils participate in the storage and release of growth factors . Now we show that fibrillin-1 also selectively binds the metalloproteinase ADAMTS-10 and some non-enzymatic ADAMTSL proteins , enabling a clustering of these protein complexes in the vicinity of the fibrillin-1 RGD site and suggesting the potential for integrin involvement in ADAMTS/ADAMTSL/fibrillin functions . The established mutual affinity of the protein components of this cluster opens varied biochemical pathways that need to be explored in the future . The genetic evidence in humans and mice shows that perturbation of such biochemical pathways can lead to significant pathobiological consequences . In addition , the genetic evidence clearly demonstrates that fibrillin-1 microfibrils , although ubiquitous structural elements in the connective tissue space , perform local functions to support tissue microenvironments . From the perspective of normal development and tissue homeostasis , we propose that the fibrillin microenvironment may enable two-way communication between a cell and its surroundings . The extended fibrillin fibril may function as a sensor for mechanical distortion of the matrix , signaling the cell when there is a need for additional , reinforcing structural components like collagens . The presence of large latent TGFβ complexes within the fibrillin microenvironment conveniently couples matrix mechanics with available signals for upregulation of collagens . Installation of new structural materials into pre-existing matrices likely requires remodeling enzymes like metalloproteinases . ADAMTS enzymes , localized to the fibrillin microenvironment as well , possibly with the help of ADAMTSL proteins , could serve such purposes , or they might participate in the activation of nearby latent growth factors . Understanding pathogenetic mechanisms underlying WMS and MFS will elucidate the local , fine adjustments required for growth , homeostasis , and repair .
Clinical studies were performed with informed consent and local OHSU Internal Review Board approval . All mouse work was approved by the OHSU IACUC committee . The family pedigree shown in Figure 1 has been previously described [2] . Dermal fibroblast cultures were established from punch biopsies obtained with informed consent and local OHSU IRB approval . All participants were evaluated for myopia , glaucoma and dislocated lenses , for musculoskeletal and skin characteristics , and for cardiac or aortic disease assessed by history and/or by auscultation . This family has been followed for more than 15 years with no clinical evidence of valvular cardiac or aortic disease . A second unrelated individual , designated WMS2 , was seen who had been previously diagnosed with WMS . At 18 years old , she had a history of early high myopia and presented with headaches secondary to glaucoma . She had short stature , mild brachydactyly , microspherophakia , and apparently normal joints and skin . A punch biopsy was obtained with informed consent . Genomic DNA was extracted from cultured WMS 5016 or normal skin fibroblasts ( NSF ) or EDTA whole blood using standard procedures . Individual FBN1 exons were amplified by PCR of genomic DNA using intronic primers . Overlapping FBN1 cDNAs were amplified by PCR using exonic primers ( for sequences see Table S3 ) . PCR products were sequenced . Genomic DNA was digested ( using HindIII , Bsu36I , NcoI , SpeI , SspI ) , separated electrophoretically and transferred to a nylon membrane . The membrane was probed with a 727 bp cDNA fragment of FBN1 encompassing exons 8–12 , radiolabeled with 32P-αdCTP in the presence of random and specific primers , and exposed after stringent washing to film for autoradiography . Primers flanking FBN1 exons 9 , 11 , 21 , and 37 were used to generate fragments of genomic NSF DNA , and the products were cloned into the pGEM-T-Easy vector ( Promega ) . Radiolabeled probes were hybridized to total RNA of normal or WMS patient fibroblasts , followed by ribonuclease digestion to degrade unhybridized regions ( RPA III kit , Ambion ) . Protected fragments were separated by acrylamide gel electrophoresis , and quantitated by phosphorimager ( STORM , Molecular Diagnostics ) . All materials used for the generation of the WMΔ mouse line originated from C57BL/6 mice ( see Figure 4 for design of targeting vector ) . The floxed WMΔ mouse line was generated by Ozgene Pty . Ltd . ( Bentley , Australia ) . The neomycin selection cassette was removed by breeding targeted mice to FLPe mice . Cre-mediated removal of Fbn1 exons 10–12 in all cells was accomplished by breeding floxed WMΔ mice to mice containing Cre-recombinase knocked into the Rosa26 locus ( on a C57Bl/6 background ) . For this study , heterozygous WMΔ mice were bred to yield wildtype , heterozygous , and homozygous littermates for analyses . Genotyping was by PCR using primer pairs annealing within and outside of the deleted genomic region ( for sequences see Table S4 ) . All procedures performed on mice were approved by OHSU IACUC . Fibrillin-1 polyclonal antibody ( pAb 9543 ) and monoclonal antibodies ( mAb ) 15 , 78 , 201 , and 69 have been previously described [9]–[12] , [33] . Polyclonal antibody specific for ADAMTSL-6 was generated as described [24] . Antibody to α-smooth muscle actin was purchased from Sigma . CRL2418 , a normal dermal fibroblast cell line , was purchased from American Type Culture Collection . WMS fibroblasts were established from a punch biopsy of skin from family member 5010 . Explant cultures of P4 mouse skin were established from WMΔ wildtype , heterozygous and homozygous littermates . 1 ml chamber slides were seeded at a density of 200 , 000 cells/ml and incubated in DMEM , including 10% fetal bovine serum , for 3 to 10 days , as indicated in the figures . Media from the 3-day incubation was collected and stored at −20°C for sandwich ELISAs . Cell layers were analyzed by immunofluorescence . 96-well ELISA plates ( Corning ) were coated with 100 µl of 5 µg/ml streptavidin ( Pierce ) and incubated overnight at 4°C . Excess streptavidin was removed by extensive washing , and 100 µl/well of 0 . 25 µg/ml biotinylated monoclonal antibodies ( B15 or B201 ) were incubated at 25°C for 1 hour . After washing , wells were incubated overnight at 4°C with culture medium samples ( from 3-day chamber slide cultures of WMS or control fibroblasts ) . In addition , serially diluted protein standards ( rF11 ) were applied separately to wells coated with biotinylated antibodies and incubated overnight at 4°C . Unbound proteins were removed by washing , and alkaline phosphatase-conjugated monoclonal antibodies ( AP201 0 . 05 µg/ml or AP78 0 . 5 µg/ml ) were incubated in the wells for 1 hour at 25°C . Invitrogen's ELISA amplification system was used for colorimetric detection , according to the manufacturer's protocol . Absorbance was recorded using a Molecular Devices Emax plate spectrophotometer and was then converted to µg/ml , according to standard curve values . Calculations to determine concentration were performed on Excel software . Skin was obtained by punch biopsy from a son of family member 5010 , following informed consent . Skin was also obtained from WMΔ wildtype and mutant mouse littermates in accordance with OHSU approved IACUC procedures . Immunofluorescence of skin as well as cultured fibroblasts was performed as previously described [10] , [11] . Histology was performed by the OHSU Histology Core , using standard procedures for Hematoxylin and Eosin and Masson Trichrome stains ( Sigma , St . Louis , MO ) . For μCT analyses , mice were sacrificed at specified time points . μCT measurements and analyses were performed with a Scanco μCT 35 ( Scanco Medical , Basserdorf , Switzerland ) scanner , according to the manufacturer's instructions . qPCR using RNA from WMΔ and wildtype control mouse skin was performed as previously described [11] . Primers for mouse Col1A1 , Col1A2 , and Col3A1 were purchased from SABiosciences ( Frederick , MD ) . The primers for mouse Periostin ( Postn; forward: 5′-catcttcctcagcctccttg-3′; reverse: 5′-tcagaagctccctttcttcg-3′ ) , Plasminogen activator inhibitor-1 ( Pai1; forward: 5′-ctttacccctccgagaatcc-3′; reverse: 5′-gacacgccatagggagagaa-3′ ) , and Connective tissue growth factor ( Ctgf; forward: 5′-ctgcctaccgactggaagac-3′; reverse: 5′-ttggtaactcgggtggagat-3′ ) were individually designed and tested for amplification efficiency . Immunoelectron microscopy of tissues from WMΔ mouse littermates was performed as described [11] . Tissues were labeled en bloc with anti-fibrillin-1 ( pAb 9543 ) followed by 5 nm secondary gold conjugated antibodies ( Amersham Biosciences ) . Aligned tilt series were acquired from 500 nm thick sections as described [11] . Expression vectors carrying full length human LTBP-4S and LTBP-1S were kindly provided by Dr . Jorma Keski-Oja and Dr . Daniel Rifkin . The rF90WMΔ and rF84WMΔ expression constructs were cloned from WMS fibroblast cDNA . Full-length ADAMTSL1 was obtained from human fibroblast cDNA . For cloning of ADAMTSL2 , a mouse full length cDNA clone ( ID RIKEN cDNA F83011122 ) was obtained . Constructs for ADAMTSL3 were made using a clone ( RIKEN ) and mouse lung cDNA . Constructs for Papilin were cloned from mouse fibroblast cDNA . A full length human ADAMTS10 clone ( SC309981 ) was purchased from Origene , and mutations in this clone were corrected . Generation of recombinant polypeptides representing fragments of LTBP-1 was previously described [13] . The generation of rF90 was described before [16] . All fibrillin and ADAMTSL expression constructs were transfected into 293/EBNA cells for protein expression . All proteins were purified using metal ion affinity chromatography . Protein domain boundaries for the constructs are depicted in Figure 1 and Figure S3 . Binding analyses were performed using a BIAcoreX ( BIAcore AB , Uppsala , Sweden ) . Recombinant full length ADAMTSL-1 , -2 , LTBP-1 , -4 , and polypeptides ADAMTS-10 C-term , rL1K , rLM , rLN , rF6 , rF90 , and rF90WMΔ were covalently coupled to CM5 sensor chips ( research grade ) using the amine coupling kit following the manufacturer's instructions ( BIAcore AB ) . Binding assays were performed at 25°C in 10 mM Hepes buffer , pH 7 . 4 , containing 0 . 15 M NaCl , 3 mM EDTA , and 0 . 005% ( v/v ) P20 surfactant ( HBS-EP buffer , BIAcore AB ) . Kinetic constants were calculated by nonlinear fitting of association and dissociation curves ( BIAevaluation 3 . 0 software ) . Equilibrium dissociation constants ( KD ) were then calculated as the ratio of kd/ka . Cell culture media ( 1 ml ) from stably transfected 293/EBNA cells expressing ADAMTS-10 with a C-terminal His6-tag , and media from untransfected EBNA cells as a control were adjusted to 20 mM Tris pH 7 . 8 , 5 mM imidazole and incubated for 1 h with 5–20 µg of rF11 ( N-terminal half of fibrillin-1 , without a His6-tag ) . Subsequently , 50 µl of a 50% Ni-NTA slurry in water was added and incubated for 2 hs . The resin was washed and boiled in 50 µl 1× SDS loading buffer . Eluted proteins were subjected to SDS-PAGE followed by immunoblotting with polyclonal anti fibrillin-1 antibody 9543 . The quantity of TGF-β1 in 100 µl culture medium from confluent fibroblasts ( 200 , 000 cells/ml grown for 72 h in 1 ml chamber slides ) was determined using the TGF-β1 EMax Immnunoassay kit ( Promega ) . WMS and control fibroblasts were utilized . Prism 5 . 02 for Windows ( GraphPad , San Diego , CA ) was used to perform One-way Analysis of Variance ( 1-way ANOVA ) followed by post-test analysis with Tukey's multiple comparison test . p-values<0 . 05 were considered significant .
|
The microenvironment is specified by cell-surface molecules , growth factors , and the extracellular matrix . Here we report genetic evidence that implicates fibrillin-1 , a ubiquitous extracellular matrix molecule that sequesters latent growth factor complexes , as a key determinant in the local control of musculoskeletal and skin microenvironments . A novel mutation in fibrillin-1 demonstrates that modulation of the fibrillin microfibril scaffold can influence tissue microenvironments and result in the clinical features of Weill-Marchesani syndrome ( WMS ) , including thick skin , short stature , and brachydactyly . Dysregulated WMS microenvironments diverge from Marfan pathogenetic mechanisms , which lead to broad activation of TGFβ signaling in multiple tissues .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"biochemistry",
"histology",
"dermatology",
"clinical",
"genetics",
"model",
"organisms",
"ophthalmology",
"genetics",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics",
"cardiovascular"
] |
2012
|
Microenvironmental Regulation by Fibrillin-1
|
Theileria annulata is an apicomplexan parasite that infects and transforms bovine macrophages that disseminate throughout the animal causing a leukaemia-like disease called tropical theileriosis . Using deep RNAseq of T . annulata-infected B cells and macrophages we identify a set of microRNAs induced by infection , whose expression diminishes upon loss of the hyper-disseminating phenotype of virulent transformed macrophages . We describe how infection-induced upregulation of miR-126-5p ablates JIP-2 expression to release cytosolic JNK to translocate to the nucleus and trans-activate AP-1-driven transcription of mmp9 to promote tumour dissemination . In non-disseminating attenuated macrophages miR-126-5p levels drop , JIP-2 levels increase , JNK1 is retained in the cytosol leading to decreased c-Jun phosphorylation and dampened AP-1-driven mmp9 transcription . We show that variation in miR-126-5p levels depends on the tyrosine phosphorylation status of AGO2 that is regulated by Grb2-recruitment of PTP1B . In attenuated macrophages Grb2 levels drop resulting in less PTP1B recruitment , greater AGO2 phosphorylation , less miR-126-5p associated with AGO2 and a consequent rise in JIP-2 levels . Changes in miR-126-5p levels therefore , underpin both the virulent hyper-dissemination and the attenuated dissemination of T . annulata-infected macrophages .
Theileria annulata is an apicomplexan parasite causing a widespread disease called tropical theileriosis that is endemic to North Africa , the Middle East , vast parts of India and China [1] . The parasite can infect bovine B cells , but in the natural environment , predominantly infects macrophages . Theileria-infected leukocytes are transformed into tumour-like leukemias that display uncontrolled proliferation and increased ability to disseminate and invade organs and tissues [2 , 3] . As the transformed state of Theileria-infected leukocytes can be reversed by drug ( buparvaquone ) -induced parasite death , molecular events related to Theileria-induced host cell transformation have been proposed to have an epigenetic basis [4] . Since attenuated vaccine lines used to fight tropical theileriosis are derived by long-term culture of virulent infected macrophages , and since many virulence traits can be restored by TGF-β2–stimulation of attenuated macrophages [5] , loss of Theileria-infected macrophage virulence such as activation of the Activator Protein 1 ( AP-1 ) transcription factor [6 , 7] may also have an epigenetic element [8 , 9] . Micro-ribonucleic acids ( miRNAs ) are small ( 17–25 bases long ) single-stranded , non-coding RNAs [10 , 11] that modulate diverse biological processes by normally binding to the 3′-untranslated region ( 3’-UTR ) of target mRNAs , thus altering the post-transcriptional regulation of numerous genes [12–14] . However , miRNA can also bind to 5'-UTRs , introns and coding sequence of mRNA [15–17] . Post-transcriptional control of gene expression by miRNAs is increasingly recognized as a central part of host/pathogen interactions [18 , 19] . The role of miRNAs in bacterial [20 , 21] , viral [22] and protozoan [23] infections is now well established . A role for miR-155 in the virulence of T . annulata-infected leukocytes occurs via its suppression of De-Etiolated Homolog 1 ( DET-1 ) expression that diminishes c-Jun ubiquitination [9] . In order to obtain a global view of all bovine miRNAs expressed in different types of T . annulata-infected leukocytes and how they might contribute to parasite-induced leukocyte transformation and tumour dissemination we determined the miRNomes of infected macrophages ( both virulent and attenuated ) and infected B cells ( TBL20 and TBL3 ) versus non-infected B cells ( BL20 and BL3 ) . Defining the miRNomes of different types of T . annulata-transformed leukocytes allowed us to observe that infection alters the expression of many known miRNAs . However , to identify miRNAs implicated in Theileria-transformed leukocyte dissemination , as opposed to immortalisation , we queried our datasets only for miRNAs whose expression is activated by infection , but dampened upon loss of transformed macrophage dissemination . Here , we characterise the role of miR-126-5p in T . annulata-induced leukocyte transformation and attenuation of infected macrophage dissemination . The mixed lineage kinase dual leucine zipper kinase-1 ( DLK1 ) is an established target of miR-126-5p [24] . DLK1 can selectively regulate the JNK-based stress response pathway via its interaction with the scaffolding protein JIP to form a specialized JNK signalling complex [25] . JIP-1 and JIP-2 bind selectively to JNK , but not to other related MAP kinases including p38 [26 , 27] and their over-expression causes cytoplasmic retention of JNK; thereby preventing its nuclear translocation and its ability to phosphorylate specific nuclear substrates such as c-Jun [26] . Moreover , under basal conditions DLK1 is bound to JIP , but upon stimulation , DLK1 dissociates from JIP resulting in JNK translocation to the nucleus [28] . Within the nucleus Jun and Fos family members form homo- and hetero-dimers as part of the AP-1 transcription factor that binds to specific DNA sequences to drive target gene expression; a notable example being mmp9 . The matrix metallo-proteinase MMP9 is associated with metastasis [29] and AP-1 is implicated in various cellular processes such as apoptosis [30] , growth control [31] and cellular transformation [32] . Upon activation , JNK translocates to the nucleus to phosphorylate c-Jun that is a key transcription factor in the virulence-associated hyper-dissemination phenotype of Theileria-transformed leukocytes [2 , 7] . Importantly , upon attenuation of Theileria-infected macrophage dissemination JNK activity decreases resulting in reduced c-Jun phosphorylation and decreased AP-1-driven transcription of mmp9 [33] . We now demonstrate that T . annulata infection of B cells and macrophages leads to the up-regulation of miR-126-5p that ablates JIP-2 expression liberating cytosolic JNK1 to translocate to the nucleus and phosphorylate c-Jun . Conversely , in attenuated macrophages miR-126-5p levels drop , JIP-2 complexes reform retaining JNK in the cytosol leading to reduced nuclear c-Jun phosphorylation , dampened AP-1-driven transcription of mmp9 and reduced traversal of matrigel . Thus , high miR-126-5p levels contribute to Theileria-transformed leukocyte dissemination and reduced miR-126-5p levels contribute to attenuation of the virulent hyper-dissemination phenotype .
To identify miRNAs whose expression is altered upon infection by T . annulata we determined the miRNomes of both infected versus non-infected B cells and virulent versus attenuated macrophages . The comparison between T . annulata-infected TBL20 B lymphocytes and their uninfected counterparts revealed potential involvement of many miRNAs in the transformation of the host cells , as reflected by the changes in their expression levels ( Fig 1A ) . We analyzed the differential expression ( DE ) of our miRNA sequencing data using DESeq2 [34] . 115 miRNAs are differentially expressed in TBL20 as compared to BL20 with a cutoff of adjusted p value < 0 . 05 . In order to increase the confidence level and limit our analysis to the most significantly DE miRNAs we used a second pipeline , baySeq [35] . Both DESeq2 and baySeq are highly specific and sensitive tools for the detection of differential expression [36] . We consider a miRNA differentially expressed ( DE ) following two criteria: a ) fold change ( FC ) greater than 2 and b ) adjusted p value less than 0 . 05 ( DESeq2 ) and FDR less than 0 . 1 ( baySeq ) . Finally , we compared the lists of up- and down-regulated miRNAs from both DESeq2 and baySeq and retained only miRNAs identified as DE in both pipelines ( S1 Table ) . In addition to TBL20 , we characterize the DE miRNAs in a different B cell line: TBL3 . Similarly , we identified the DE miRNAs in infected TBL3 as compared to their uninfected counterparts , BL3 , using the same pipelines and criteria ( S2 Table ) . The comparison between the lists of DE miRNAs in TBL20 and TBL3 shows that there are 19 common differentially expressed miRNAs: 9 up- and 10 down-regulated . We followed the expression of these 19 DE miRNAs in virulent compared to attenuated macrophages ( Fig 1B ) . This identified miR-126-5p as a miRNA upregulated after T . annulata infection of B lymphocytes in 2 independent cell lines ( TBL20 and TBL3 ) and down-regulated in attenuated macrophages that have lost their hyper-disseminating phenotype . As expected for transformed leukocytes the biological functions of the DE miRNAs are annotated as being associated with “oncogenesis” , with the exception of miR-6526 and miR-30f that are not well characterized . The reported functions of the DE miRNAs are therefore consistent with the cancer-like phenotype of T . annulata-infected leukocytes . To confirm the sequencing results , we randomly selected 10 DE miRNAs , 5 up- and 5 down-regulated and verified their expression qRT-PCR ( Fig 1C , left ) . All tested miRNAs , including miR-126-5p , confirmed the miRNA sequencing data for DE .
Unlike T . parva , the causative agent of East Coast Fever that results from infection and transformation of T and B cells , T . annulata-transformed macrophages lose their virulent hyper-disseminating phenotype following long-term culture , and attenuated macrophages with diminished dissemination are used as live vaccines against tropical theileriosis . For these reasons , we used next generation sequencing ( NGS ) to profile the miRNomes of non-infected B cells , T . annulata-infected B cells and virulent versus attenuated infected macrophages . The different miRNomes allowed us to compare the miRNA expression of B cells before and following infection and concomitant with loss of T . annulata-infected macrophage virulence . To focus on miRNAs of particular relevance to parasite-induced leukocyte tumour virulence we asked that expression of a given miRNA be upregulated by infection yet downregulated in attenuated macrophages that have lost their hyper-disseminating phenotype . These criteria identified miR-126-5p as a prime candidate and led to our characterization of its contribution to the transformed phenotype of T . annulata-infected B cells and macrophages . We demonstrated that JIP-2 is a novel miR-126-5p target gene and that infection by increasing miR-126-5p levels suppresses JIP-2 expression in virulent macrophages . Loss of JIP-2 released cytosolic JNK1 to translocate to the nucleus and phosphorylate c-Jun , contributing to constitutive AP-1-driven MMP production that is characteristic of Theileria-induced leukocyte dissemination . By contrast , in attenuated macrophages , where miR-126-5p expression is reduced , augmented JIP-2 retains JNK1 in the cytosol leading to decreased nuclear c-Jun phosphorylation , ablated MMP9 production and dampened traversal of matrigel . Thus , miR-126-5p-provoked reduction in JIP-2 levels activates JNK1>AP-1 signalling and provides an epigenetic explanation for both T . annulata-induced leukocyte transformation , and for the attenuated phenotype of live vaccines against tropical theileriosis . By demonstrating that infection-induced miR-126-5p expression ablates JIP-2 and diminishes the cytosolic localisation of JNK1 we provide a mechanism that contributes to constitutive c-Jun phosphorylation , increased MMP9 production and a greater capacity of Theileria-transformed leukocytes to disseminate ( Fig 7 ) . miR-126 is located within the 7th intron of the EGFL7 gene [37–39] and EGFL7 is equivalently expressed in virulent and attenuated macrophages ( S1 Fig ) . In T . annulata-infected macrophages miR-126-5p levels therefore do not depend on the degree of EGFL7 expression [48] , nor on the amount of precursor miR-126 , rather infection impacts on the capacity of AGO2 to load miR-126-5p , where it’s protected from degradation , while miR-126-3p is not loaded and is consequently degraded . In virulent macrophages Grb2 recruits PTP1B to de-phosphorylate AGO2 that facilitates uptake of miR-126-5p , whereas in attenuated macrophages the amount of PTP1B associated with AGO2 diminishes with a concomitant increase in AGO2 phosphorylation and decrease in bound miR-126-5p ( Fig 7 ) . Inflammation stemming from T . annulata infection likely explains induction of EGFL7 and pre-miR-126 expression , but why miR-126-5p , rather than miR-126-3p , is loaded onto AGO2 is unknown and will animate future studies . Finally , given that miR-126-5p is deregulated in many cancers; reagents that manipulate miR-126-5p levels could be discussed as tools for cancer therapy .
Cells used in this study are T . annulata-infected Ode macrophages [49] , where virulent macrophages used correspond to passage 62 and attenuated macrophages to passage 364 . The different B cell lines used were non-infected immortalized B sarcoma lines ( BL3 and BL20 ) and T . annulata-infected BL3 ( TBL3 ) and BL20 ( TBL20 ) cells , and all have been previously described [50–53] . All cells were incubated at 37°C with 5% CO2 in Roswell Park Memorial Institute medium ( RPMI ) supplemented with 10% Fetal Bovine Serum ( FBS ) , 2mM L-Glutamine , 100 U penicillin , 0 . 1mg/ml streptomycin , and 4- ( 2-hydroxyethyl ) -1-piperazineethanesulfonic acid ( HEPES ) and 5% 2-mercapthoethanol for BL20 and TBL20 . Total RNA of Theileria-infected leukocytes was isolated using the miRNA isolation kit ( #AM1560 , Thermo fisher scientific , Villebon , France ) according to the manufacturer’s instructions . Total RNA designated for miRNA experiments was extracted using the mirVana miRNA isolation kit ( Thermo Fisher ) using the manufacturer’s instructions . The quantity and quality checked by Qubit and Bioanalyzer 2100 , respectively . The miRNA libraries were prepared using the Illumina Truseq Small RNA Sample Preparation kit ( RS-200-0012 ) following the manufacturer’s instructions . Briefly , 1 following the manufactupter ligated at the 3’ and 5’ ends , reverse transcribed , barcoded then amplified with 11 cycles of PCR amplifications . Then the cDNA was run on a 6% TBE PAGE gel ( Novex , Thermo Fisher ) . After staining with SYBR Green the gel is visualized on a UV transluminator ( Doc-It imaging system , UVP ) and the cDNA constructs of a size between 145–160 bp were cut out and eluted from the gel , concentrated and the libraries validated , quantified and finally pooled and sequenced on a Hiseq 2000 and Hiseq 4000 . Raw miRNAseq reads are quality checked by fastqc [54] . mirTools2 . 0 [55] pipeline is used for trimming ( “Adaptor_trim . pl” script ) and downstream analysis of known miRNAs . Sequencing reads are aligned to the bosTau7 genome using SOAP [56] . Annotations are added from miRBase 21 [57] and Rfam [58] databases . The differential expression of each known miRNAs from their absolute read counts are analysed by DESeq2 [34] . Differential expression is tested based on the negative Binomial distribution and miRNAs with adjusted p-value < 0 . 05 considered as statistically significant . Bovine genomic DNA of mapk8 ( jip2 ) was purified from Theileria-infected macrophages and used as a template for PCR . For construction of the jip2 , the 3′-UTRs of bovine JIP2 was amplified using the following primers: Forward: GGCCCTCGAGAGCAGAAAGTTTATTGAGGTGCT Reverse: GGCCGCGGCCGCTGGGGTCGGAACTGGGAG The PCR-amplified fragments were digested by XhoI and NotI and inserted in the psiCHECK-2 vector . Cells were transiently transfected with 2 μg of firefly/renilla luciferase reporter plasmid . Protein extracts were prepared using the Passive Lysis Buffer provided in the Dual-Luciferase Assay ( Promega ) . Equal amounts of protein extracts were plated into a 96-well plate . Firefly luciferase activity was measured for 12 seconds using the LB 960 luminometer ( Berthold Technologies , Thoiry , France ) . To assess the internal standard activity , Stop and Glo reagent was added ( Promega ) , and the peak of the renilla luciferase activity was then measured . Normalized relative luciferase units ( RLU ) were then calculated as firefly luciferase units of protein extracts of treated or untreated cells divided by renilla luciferase units of protein extracts of untreated cells . Data represent the mean ± SEM of three independent experiments , each performed in duplicate . Cells were transfected by electroporation using the Nucleofector system ( Lonza , Basel , Switzerland ) . 2 . 5x105 cells were suspended in 100μl of nucleofector solution mix with 2μg of plasmids and subjected to electroporation using the cell line—specific programme: T-O17 . After transfection , cells were suspended in fresh complete medium and incubated at 37°C , 5% CO2 for 24 h and RNA extracted after 48 h post transfections . Measurements of luciferase and β-galactosidase activities were performed using the Dual Light Assay system ( Thermo Fisher scientific ) and luminometer Centro LB 960 ( Berthold ) according to the manufacturer’s instructions . RNA extracted with the mirVana kit was used . cDNA was synthesized using the miScript II RT kit ( Qiagen ) following the manufacturer’s instructions . Briefly , a 20 μl was set for each biological replicate of each tested miRNA . The RT contained 8 μl MMIX ( 4 μl of 5x miScript HiFlex Buffer , 2 μl of 10x miScript Nucleics Mix and 2 μl miScript Reverse Transcriptase Mix ) , 500 ng of total RNA in RNase-free water . The RT was performed at 37°C for 60 min and 95°C for 5 min . The qPCR was performed in a 96-well plate as a 25 μl volume containing 2μl of RT product , 12 . 5 μl of 2x QuantiTect SYBR Green PCR Master Mix , 2 . 5 μl of 10x miScript Precursor Assay and RNase-free water . The qPCR thermal cycle was set to 95°C for 15 min and 40 cycles of 94°C for 15 secs , 55°C for 30 secs and 70°C for 30 sec . Data were analysed using the 2−ΔΔCT , or the relative expression method by normalization to HRPT ( ENSBTAT00000019547 . 4 ) as a reference gene . RNA extracted with the mirVana kit was used . cDNA was synthesized using the TaqMan microRNA RT kit ( Thermo Fisher scientific ) following the manufacturer’s instructions . Briefly , a 15 μl was set for each biological replicate of each tested miRNA . The RT contained 7 μl MMIX ( 100 mM dNTPs , 50 U/μl MultiScribe reverse transcriptase , 10X RT buffer , 20 U/μl RNase inhibitor ) , 10 ng of total RNA in 5 μl and 3 μl of 5X miRNA-specific RT primers . The RT was performed at 16°C for 30 min , 42°C for 30 min and 85°C for 5 min . The qPCR was performed in a 384well plate as a 10 μl volume containing 1 . 33 μl of RT product , 5 μl TaqMan 2X universal PCR MMIX , 1 μl of miRNA-specific 20X TaqMan MicroRNA assay . The qPCR thermal cycle was set to 95°C for 10 min and 40 cycles of 95°C for 15 secs and 60°C for 60 sec . Data was analysed using the 2−ΔΔCT , or the relative expression method by normalization to U6b as a reference gene for miRNA . bta-miRNA primers were purchased from Thermo Fisher Scientific . Total RNA of Theileria-infected leukocytes was isolated using the RNeasy mini kit ( Qiagen ) , according to the manufacturer’s instructions . The quality and quantity of RNA were measured by Bioanalyzer 2100 and Qubit , respectively . For reverse transcription , 1μg isolated RNA was diluted in water to a final volume of 12 μl , warmed at 65°C for 10 min , then incubated on ice for 2 min . Afterwards , 8 μl of reaction solution ( 0 . 5 μl random hexamer , 4 μl 5x RT buffer , 1 . 5 μl 10mM dNTP , 1 μl 200U/μl RT-MMLV ( Promega , Charbonnières-les-Bains , France ) and 1 μl 40U/μl RNase inhibitor ( Promega ) was added to get a final reaction volume of 20 μl and incubated at 37°C for 2 h . The resultant cDNA was stored at -20°C . mRNA expression levels were estimated by qPCR on Light Cycler 480 ( Roche , Meylan , France ) using SYBR Green detection ( Thermo Fisher Scientific ) . The detection of a single product was verified by dissociation curve analysis and relative quantities of mRNA calculated using the method described by [59] . gapdh was used as reference gene to normalize for mRNA levels . The specificity of PCR amplification was confirmed by melting curve analysis . Sequence primers used are as follows: gapdh: FO 5’-AGGACAAAGCTCAGGGACAC-3’ , Rev 5’- CCCCAGGTCTACATGTTCCA-3’ mmp9: FO 5’-CCCATTAGCACGCACGACAT-3’ , Rev 5’- TCACGTAGCCCACATAGTCCA-3’ dlk1: FO 5’- ATGGGCATCGTCTTCCTCAA -3’ , Rev 5’- CAGGATGGTGAAGCAGATGG -3’ jip2: FO 5’- TCTTCCCTGCCTTCTATGCC -3’ , Rev 5’- CAGGTGGACGGTCAGTTT -3’ For total cell extraction , cells were harvested and extracted by lysis buffer ( 20mM Hepes , Nonidet P40 ( NP40 ) 1% , 0 . 1% SDS , 150mM NaCl , 2mM EDTA , phosphatase inhibitor cocktail tablet ( PhosSTOP , Roche ) and protease inhibitor cocktail tablet ( Complete mini EDTA free , Roche ) ) . For cytoplasmique extraction , cells were harvested and extracted by lysis buffer ( HEPES [10 mM] pH 7 . 9 , KCl [10 mM] , EDTA [0 . 1 mM] , NP-40 0 . 3% , protease inhibitors 1x , protease and phosphatase inhibitor cocktail ) . For nuclear extraction , cell pellets were lysed and extracted by lysis buffer ( HEPES [20 mM] pH 7 . 9 , NaCl [0 . 4 M] , EDTA [1mM] , Glycerol 25% and protease Inhibitors 1x . Protein concentration was determined by the Bradford protein assay [60] . Cell lysates were subjected to Western blot analysis using conventional SDS/PAGE and protein transfer to nitrocellulose filters ( Protran , Whatman ) . The membrane was blocked by 5% non-fat milk-TBST ( for anti-DLK , anti-JIP-2 , anti-c-JUN , anti-JNK ) , or 3% non-fat milk-PBST ( for anti-actin antibody ) for 2 h at room temperature ( RT ) . Antibodies used in immunoblotting were as follows: goat polyclonal antibody anti-JIP-2 ( Santa Cruz Biotechnologies , Heidelberg , Germany # sc-19740 ) , rabbit polyclonal antibody anti-JIP-2 ( Abcam # ab-154090 ) , rabbit polyclonal antibody anti-DLK ( Santa Cruz Biotechnologies # sc-25437 ) , rabbit polyclonal antibody anti-JNK ( Santa Cruz Biotechnologies # sc-571 ) , goat polyclonal anti-PTP1B ( Santa Cruz Biotechnologies # sc-1718 ) , mouse monoclonal antibody anti-AGO2 ( Abcam # ab-57113 ) , rabbit monoclonal anti-AGO2 ( Cell signalling # 2987 ) , mouse anti-phospho tyrosine antibody ( Transduction laboratories # P1120 ) , goat anti-GST antibody ( GE Healthcare # 27-4577-01 ) and goat polyclonal antibody anti-actin ( Santa Cruz Biotechnologies I-19 ) . After washing , proteins were visualized with ECL western blotting detection reagents ( Thermo Scientific ) . The β-actin level was used as a loading control throughout all experiments . Co-immunoprecipitations and GST-Grb2 pull down assay were conducted with protein extracts of Theileria-infected macrophages . JIP-2 , AGO2 , GST-Grb2 and PTP1B precipitates were transferred to western blots and probed respectively with a rabbit polyclonal anti-DLK , mouse monoclonal anti-AGO2 , mouse anti-phospho tyrosine , rabbit anti-PTP1B ( Abcam #ab88481 ) and goat anti-PTP1B antibodies . Normal IgG was used as a negative control , cell lysates from virulent and attenuated macrophages were treated with IgG and the whole cell lysate without IP was included as positive control . 1×105 cells were centrifuged on glass slide with the Cellspin I ( Tharmac ) at 1500 rpm for 10 min and fixed by 4% paraformaldehyde for 10–15 min at room temperature . Fixed cells were permeabilized by 0 . 2% Triton X-100 for 10 min and blocked with 1% BSA for 30 min . These cells were incubated with primary antibodies against Ser-phospho-73 c-Jun ( 1/200 , Santa Cruz Biotechnologies #sc-7981 ) overnight , sequentially stained with secondary antibodies conjugated with Alexa 488 ( 1/1000 , Molecular Probes ) for 45 min at room temperature . Stained cells were mounted in ProLong Diamond Antifade Mountant with DAPI ( Thermo Fisher Scientific ) . Acquisitions were made by inverted microscopy ( Leica DMI6000s ) with metamorphous software . Images were taken at x100 magnification . The invasive capacity of Theileria-infected macrophages and B cells were assessed in vitro using matrigel migration chambers [7] . Culture coat 96-well medium BME cell invasion assay was obtained from Culturex instructions ( 3482-096-K ) . Fifty thousand cells were added to each well and after 24 h of incubation at 37°C , each well of the top chamber was washed once in buffer . The top chamber was placed back on the receiver plate . 100 μl of cell dissociation solution/Calcein AM were added to the bottom chamber of each well , incubated at 37°C for 1 h to fluorescently label cells and dissociate them from the membrane before reading at 485 nm excitation , 520 nm emission using the same parameters as the standard curve . Data were analysed with the Student’s t-test . All values are expressed as mean+/-SEM . Values were considered to be significantly different when two-sided p values were < 0 . 05 . The miRNA expression dataset has been assigned series record GSE97706 in the GEO repository . The work was conducted under the approval number 15IBEC11 by the Institutional Biosafety and Ethics Committee ( IBEC ) in KAUST .
|
Theileria annulata-infected bovine macrophages lose their hyper-disseminating virulent phenotype during long-term culture and are used as attenuated live vaccines to fight tropical theileriosis . Deep microRNA sequencing revealed that infection of both B cells and macrophages alters the expression of a large number of host cell microRNAs . We focused on miR-126-5p as its expression was induced by infection , but diminished in attenuated macrophages that had lost their disease causing disseminating phenotype . We show that miR-126-5p in virulent macrophages directly targets and suppresses a cytosolic scaffold protein called JNK-Interacting Protein-2 ( JIP-2 ) , so liberating JNK1 to enter the nucleus and phosphorylate c-Jun . This activates AP-1-driven transcription of mmp9 that promotes tumour dissemination . In virulent macrophages , an adaptor protein called Grb2 recruits the tyrosine phosphatase PTP1B to AGO2 so decreasing AGO2 phosphorylation to increase miR-126-5p levels . By contrast , in attenuated macrophages AGO2 tyrosine phosphorylation increases and miR-126-5p levels drop leading to a regain in JIP-2 expression that retains JNK1 in the cytosol . This leads to decreased nuclear c-Jun phosphorylation and reduced mmp9 production . Thus , variations in miR-126-5p levels underpin both virulent hyper-dissemination and attenuation of T . annulata-transfected macrophages .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"transfection",
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"chemical",
"compounds",
"gene",
"regulation",
"immunology",
"organic",
"compounds",
"micrornas",
"tyrosine",
"immunoprecipitation",
"amino",
"acids",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"aromatic",
"amino",
"acids",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"gene",
"expression",
"chemistry",
"molecular",
"biology",
"precipitation",
"techniques",
"biochemistry",
"rna",
"antibody-producing",
"cells",
"cell",
"biology",
"nucleic",
"acids",
"b",
"cells",
"post-translational",
"modification",
"organic",
"chemistry",
"genetics",
"hydroxyl",
"amino",
"acids",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"macrophages",
"non-coding",
"rna"
] |
2018
|
miR-126-5p by direct targeting of JNK-interacting protein-2 (JIP-2) plays a key role in Theileria-infected macrophage virulence
|
Resistance and susceptibility to Leishmania major infection in the murine model is determined by the capacity of the host to mount either a protective Th1 response or a Th2 response associated with disease progression . Previous reports involving the use of cysteine cathepsin inhibitors indicated that cathepsins B ( Ctsb ) and L ( Ctsl ) play important roles in Th1/Th2 polarization during L . major infection in both susceptible and resistant mouse strains . Although it was hypothesized that these effects are a consequence of differential patterns of antigen processing , the mechanisms underlying these differences were not further investigated . Given the pivotal roles that dendritic cells and macrophages play during Leishmania infection , we generated bone-marrow derived dendritic cells ( BMDC ) and macrophages ( BMM ) from Ctsb−/− and Ctsl−/− mice , and studied the effects of Ctsb and Ctsl deficiency on the survival of L . major in infected cells . Furthermore , the signals used by dendritic cells to instruct Th cell polarization were addressed: the expression of MHC class II and co-stimulatory molecules , and cytokine production . We found that Ctsb−/− BMDC express higher levels of MHC class II molecules than wild-type ( WT ) and Ctsl−/− BMDC , while there were no significant differences in the expression of co-stimulatory molecules between cathepsin-deficient and WT cells . Moreover , both BMDC and BMM from Ctsb−/− mice significantly up-regulated the levels of interleukin 12 ( IL-12 ) expression , a key Th1-inducing cytokine . These findings indicate that Ctsb−/− BMDC display more pro-Th1 properties than their WT and Ctsl−/− counterparts , and therefore suggest that Ctsb down-regulates the Th1 response to L . major . Moreover , they propose a novel role for Ctsb as a regulator of cytokine expression .
Worldwide , 2 million new cases of leishmaniasis occur every year . It is endemic in 98 countries , where 350 million people are considered to be at risk , largely affecting “the poorest of the poor” [1] , [2] . The cutaneous form of leishmaniasis is characterized by lesions that heal over the course of months or years , and leave permanent scars that can be disfiguring or disabling [1] , [2] . Control of Leishmania within the host is mediated by innate and adaptive immune responses . Experimental mouse models of Leishmania major infection first documented the relevance of Th1/Th2 polarization for resistance and susceptibility to the disease in vivo [3] . In resistant mouse strains , such as C57BL/6 , a contained and self-healing development of the disease appears , mediated by a protective Th1 immune response . Th1 cells secrete IFN-γ , which induces the nitric oxide ( NO ) -mediated killing of the amastigote form of the parasite within phagosomes in macrophages . In contrast , infection of BALB/c mice with L . major causes a non-healing Th2 form of the disease , characterized by expression of the cytokines IL-4 , IL-13 , and IL-10 . The key role of dendritic cells ( DC ) in inducing cell-mediated immune responses against leishmaniasis has been extensively documented [4] , based on their capacity to migrate to draining lymph nodes after capture of Leishmania parasites and to induce Th cell polarization . Several subsets of DC have been reported to perform this function , including Langerhans cells [5] , dermal DC [6] , lymph node resident DC [7] , and monocyte-derived DC [8] , [9] . In order to instruct Th cell polarization , DC use three main signals: ( 1 ) antigen presentation via MHC class II molecules , ( 2 ) the expression of co-stimulatory molecules and ( 3 ) cytokine secretion . Quantitative and qualitative differences in these signals are crucial for Th cell polarization [10] . Among these signals , IL-12 is a key cytokine for the development of a protective Th1 immune response . Neutralization of IL-12 by antibodies leads to susceptibility to Leishmania infection in otherwise resistant mice [11] , [12] . Conversely , treatment of BALB/c mice with IL-12 resulted in a protective Th1 response [13] . DC have been reported to be the primary source of IL-12 in lymphoid tissues [14] , with variations depending on the DC subset , maturation status , and whether promastigotes or amastigotes are used [15] . Macrophages , on the other hand , are considered as main host cells for Leishmania parasites , where freshly inoculated promastigotes find a niche for differentiating into amastigotes and proliferating . Macrophages are not able at all to produce IL-12 in response to L . major [16] , [17] , even if they are further stimulated with lipopolysaccharide ( LPS ) [18] , [19] , reflecting the extent of the silencing that L . major induces in its host . Silencing of infected cells has been attributed to different virulence factors . Some of them are cysteine proteases [20] , which impair NF-κB signaling in macrophages [21] and are also important for autophagic and differentiation processes in the parasite [22] . Therefore , they are interesting targets for drug development [23] , [24] . However , they also have homologs in mammals . Few studies have addressed the effects that unspecific inhibition of host cathepsins would have on the immune response against L . major . Maekawa et al . reported that treatment of L . major-infected BALB/c mice with the Ctsb inhibitor CA074 triggered a protective Th1 immune response [25] , [26] . Treatment of these mice with the Ctsl inhibitor CLIK148 , on the other hand , caused a stronger Th2 response [27] , even in resistant mouse strains [28] . The authors showed that the inhibitors had no direct effect on the proliferation of the parasite but that the host cell cathepsins were inhibited , and hypothesized that lack of Ctsb or Ctsl would lead to different patterns for proteolytic processing of L . major antigens . It had remained unclear , however , how the inhibition of Ctsb and Ctsl activity could have such effects in Th polarization . Thus , further investigation is needed to understand the involvement of Ctsb and Ctsl in the immune response during leishmaniasis . In the present study , we used cathepsin B ( Ctsb−/− ) - and cathepsin L ( Ctsl−/− ) -deficient bone marrow-derived DC ( BMDC ) and bone marrow-derived macrophages ( BMM ) to determine the role of these proteases for the signals that DC use to instruct Th cell polarization in response to L . major promastigotes: the expression of MHC class II and co-stimulatory molecules , and the production of cytokines . Furthermore , we studied the effect of these cathepsins on parasite proliferation in infected cells . Our results indicate that in this model of infection , cathepsin B plays a significant role not only in the expression of antigen-presenting MHC class II molecules , but also in the regulation of IL-12 production .
Complete RPMI medium was prepared by supplementing RPMI 1640 medium ( with phenol red or phenol red-free , as indicated; Invitrogen , Darmstadt , Germany ) with heat-inactivated fetal calf serum ( FCS , 10% v/v; PAA Laboratories , Pasching , Austria ) , L-glutamine ( final concentration 2 mM; Biochrom AG , Berlin , Germany ) , HEPES ( pH 7 . 2 , 0 . 01 M; Invitrogen , Darmstadt , Germany ) , penicillin G ( 0 . 2 U/ml; Sigma-Aldrich , Taufkirchen , Germany ) , gentamicin ( 0 . 05 mg/ml; Sigma-Aldrich ) , and 2-mercaptoethanol ( 0 . 05 mM; Sigma-Aldrich ) . In addition , for generation of BMM , a conditioned medium was used containing Dulbecco's Modified Eagles Medium ( DMEM; Invitrogen ) , heat-inactivated FCS ( 10% v/v; PAA Laboratories ) , heat-inactivated horse serum ( 0 . 5%; Invitrogen ) , 2-mercaptoethanol ( 0 . 05 mM; Sigma-Aldrich ) , nonessential amino acids ( Invitrogen ) , HEPES ( 0 . 01 M; Invitrogen ) , L-glutamine ( 4 mM; Biochrom ) and L929 supernatant ( 15% v/v ) . L . major promastigotes were cultured in a biphasic medium consisting of a solid base of rabbit-blood agar ( Elocin-lab , Gladbeck , Germany ) plus a liquid phase of RPMI medium without phenol red . BMDC and BMM were generated from bone marrow progenitors as described previously [23] , [29] , [30] from female BALB/c , C57BL/6 , C57BL/6 Ctsb−/− and C57BL/6 Ctsl−/− mice ( 6–12 weeks old ) . The generation of Ctsb−/− and Ctsl−/− mice has been described previously [31]–[33] . Briefly , total bone marrow cells were flushed from femurs and tibiae . To generate DC , the cell number was determined by trypan blue staining , and 0 . 2×106 cells/ml bone marrow cells were cultured in complete RPMI 1640 medium in the presence of recombinant murine granulocyte-macrophage colony-stimulating factor ( GM-CSF , 0 . 04 µg/ml; Invitrogen ) at 37°C , 5% CO2 . Cultures were fed with complete RPMI medium supplemented with GM-CSF on days 3 and 6 . At day 8 , the non-adherent cells were collected , washed with complete RPMI medium and resuspended at 2×106 cells/ml in complete RPMI medium . BMM were generated by culturing 0 . 67×106 cells/ml of total bone marrow progenitors as described above in conditioned DMEM at 37°C and 5% CO2 . On day 6 , the culture medium was removed carefully and replaced with cold RPMI complete medium , and the petri dishes were kept on ice for 10 min . Thereafter , the macrophages were removed with a cell scrapper , washed with fresh complete medium without phenol red , and resuspended at 2×106 cells/ml . As quality control , the morphology of the obtained BMDC and BMM was analyzed . Part of the cells was used for cytospin preparations stained with Diff-Quik II dye ( Medion Diagnostics , Düdingen , Switzerland ) according to the manufacturer's instructions , and observed under the light microscope . Furthermore , the expression of the phenotypic markers CD11c in DC and F4/80 in macrophages was assessed by flow cytometry as described below . The morphology of BMDC and BMM from WT , Ctsb−/− and Ctsl−/− mice was additionally analyzed by TEM . Samples of the obtained cells were prepared for TEM using OsO4 and uranyl acetate as contrasting agents , following the protocol previously described by Schurigt et al . [24] . The L . major isolate MHOM/IL/81/FE/BNI was maintained by continuous passage in BALB/c mice , and promastigotes were grown in vitro in blood-agar cultures as described previously [34] at 27°C , 5% CO2 and 95% humidity . In order to preserve maximal infectivity , only promastigotes passaged 5 to 8 times were used for in vitro infection experiments . In addition , two different transgenic L . major strains were used in some experiments: a luciferase-transgenic strain ( Luc-tg ) previously described [30] , and an eGFP-transgenic strain ( eGFP-tg ) . For the preparation of L . major antigen ( LmAg ) , stationary-phase WT promastigotes were washed three times in phosphate-buffered saline ( PBS ) , resuspended at 1×109/ml in PBS , and subjected to three cycles of freezing in liquid nitrogen and thawing . The aliquots were stored at −80°C and each aliquot was thawed not more than twice . Heat-killed parasites ( HK ) were prepared by incubating a parasite suspension of 1×109/ml in RPMI medium for 30 min at 80°C . The enhanced-green fluorescent protein ( eGFP ) -coding region was cut from pEGFP-N1 ( Clontech , Saint-Germain-en-Laye , France ) by BamHI-NotI ( Promega , Mannheim , Germany ) and cloned into the Bglll-NotI-restricted Leishmania expression vector pLEXSY-hyg2 ( Jena Bioscience , Jena , Germany ) , which contains a marker gene for selection with hygromycin . The generated plasmids were linearized by SwaI ( New England Biolabs , Frankfurt , Germany ) , and the parasites were transfected by electroporation . eGFP and HYG were integrated into the 18S rRNA locus of L . major by homologous recombination . For in vitro experiments , promastigotes were grown in blood-agar cultures supplemented with 50 µg/ml hygromycin under the same conditions as WT and Luc-tg L . major promastigotes . In order to maintain their virulence , eGFP-tg L . major parasites were passaged in female BALB/c mice . The stability of the integrated eGFP without further selection by hygromycin was assessed in vitro and in vivo by flow cytometry . Intracellular amastigotes in BMM from Ctsb−/− and Ctsl−/− mice , and their WT C57BL/6 counterparts , were measured with the method described by Bringmann et al . [30] . Briefly , 200 µl of a 2×105 cells/ml suspension of BMM in phenol red-free complete medium were seeded into 96-well plates with clear bottoms ( Greiner Bio-One , Frickenhausen , Germany ) and were incubated for 4 hours to allow cell adhesion . The medium was then removed , and 200 µl of a 3×106 cells/ml suspension of Luc-tg L . major promastigotes were added at an infection ratio of 1∶15 and incubated for 24 hours at 37°C , 5% CO2 . Any remaining extracellular parasites were eliminated by washing 3 times with medium , and 200 µl of the phenol red medium were added . After further 24 hours incubation at 37°C , 5% CO2 , 50 µl of the luciferin-containing lysis buffer Britelite Plus ( PerkinElmer , Waltham , USA ) were added to each well . The plate was incubated in the dark for 5 min at room temperature ( RT ) , and the resulting luminescence was measured as counts per second ( CPS ) , with a Victor×Light 2030 luminometer ( PerkinElmer ) . 5×105 BMM from WT , Ctsb−/− and Ctsl−/− mice were seeded in duplicates into chambered cover glasses , in a final volume of 250 µl of complete medium without phenol red , and incubated at 37°C , 5% CO2 for 4 hours to promote cell adhesion . The culture medium was removed , replaced by an equivalent volume of eGFP-tg L . major promastigotes at an infection ratio of 1∶15 , and the cells were further incubated for 24 hours at 37°C , 5% CO2 . The cells were then washed 3 times with warm PBS; part of the cells were incubated with Hoechst solution 0 . 5% v/v ( Immunochemistry Technologies , Bloomington , USA ) for 15 min at 37°C protected from the light , followed by washing 3 times with warm PBS and addition of 250 µl of complete medium . Then , they were observed under a fluorescence microscope ( Leica Microsystems ) . The rest of the cells were incubated in fresh medium for further 24 hours , stained and observed under the fluorescence microscope as described above . The amount of cells and L . major bodies were quantified with the Cell Counter plug-in from the ImageJ software [35] , and the average parasite count per infected cell was calculated as a geometric mean ( G ) using the formula , where represents the sequence of parasites counted for every infected cell . 1×106 BMDC/ml were harvested at day 7 of culture , plated in 6-well plates and incubated overnight at 37°C , 5% CO2 . For some experiments , BALB/c BMDC were pre-incubated with 10 µM CA074Me ( Bachem , Bubendorf , Switzerland ) , 10 µM CLIK148 ( kindly provided by Prof . Tanja Schirmeister ) , or 10 µM of Z-Arg-Leu-Arg-α-aza-glycyl-Ile-Val-OMe ( ZRLR , kindly provided by Dr . Timo Burster , University of Ulm , and Dr . Ewa Wieczerzak , University of Gdansk ) for 4 hours prior to infection . eGFP-L . major promastigotes were harvested , washed 3 times in warm PBS , added to the BMDC at a 1∶5 infection ratio , and further incubated at 37°C . After 2 hours of exposure of the BMDC to the parasites , the cells were washed with warm PBS and resuspended in fresh medium at a concentration of 1×106 cells/ml . Part of the cells was fixed in paraformaldehyde ( PFA , 4%; Applichem , Darmstadt , Germany ) . The remaining cells were incubated for a total of 4 hours or 24 hours post infection , fixed , and the amount of infected cells at the different time points was determined by flow cytometry , together with the expression of maturation markers as described next . BMDC infected with e-GFP L . major or stimulated with LmAg ( 30 µl LmAg/ml , equivalent to 30 parasites per BMDC ) were fixed with 4% PFA and resuspended in FACS buffer containing the following antibodies ( Ab ) : phycoerythrin-cyanine 7 ( PECy7 ) –conjugated anti-CD11c ( BD Biosciences , Heidelberg , Germany ) , phycoerythrin ( PE ) -conjugated anti-CD86 ( BD Biosciences ) , allophycocyanin ( APC ) -conjugated anti-MHC class II ( Miltenyi , Bergisch Gladbach , Germany ) . For some assays , BMDC were infected with WT L . major promastigotes instead , and fluorescein isothiocyanate ( FITC ) -conjugated anti-CD40 ( Biolegend , San Diego , USA ) and FITC-conjugated anti-CD80 ( eBioscience , San Diego , USA ) Ab were used . Data was obtained using the MACSQuant flow cytometer ( Miltenyi ) and analyzed using FlowJo ( Tree Star Inc . , CA , USA ) . The expression of F4/80 in BMM was determined using FITC-conjugated anti-F4/80 Ab ( Biolegend ) . The expression of intracellular IL-12 was analyzed in BMDC after 24 hours of stimulation with LPS ( 1 µg/ml ) at 37°C , 5% CO2 , in the presence or absence of 10 µM CA074Me or 10 µM ZRLR , and brefeldin A ( 3 µg/ml , eBioscience ) . The cells were then incubated for 20 min in 4% PFA fixation buffer , permeabilized for 20 min at 4°C using 0 . 1% saponin , 1% FCS permeabilization buffer , and incubated for 1 hour with ( PECy7 ) -conjugated anti-CD11c and PE-conjugated anti-IL-12 ( p40/p70 , BD Biosciences . Data was obtained using the MACSQuant flow cytometer . Furthermore , cells from polarization assays described below were fixed with 2% formaldehyde for 20 min at 4°C , permeabilized for 20 min at 4°C , and stained with the following Ab diluted in permeabilization buffer: Pacific Blue-conjugated anti-CD4 ( Biolegend ) , FITC-conjugated anti-IFN-γ ( BD Biosciences ) , PE-conjugated anti-IL-4 ( BD Biosciences ) , and allophycocyani-conjugated anti-IL10 ( Biolegend ) . Data was obtained using a LSR-II flow cytometer ( BD Biosciences , San Jose , USA ) . All results were analyzed using the software FlowJo . BMM from WT and cathepsin-deficient mice were seeded and infected as described for the proliferation assay . After 24 hours of incubation , the cells were washed with phenol-free complete RPMI medium to eliminate any extracellular parasites and incubated for further 48 hours in the absence or presence of 1 µg/ml LPS . The supernatants were collected , and the concentration of nitrite ( ) was determined by addition of 100 µl of culture supernatant to 100 µl of Griess reagent ( Sigma-Aldrich ) and incubation for 15 min at RT . The resulting absorbance at 540 nm was measured with the Multiskan Ascent ELISA reader ( Thermo Electronic Corporation ) . The nitrite concentrations were determined using sodium nitrite ( NaNO3 ) as a standard , and reflect the NO levels released by macrophages . 1×106 BMDC were seeded in a final volume of 1 ml in 24-well plates , and were stimulated with 5×106 L . major WT promastigotes ( infection ratio 1∶5 ) , LmAg ( 30 µl/ml ) , LPS ( 1 µg/ml; Sigma-Aldrich ) , or CpG ODN 1668 ( 5′-TCCATGACGTTCCTGATGCT-3′ , Qiagen Operon , Cologne , Germany ) . The cells were further incubated for 24 or 48 hours , and the supernatants were collected . The concentration of the cytokines in the supernatants was determined by sandwich ELISA , using capture-detection Ab pairs purchased from BD Biosciences for IL-12p40 , IL-6 and tumor necrosis factor alpha ( TNF-α ) , and R&D Systems for IL-10 ( Wiesbaden , Germany ) following the suppliers' instructions . In addition , IL-12p70 was measured by using the IL-12p70 ELISA Ready-SET-Go kit from eBioscience according to the manufacturer's instructions . To analyze the cytokine production in BMM , 1×106 cells were seeded in 500 µl into 24-well plates , together with 15×106 L . major WT promastigotes ( infection ratio 1∶15 ) , in the presence or absence of LPS ( 1 µg/ml ) . The cells were incubated for 24 and 48 hours , and the supernatants were collected . Cytokine measurements by ELISA were performed as described above . Total RNA from 2×106 BMDC or BMM , stimulated as described above , was isolated using the RNeasy kit ( Qiagen , Hilden , Germany ) according to the manufacturer's instructions . cDNA synthesis was performed using the iScript cDNA synthesis kit ( BioRad , Munich , Germany ) and the resulting cDNA was used at a 1∶8 dilution to assess the expression of IL-12a ( p35 ) by real-time PCR . The real-time PCR was performed in a final volume of 25 µl per well using Maxima SYBR Green/Fluorescein qPCR Master Mix ( Thermo Scientific , Schwerte , Germany ) and run with a CFX96 Touch real-time PCR detection system ( BioRad ) for 40 cycles . The primer pairs used were: Il12p35 forward: TGGCTACTAGAGAGACTTCTTCCACAA , Il12p35 reverse: GCACAGGGTCATCATCAAAGAC; Il12p40 forward: CGTGCTCATGGCTGGTGCAAA , Il12p40 reverse: ACGCCATTCCACATGTCACTGCC . The housekeeping gene β-actin was used for normalization of the samples: β-actin forward: CATTGCTGACAGGATGCAGA , β-actin reverse: TTGCTGATCCACATCTGCTG . Non-treated ( NT ) BMM were used as negative control , and LPS-stimulation ( 1 µg/ml ) was used as positive control . Relative gene expression values were calculated with the 2− ΔΔCT method [36] using WT NT BMM at t = 6 hours as a reference . Lymph nodes and spleens were removed from OVA-specific T-cell receptor ( TCR ) -transgenic OT-II mice , and kept in ice-cold complete RPMI medium in 60×15 mm petri dishes . Lymphocytes and splenocytes were isolated by mechanical dissociation using the sterile plunger of a 5-ml syringe and a cell strainer ( 70 µm , BD Falcon , Durham , USA ) . Red blood cells from spleen suspensions were eliminated with ammonium chloride lysis buffer for 5 min at 37°C . Naïve CD4+ T cells were isolated by negative selection using the CD4+ T-cell enrichment kit ( StemCell Technologies , Grenoble , France ) following the manufacturer's instructions . The enriched cells ( 1×104 ) were co-cultured with day 8 BMDC ( 5×104 ) from WT C57BL/6 , Ctsb−/− or BALB/c mice , together with 1 mg/ml ovalbumine ( OVA , Hyglos , Bernried , Germany ) or 100 ng/ml OVA-peptide 327–339 ( Activotec , Cambridge , UK . ) , and LPS ( 0 . 1 µg/ml ) , in U-bottom 96-well plates , with a final volume of 200 µl/well at 37°C , 5% CO2 . After 5 days of culture , the cells were harvested , counted , and adjusted to a concentration of 1×106 cells/ml for re-stimulation with phorbol 12-myristate 13-acetate ( PMA , 10 ng/ml , Sigma-Aldrich ) , ionomycin ( 1 µg/ml , Sigma-Aldrich ) , and brefeldin A ( 3 µg/ml ) , for 5 hours at 37°C , 5% CO2 . The cells were then washed , fixed in 2% formaldehyde , incubated for 20 min in saponin buffer , and the expression of Th1 cytokines was assessed by staining of intracellular cytokines , and flow cytometry as described above . 5×106 BMM were seeded in 6-well cell culture plates , and incubated for 4 hours at 37°C , 5% CO2 to promote adherence . The cells were thereafter infected with L . major promastigotes using a 1∶15 ratio , and further incubated at 37°C , 5% CO2 . At different time points ( t = 0 , t = 15 min , t = 30 min , and t = 1 h ) , lysates were prepared as follows: two different buffers were prepared , cytoplasmic cell fractionation buffer ( 10 mM HEPES , 10 mM KCl , 1 . 5 mM MgCl2 , 0 . 34 M D-sucrose , 10% glycerin and 1 mM dithiothreitol ( DTT ) , and nuclear cell fractionation buffer ( 3 mM ethylenediaminetetraacetic acid ( EDTA ) , 0 . 2 mM ethylene glycol tetraacetic acid ( EGTA ) , and 1 mM DTT ) . Directly prior to use , both buffers were supplemented with DTT ( final concentration 0 . 5 mM ) , protease inhibitor cocktail ( 1∶100 dilution , Sigma-Aldrich ) , and Na3VO4 ( final concentration 1 mM ) . At each time point , the stimulated cells were washed twice with cold PBS , and resuspended in 90 µl of ice-cold cytoplasmic cell fractionation buffer . 10 µl of 1% Triton X-100 in cytoplasmic cell fractionation buffer were added to the samples , and they were further incubated for 5 min on ice with gentle agitation . The samples were then centrifuged at 2000× g for 5 min at 4°C , and the supernatants were collected as cytoplasmic fraction , and stored at −20°C . The pellets were then washed with 100 µl of cytoplasmic cell fractionation buffer , and the samples were centrifuged again as described above . The supernatants were discarded , and the pellets were resuspended in 60 µl of nuclear cell fractionation buffer . The samples were further incubated for 30 min on ice . Then , they were sonicated on ice ( Sonoplus , Bandelin , Berlin , Germany ) using two cycles of 20 s each , with 40% of amplitude . The resulting suspensions were collected as nuclear fraction , and were stored at −20°C . The protein concentration of each sample was determined using a microplate bicinchoninic acid ( BCA ) protein assay kit ( Thermo Scientific ) , following the manufacturer's instructions . 40 µg of protein from each sample were separated by SDS-PAGE ( 10% acrylamide gels ) , and transferred to poly ( vinylidene difluoride; PVDF ) membranes . The membranes were incubated overnight with primary Ab against the p65 subunit of NFκB ( 1: Santa Cruz , Dallas , USA , 2: Cell Signaling , Danvers , USA ) , MEK1/2 , Lamin A/C ( Cell Signaling ) , and Ctsb ( R&D Systems , Minneapolis , USA ) . For detection , the membranes were incubated for 1 hour at RT with their corresponding horseradish peroxidase ( HRP ) -conjugated secondary Ab ( Cell Signaling ) , and developed using a chemiluminescence kit ( GE Healthcare , Munich , Germany ) . The membranes were then visualized using a FluorChem Q imager ( Biozym Scientific , Oldendorf , Germany ) . Values are provided as mean ± standard deviations from at least 3 independent experiments . Statistical significance was determined by the unpaired 2-tail Student's t test ( Microsoft Excel Software ) comparing , for each treatment , the results from Ctsb−/− or Ctsl−/− cells with their WT counterparts .
We used bone marrow stem cell progenitors from Ctsb−/− , Ctsl−/− and WT ( C57BL/6 ) mice to generate BMDC . At day 8 of culture with GM-CSF , the cells from all these mice displayed a typical myeloid immature DC morphology ( Figure 1A , 1–6 ) . These cells presented similar levels of CD11c expression , and comparable yields of CD11c+ cells were obtained ( Figure 1B ) . Similarly , BMM generated from these cathepsin-deficient mice did not show any significant differences in morphology ( Figure 1C , 1–6 ) and levels of F4/80 expression ( Figure 1D ) in comparison with BMM from WT mice . Next , eGFP-tg L . major promastigotes were used to analyze the kinetics of parasite uptake and processing by BMDC . After 2 hours of culture with L . major promastigotes at a parasite-to-BMDC ratio of 5 to 1 , most of the WT cells ( 70%±9 . 8% ) were infected with L . major , and rapidly processed the phagocytosed promastigotes ( Figure 2A ) . At 24 hours after infection , only around 13% of the cells remained infected . BMDC from Ctsb−/− and Ctsl−/− mice showed no significant differences neither in the uptake of eGFP-tg L . major promastigotes nor in their kinetics for processing the parasites in comparison with WT BMDC ( Figure 2B ) . These results indicate that cathepsins B and L are not relevant for the generation of BMM and BMDC , and that the capacity of the latter to phagocytize and process L . major promastigotes is not altered by the lack of cysteine cathepsins . We examined the survival of eGFP-tg L . major in macrophages . At 24 and 48 hours after infection , BMM from WT , Ctsb−/− and Ctsl−/− mice showed no significant differences in terms of percentage of infected cells , and average number of parasites per infected cell ( Figures 3A and 3B , and Figure S1 ) . To confirm these results , we also infected BMM from these mice with Luc-tg . L . major promastigotes , and measured the luminescence produced after addition of a luciferin substrate as a read-out for intracellular parasites . Again , no significant differences were found between WT BMM and cathepsin-deficient-BMM ( Figure 3C ) . In addition , we measured the nitrite concentrations in supernatants of BMM infected with L . major in the presence or absence of LPS , and in response to LPS alone ( Figure 3D ) . We found that 48 hours of infection with L . major promastigotes alone did not result in higher NO production compared to non-infected cells from WT , Ctsb−/− and Ctsl−/− BMM . Furthermore , we found no significant differences in nitrite levels in the supernatants of these cells neither after stimulation with LPS alone nor after infection with L . major promastigotes and further stimulation with LPS . These results show that cathepsin B and cathepsin L are dispensable for the control in vitro of L . major in BMM , and that absence of either of them does not affect the capacity of BMM to produce NO in response to neither L . major nor LPS . Upon encounter with pathogens , immature DC become activated and mature , up-regulating the expression of the antigen-presenting molecules MHC class II as well as co-stimulatory molecules such as CD86 , CD80 and CD40 [37] . We analyzed the maturation profile of BMDC from WT , Ctsb−/− and Ctsl−/− mice 24 hours after uptake of L . major promastigotes . We found that the expression of MHC class II molecules was greatly enhanced in BMDC from Ctsb−/− mice and , to a lesser extent , in BMDC from Ctsl−/− mice in response to L . major compared to WT BMDC ( Figure 4A ) . The expression of the co-stimulatory molecules CD40 , CD86 , and CD80 , on the other hand , was comparable among WT , Ctsb−/− and Ctsl−/− BMDC ( Figures 4B and 4C ) . In addition , we tested the expression of MHC class II and co-stimulatory molecules of BMDC pre-incubated with CLIK148 and CA074Me , a modified form of CA074 with increased cell permeability . BMDC treated with CA074Me showed an up-regulation of MHC class II molecules in response to L . major promastigotes higher than that observed for BMDC treated with CLIK148 or DMSO . On the other hand , we found no effect on CD86 expression , similar to the results obtained with the use of cathepsin-deficient cells ( Figure S2 A and B ) . These higher expression levels of MHC class II molecules in Ctsb−/− BMDC appeared to be a specific response to living parasites , since stimulation of Ctsb−/− BMDC with LmAg or HK parasites did not result in enhanced expression levels of MHC class II molecules and co-stimulatory molecules in comparison with WT and Ctsl−/− mice ( Figure 4 ) . In addition , no significant differences in the expression of co-stimulatory and MHC class II molecules were found in WT and cathepsin-deficient BMDC upon stimulation with LPS . The activation by inflammatory stimuli is known to recruit cathepsins B , L and S to late endosomes [38] , [39] , and , therefore , it is possible that different stimuli would lead to different profiles of active cathepsins for antigen processing . Our results demonstrate that L . major-stimulated Ctsb−/− BMDC , on the basis of MHC class II expression , display an enhanced maturation compared to WT and Ctsl−/− BMDC in response to L . major promastigotes . On the other hand , no significant differences in the expression of the co-stimulatory molecules CD86 , CD80 , and CD40 were observed . This effect was not elicited by stimulation with LmAg , HK parasites , or with LPS . Another important signal that BMDC use to instruct Th cell polarization is cytokine production . We analyzed the concentrations of the cytokines IL-12p70 , IL-12p40 , IL-10 , IL-6 and TNF-α in supernatants of BMDC 48 hours after infection with L . major promastigotes , or stimulation with LmAg . We found a significant increase in both IL-12p70 and IL-12p40 levels in Ctsb−/− BMDC in response to L . major in comparison with WT and Ctsl−/− BMDC ( Figures 5A and 5B ) . We also found that IL-10 expression was enhanced in Ctsb−/− BMDC ( Figure 5C ) , resembling the production of IL-10 in response to an up-regulation of IL-12 observed in DC after stimulation with LPS , a Th1 inducer . Moreover , the IL-12 up-regulation required living parasites , since stimulation of BMDC with LmAg or HK parasites did not induce higher levels of IL-12 production . In addition , we found no differences in IL-6 and TNF-α production between WT and cathepsin-deficient BMDC in response to L . major parasites ( Figure S3 ) . In contrast , Ctsb−/− BMDC presented an impaired IL-12p70 expression in response to CpG in comparison with WT BMDC , which reflects the importance of Ctsb in Toll-like receptor 9 ( TLR9 ) signaling ( Figure S4 ) . The IL-12 up-regulation in response to L . major observed with Ctsb−/− BMDC could not be replicated using the inhibitor CA074Me ( Figure S2 , C ) , and this inhibitor caused a dose-dependent decrease in IL-12p70 in LPS-stimulated cells ( Figure S2 , D ) . However , when BMDC from BALB/c and C57BL/6 mice were pre-treated with the peptide-based cathepsin B inhibitor ZRLR , they up-regulated their expression of IL-12p70 in response to L . major promastigotes resulting in levels comparable to those observed in Ctsb−/− BMDC ( Figure S5 , A ) . Pre-treatment of Ctsb−/− BMDC with ZRLR did not cause significant differences in the expression of IL-12 in response to L . major in comparison with DMSO pre-treated Ctsb−/− BMDC . In addition , we found a very similar pattern of increased IL-12p70 , IL-12p40 and IL-10 expression in Ctsb−/− BMM in comparison with WT and Ctsl−/− BMM ( Figures 5D to F ) . It should be noticed that the levels of IL-12p40 were considerably lower in BMM than in BMDC . Moreover , Ctsb−/− BMDC stimulated with LPS also produced slightly more IL-12p70 and IL-12p40 , but not IL-10 , compared to WT and Ctsl−/− BMDC . However , TNF-α expression was greatly impaired in Ctsb−/− BMDC , and no significant differences in IL-6 production among WT , Ctsb−/− and Ctsl−/− were found ( Figure 6 ) . This effect in IL-12 expression was also found in BMDC from BALB/c and C57BL/6 mice pre-treated with ZRLR ( Figure S5 , B to D ) . When co-cultured with naïve T cells from OT-II mice , we found that Ctsb−/− BMDC having LPS as a maturation stimulus resulted in higher frequencies of IFN-γ+ T cells , but not of IL-4+ T cells , indicating a Th1 polarization ( Figure 7 ) . Furthermore , the enhanced IL-12 production in response to L . major and LPS that we observed was found also at the transcriptional level , since Ctsb−/− BMM presented an up-regulation in the expression of Il12p35 and Il12p40 in response to both stimuli ( Figure 8 ) . Next , we investigated if this up-regulation of IL-12 was dependent on the NF-κB signaling pathway by assessing the translocation of the p65 subunit from the cytoplasm to the nucleus ( Figures S6 and S7 ) . We tested separately two different monoclonal antibodies to detect the NF-κB p65 subunit by Western Blot , and used for analysis the protein bands with the expected molecular weight ( 65 kDa ) . We found that the different levels of p65 for all the treatments in nuclear fractions from WT and Ctsb−/− had no statistical significance . We should point out that we found with both antibodies multiple protein bands with molecular weights other than 65 kDa . In particular , we observed a non-identified protein with a molecular size around 30 kDa in WT LPS-stimulated BMM but not in LPS-stimulated Ctsb−/− BMM . While these results suggest that NF- κB is not responsible of the Ctsb−/− -mediated regulation of IL-12 , more experiments with different approaches would be needed to confirm this observation . Altogether , our results show that Ctsb−/− BMDC and BMM presented a significant up-regulation of the Th1 promoter cytokine IL-12 in response to L . major and LPS , in comparison with WT and Ctsl−/− cells . This effect could be replicated using the peptide-based cathepsin B inhibitor ZRLR in BMDC from mouse strains susceptible and resistant to L . major . Furthermore , the observed up-regulation of IL-12 was present already at the transcriptional levels , and Ctsb−/− BMDC induced higher frequencies in vitro of Th1-polarized T cells .
In the present study , we investigated the differences in the signals that BMDC use to instruct Th cell polarization , i . e . , antigen presentation , expression of co-stimulatory molecules and cytokines , as a response to L . major promastigotes in the absence of Ctsb and Ctsl . In addition , we analyzed the impact of the lack of these proteases on the proliferation of L . major in infected BMM . We found that Ctsb−/− BMDC express higher levels of MHC class II molecules and of IL-12 in response to L . major promastigotes , and that this up-regulation of IL-12 expression was also present in BMM . These results indicate a novel role for Ctsb in the regulation of cytokine expression in response to L . major . Stem cell progenitors from Ctsb−/− and Ctsl−/− mice are able to generate BMDC and BMM with comparable yields and phenotypes as WT mice . These cells presented similar rates of parasite uptake , and BMDC showed similar kinetics of parasite processing . Moreover , parasite survival was similar in cathepsin-deficient and WT BMM . These results indicate that the up-regulation in MHC class II molecules and IL-12 expression that we observed was not due to differences in the parasite load between WT , Ctsb−/− and Ctsl−/− BMDC . DC present an incomplete maturation after uptake of L . major promastigotes [40] and L . amazonensis [41] . Previous studies using the Ctsb-selective inhibitor CA074 and the Ctsl-specific inhibitor CLIK148 showed drastic changes in the Th cell response of mice infected with L . major [25] , [28] , [42] , and it was hypothesized that these effects resulted from differences in antigen processing . Later reports showed that cathepsin S is indispensable for the degradation of the invariant chain in antigen-presenting cells [43]–[45] , while Ctsl was relevant for the processing of antigens only in cortical thymic epithelial cells [46] . Ctsb and cathepsin D ( Ctsd ) were shown to be dispensable for the maturation of MHC class II molecules and the presentation of several antigens [31] . However , this study also reported that splenocytic antigen-presenting cells from Ctsb- or Ctsd-deficient mice were actually more efficient to present certain antigens to T-cell hybridomas , in agreement with reports using different inhibitors with murine splenocytes [47] and primary human DC [48] . The latter study described the use of the peptide-based cathepsin inhibitor ZRLR which was shown to have superior specificity towards Ctsb compared to CA074Me . Deussing et al . suggested that some antigenic determinants may present different degrees of susceptibility to degradation by cathepsins before being able to bind to MHC class II molecules , and , therefore , would benefit from the absence of Ctsb or Ctsd [31] . We found higher levels of MHC class II molecules in Ctsb−/− BMDC than in WT and Ctsl−/− BMDC in response to L . major . Similar results were found with the use of the inhibitor CA074Me , while no significant differences were found when we used LmAg or heat-killed parasites as stimulus . The different results observed with promastigote- and LmAg-mediated stimulation could reflect the interaction of the living parasite with the host cell , and the differences in uptake mechanism and subsequent processing [47] , [48] . Stimulation with heat-killed parasites led to comparable levels of MHC class II molecules and co-stimulatory molecules as observed with infected BMDC . However , we found no significant differences between WT and cathepsin-deficient BMDC . This could indicate that the higher levels of MHC class II molecules in infected Ctsb−/− BMDC in comparison with WT BMDC are related to the active manipulation that the living parasite exerts in its host cell . Incomplete BMDC maturation , such as after stimulation with Trypanosoma brucei antigens , has been shown to induce activation of genes correlating with the induction of Th2 polarization [49] . In contrast , higher levels of antigen presented [50] are associated with induction of Th1 responses . Upon in vitro infection with Leishmania promastigotes , BMDC present poor cytokine expression [51] . We found that Ctsb−/− BMDC were able to express significantly higher levels of IL-12 ( both p70 and p40 forms ) than WT and Ctsl−/− BMDC in response to L . major promastigotes . We did not detect significant differences in IL-6 and TNF-α , which would indicate that the observed effect was not a generalized up-regulation of cytokine expression , but a rather specific mechanism . Likewise , infection of BMM with L . major promastigotes induces poor cytokine expression [17] , [19] . Ctsb−/− BMM also presented a significant increase in IL-12 expression , with similar levels of IL-12p70 as Ctsb−/− BMDC , although the up-regulation of IL-12p40 was not as high . Pompei et al . reported a differential release of IL-12 in BMDC and BMM in response to Mycobacterium tuberculosis , and suggested that this was dependent upon the level of engagement of different TLR , particularly TLR9 in BMDC [52] . TLR9 requires processing by endosomal cathepsins to initiate signaling [53] , [54] . In agreement with Matsumoto et al . [53] , we observed a great impairment in IL-12 expression in Ctsb−/− BMDC upon CpG stimulation . Therefore , the up-regulation in IL-12 expression by Ctsb−/− BMDC and BMM that we observed here is independent from TLR9 signaling . In addition , we tested the response of Ctsb−/− and Ctsl−/− BMDC to LPS , which is recognized by TLR4 . Ctsb−/− BMDC stimulated with LPS did not show a significantly higher expression of MHC class II molecules in comparison with WT BMDC or Ctsl−/− BMDC but they did display higher levels of IL-12 . Moreover , the expression of TNF-α was greatly impaired in Ctsb−/− BMDC , in agreement with Ha et al . [55] who reported that LPS-treated BMM secrete significantly less TNF-α in response to LPS upon lack of Ctsb , due to an accumulation of TNF-α-containing vesicles that could not reach the plasma membrane . Schotte et al . reported an impairment of cytokine production in macrophages stimulated with LPS upon treatment with the cathepsin B inhibitor z-FA . fmk [56] . Our results with Ctsb−/− BMDC and BMM do not show an inhibition of IL-12 expression but rather an enhancement . We obtained similar results using CA074Me , the methyl ester form of CA074 . Upon uptake by the cell , CA074Me is hydrolyzed to CA074 , but if this hydrolysis is incomplete , inhibition of other cysteine proteases besides Ctsb takes place [57] . In contrast , pre-treatment of BMDC from susceptible BALB/c or resistant C57BL/6 mice with ZRLR induced IL-12 expression levels comparable to those observed with Ctsb−/− BMDC . Although we did not have Ctsb−/− mice on a BALB/c background available , these results suggest that the observed role of Ctsb in L . major infection would be independent of the mouse strain . Upon infection with L . major promastigotes , Ctsl−/− BMDC and BMM did not present significant differences in the production of cytokines in comparison with WT cells . However , they produced higher levels of IL-10 and TNF-α , but not IL-12 , in response to LPS . These results alone would not explain the observations made by Onishi et al . [58] , where use of the cathepsin L inhibitor CLIK148 caused a Th2-like immune response to L . major in otherwise resistant mice . Again , it should be kept in mind that CLIK148 can also inhibit other cathepsins , including C , K , and S [59] , which could have contributed to this response . Altogether , while previous studies hypothesized that lack of Ctsb or Ctsl activities during L . major infection would lead to changes in Th cell polarization due to differences in antigen presentation [25] , [28] , [42] , our results indicate that Ctsb−/− BMDC up-regulate two of the three types of signals used for instructing Th cell polarization: expression of MHC class II molecules and cytokine expression . Thus , these cells exhibit a “pro-Th1”-like profile . Moreover , co-culture of purified naïve CD4+ T cells with Ctsb−/− BMDC resulted in a higher frequency of Th1-polarized T cells compared to WT BMDC . To the best of our knowledge , the present study is the first indicating a new role of Ctsb as a regulator of cytokine expression in response to L . major . Future work will focus on the implications of these effects in vivo , considering the infection of Ctsb−/− animals , as well as in transfer experiments of Ctsb−/− BMDC into WT animals . In which way could Ctsb influence cytokine production in BMDC and BMM ? Ben-Othman et al . reported that L . major parasites induce macrophage tolerance by a process involving MAPK and NF-κB pathways of the host [60] . These pathways , although initially activated by exposure to the parasite , become silenced when the infection is firmly established , rendering the cell unresponsive to further stimulation with LPS [61] . This silencing has been attributed to different virulence factors , including surface phosphoglycans [16] , [62] , the metalloprotease GP63 [63] , and cysteine proteases from L . mexicana [21] . It is possible that in the absence of Ctsb , one or more of these key signaling pathways are no longer silenced by L . major promastigotes . This would open a range of new questions regarding the involvement of Ctsb in L . major infection , e . g . , whether Ctsb interacts directly with the parasites , contributing to processing or activation of one or more virulence factors , or whether Ctsb directly interferes with any intermediate of key signaling pathways , such as NF-kB and MAPK . While Ctsb−/− BMDC and BMM were able to up-regulate IL-12 in response to L . major and LPS , IL-6 was not regulated . IL-6 transcription has been shown to depend on NF-κB [64] . Therefore , we hypothesize that the molecular mechanism behind the involvement of Ctsb−/− in the expression of IL-12 would not be shared by IL-6 . Our results suggest that the regulation of IL-12 expression by Ctsb is not NF-κB-dependent , although further work is necessary to confirm this observation and to explore the interaction of Ctsb with other candidate signaling pathways . In addition , finding the location of these interactions would be a key to further understand the mechanisms underlying these processes , e . g . , whether proteolytic processing of signaling intermediates takes place in the cytoplasm , or whether cleavage of transcription factors by Ctsb occurs in the nuclear space , as described in thyroid carcinoma cells [65] . The concept of “protease signaling” has gained increasing attention in different research fields [66] , especially in the context of therapeutic applications . Yet , more research is needed in order to understand the interplay between proteolytic networks and other signaling pathways . On the basis of the present study , we propose a novel role for cathepsin B during L . major infection: in addition to its involvement in antigen presentation , it is also a regulator of cytokine expression . It is tempting to speculate that pharmacological inhibition of cathepsin B may improve the Th1-mediated clearance of L . major .
|
The emergence of resistance to the available drugs against cutaneous leishmaniasis emphasizes the need of new chemotherapeutic approaches . Cysteine proteases from Leishmania are important virulence factors and , therefore , interesting drug targets . Studies on inhibitors against these enzymes during Leishmania major infection in mice had shown that host equivalents of these proteases are also affected , namely cathepsin B and cathepsin L . The inhibition of cathepsin B resulted in immune-mediated protection , while inhibition of cathepsin L caused susceptibility to the parasite . In the present study , we investigated the effect of cathepsin deficiency on the signals used by dendritic cells to orchestrate the T helper ( Th ) -mediated immune response against L . major and the control of parasite proliferation within infected macrophages . The results demonstrate that cathepsin B-deficient dendritic cells express higher levels of the antigen-presenting MHC class II molecules than WT and cathepsin L-deficient cells . Surprisingly , dendritic cells and macrophages deficient for cathepsin B showed higher expression of the protective Th1-inducing cytokine IL-12 . Therefore , we propose a novel role of this protease as a regulator of cytokine expression . Altogether , these findings suggest that cathepsin B down-regulates the Th1 response to L . major , and , in its absence , antigen-presenting cells express signals protecting against the parasite .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"blood",
"cells",
"dendritic",
"cells",
"white",
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"protozoans",
"cell",
"biology",
"animal",
"cells",
"t",
"cells",
"antigen-presenting",
"cells",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"immunology",
"tropical",
"diseases",
"neglected",
"tropical",
"diseases",
"parasitic",
"protozoans",
"organisms"
] |
2014
|
Cathepsin B in Antigen-Presenting Cells Controls Mediators of the Th1 Immune Response during Leishmania major Infection
|
In tropical Africa , where the spectrum of the bacterial pathogens that cause fevers is poorly understood and molecular-based diagnostic laboratories are rare , the time lag between test results and patient care is a critical point for treatment of disease . We implemented POC laboratory in rural Senegal to resolve the time lag between test results and patient care . During the first year of the study ( February 2011 to January 2012 ) , 440 blood specimens from febrile patients were collected in Dielmo and Ndiop villages . All samples were screened for malaria , dengue fever , Borrelia spp . , Coxiella burnetii , Tropheryma whipplei , Rickettsia conorii , R . africae , R . felis , and Bartonella spp . We identified DNA from at least one pathogenic bacterium in 80/440 ( 18 . 2% ) of the samples from febrile patients . B . crocidurae was identified in 35 cases ( 9 . 5% ) , and R . felis DNA was found in 30 cases ( 6 . 8% ) . The DNA of Bartonella spp . was identified in 23/440 cases ( 4 . 3% ) , and DNA of C . burnetii was identified in 2 cases ( 0 . 5% ) . T . whipplei ( 0 . 2% ) was diagnosed in one patient . No DNA of R . africae or R . conorii was identified . Among the 7 patients co-infected by two different bacteria , we found R . felis and B . crocidurae in 4 cases , B . crocidurae and Bartonella spp . in 2 cases , and B . crocidurae and C . burnetii in 1 case . Malaria was diagnosed in 54 cases . In total , at least one pathogen ( bacterium or protozoa ) was identified in 127/440 ( 28 . 9% ) of studied samples . Here , the authors report the proof of concept of POC in rural tropical Africa . Discovering that 18 . 2% of acute infections can be successfully treated with doxycycline should change the treatment strategy for acute fevers in West Africa .
The POC immuno-chromatographic tests ( ICT ) tests for malaria and HIV diagnostics are extensively used in Africa [21] . The malaria RDT played a significant role in reducing of malaria morbidity [5] . Other diagnostic tests are either not available , or , as in case of trachoma , not suitable for field diagnosis as the specificity decreases in hot and dry conditions [22] . Although POC tests rely mostly on ICT or agglutination assays , miniaturization and full automation of molecular methods allow for quicker real-time PCR-based detection of pathogens using simplified procedures . It was reported to be successfully applied in the POC diagnostics [23] . Although very promising experiences using molecular express diagnostic tests based on GenXpert technology were reported in South Africa for the diagnosis of tuberculosis [24] , the availability of these tests in sub-Saharan Africa is still extremely insufficient . For this reason , all of the samples were initially collected in the rural health dispensaries and then sent to Marseille ( France ) for molecular studies . Sending the samples to France , however , limited or perhaps negated the direct diagnostic benefit of our studies . The time-consuming logistics and laboratory procedures hampered the use of available data for clinical needs . The present paper provides the report of the installation of the POC laboratory based on ICT and real-time PCR-based detection of pathogens in rural Africa , the activity report and describes the methodology used to organize the running of this new biomedical technology in response to the needs of rural African healthcare .
Located in the Saloum region of Senegal 280 km southeast of Dakar and approximately 15 km north of the Gambian border , Dielmo is an approximately 350-inhabitant village where malaria has been holoendemic . Ndiop with its ∼400 inhabitants is situated 5 km away ( Figure 1; Table 1 ) . The study site has been described in detail elsewhere [3] . Malaria transmission was intense and perennial , with a mean 258 infected bites per person per year during 1990–2006 [5] . The research station is situated within 10 m of the Nema River and 30 m of the first household in the village and includes a dispensary with a laboratory and also has ten huts to accommodate the project staff and visitors ( Figure 2 ) . One nurse , two technicians and three fieldworkers are present every day in the village . The dispensary is open 24 hours per day , 7 days per week . The same type dispensary exists also in Ndiop . The dispensaries and the laboratory are equipped with standard materials for the screening and diagnosis of different conditions , including malaria and other febrile diseases ( light microscopy , thick-smear , rapid malaria diagnostic test , hemoglobin diagnostics , stethoscope , blood pressure measurement , pregnancy testing ) and a pharmacy that has essential medicines . The logistic equipment includes a generator , solar energy , one freezer at −20°C and one refrigerator at +4°C and one vehicle that is used for research operations and patient referrals . The station has a mobile telephone but does not have a reliable Internet connection . Point-of-Care ( POC ) is defined as medical diagnosis at or near the site of the patient [23] . We synthesized several principal issues concerning the installation of molecular-based POC in Dielmo: In September 2010 , we started to build the POC laboratory suite at the Dielmo field research station based on the proposed model [23] . In keeping with the design of PCR suites , the dimensions of the room of 9 m2 with a double brick-wall were designed to better isolate the room from the outside temperature . Air conditioning was assured by a split tropicalized air conditioner . In November 2010 , we started the installation of the equipment . The POC was equipped with working tables and chairs and with a portable meteorological station . The laboratory equipment included one freezer at −20°C and one refrigerator at +4°C , a safety cabinet designed for PCR rooms , a heating block for Eppendorf tubes , a manual polyvinyl chloride pump , a portable centrifuge for 1 . 5 ml tubes , a vortex , a Qiagen BioRobot EZ1 Workstation for DNA extraction , and two Smart cycler II units ( Cepheid Europe; Maurens-Scopont , France ) supplied with computers , printers and mini-centrifuges for specially designed tubes [23] . The pre-existed solar power was not usually enough to supply the energy for the molecular-based laboratory; the capacities of batteries do not allow to maintain the work of the real-time PCR machine , computer and air conditioner at the same time for at least one and half hours . A diesel power generator was installed outside to provide the energy for the POC . The total cost of the POC was estimated at 158 , 670 euros ( all taxes and transportation included ) .
During the first year of the study ( February 2011 to January 2012 ) , 440 blood specimens from febrile patients were collected in Dielmo and Ndiop villages . All samples were screened for malaria ( rapid diagnostic test ) , dengue fever ( rapid diagnostic test ) , Borrelia spp . , C . burnetii , T . whipplei , R . conorii , R . africae , R . felis , and Bartonella spp . A second round of qPCR was performed to confirm positive results of the first reaction for C . burnetii , T . whipplei and Bartonella spp . We identified at least one pathogenic bacterium DNA in 80/440 ( 18 . 2% ) of the samples from febrile patients . No mortality was associated with this group of patients during our study . Infection with B . crocidurae was identified in 35 cases ( 9 . 5% ) . Different results were found at the POC in Marseille in four cases . In two of these , the borderline Ct number ( 34 . 38 and 34 . 29 ) was identified in Dielmo , and the samples were considered positive , but the samples were counted as negative in Marseille because the qPCR presented a Ct value higher than 35 . In the other two cases , the samples were negative in Dielmo but weakly ( 37 . 88 and 37 . 77 ) positive in Marseille ( Table 3 ) . However the sequences of amplicons corresponded to B . crocidurae , and the samples were considered positive . R . felis DNA was found in 30 cases ( 6 . 8% ) of acute febrile illness in Dielmo and Ndiop . All samples that generated a fluorescent signal for R . felis in qPCRs performed in Marseille were previously R . felis positive in Dielmo . However , in 11/30 cases the Ct numbers were higher than 36 . We found discordance in only two cases between a positive result in POC and a negative result in Marseille . Interestingly that in 3 cases observed by one of the authors ( O . M . ) , the infection presented with a rash in the contrary to our previous data [9] . No cases of R . typhi infection was diagnosed . The DNA of Bartonella spp . ( including B . quintana ) was identified in 23/440 cases ( 5 . 2% ) . Unfortunately , the B . quintana-specific tests were not available at the beginning of this study , so we were only able to identify the species after the specific qPCR was performed in Marseille . In four of 23 cases , the Bartonella species was identified as B . quintana . There were no discordance between Marseille and Senegal results The DNA of C . burnetii was identified in 2 cases ( 0 . 5% ) . Both were confirmed in Marseille . T . whipplei ( 0 . 2% ) was diagnosed in one patient . No DNA of R . africae or R . conorii was found in patients . We calculated sensitivity , specificity , positive and negative predictive values for each of used tests ( Table 3 ) . Due to the low number of positive samples , the specificity and sensitivity in most cases were 100% . Among the 7 patients co-infected with two different bacteria , R . felis and B . crocidurae were present in 4 cases , B . crocidurae and Bartonella spp . were found 2 cases , and B . crocidurae and C burnetii were found in 1 case . In 54 cases we identified Plasmodium falciparum in the samples . In 5 cases DNA of R . felis was also found in the sample , in one case B . crocidurae and in one case T . whipplei . No samples were found positive for dengue fever . In total , at least one pathogen ( bacterium or protozoa ) was identified in 127/440 ( 28 . 9% ) of studied samples . The appropriate treatment for the etiological agent of disease was applied when possible [5] . All of the identified bacteria are susceptible in vitro or in vivo to doxycycline [9] , [14] , [16] , [28] . The dose for the children was calculated based on the amount of 5 mg/kg per day ( Table 4 ) . No pregnant women were diagnosed with bacterial infections . As all included patient participate in the long-term malaria surveillance program [3] we were able to monthly survey almost all patients but without retesting them for bacterial pathogens if they had no fever . Fourteen persons left the protocol due to death or migration: 1 person died of hepatic cirrhosis and 13 have left the villages . In all cases the treatment ( Table 3 ) resulted in rapid resolution of all symptoms , including the fever . No relapses and mortality associated with identified infectious diseases was registered .
Molecular diagnosis of infectious diseases supplemented and even replaced many current culture and serologic tests in microbiology [29] . Many fastidious bacteria ( including rickettsiae , C . burnetii , bartonellae , T . whipplei ) are very difficult to isolate . For many infectious diseases ( including borrelioses and T . whipplei infection ) serology is not reliable method for diagnostic [30] . Moreover , for all these infection serological methods of acute infection are difficult and require highly experienced technician . Thus , molecular methods play the primary role in diagnostics of these infectious diseases in traditional and POC-based laboratories [23] . The widespread use of qPCR that is less expensive than conventional PCR and reduces delay in the diagnosis of rickettsial infections . The development of qPCR strategies in the diagnosis of rickettsioses has previously been proposed [31] . Thus , qPCR-based on-site molecular testing POC laboratory seems to be the easiest way to diagnose infectious diseases in rural conditions . Molecular diagnostics are a very sensitive method of diagnosing infectious diseases . Materials and reagents are fragile , and the contamination of the reagents by amplicons may be a significant problem [30] , [31] . Lyophilized , ready-to-use mixes for individual tests might be the solution for the POC in rural Africa . Each flask contains the freeze-dried mix for the testing of one sample ( with dilution; positive and negative controls ) , is thermo-stable , easy-to-transport and is protected against contamination . During the entire study , we never had a problem with contamination . Our negative controls were consistently negative . The impact of POC for the health of the population may be important . Up until now , the clinical diagnosis of non-malarial , acute febrile illnesses in rural Senegal was impossible . The example of tick-borne relapsing fever is interesting because before the implementation of the POC laboratory in Dielmo , Borreliae were detected in Giemsa-stained thick blood smears during the examination for malaria . However , only very experienced laboratory personnel could confirm the presence of Borreliae in the blood; thus , in most cases , the diagnosis was made only after repeated examination performed in the URMITE Dakar laboratory . This means that there was a delay of several weeks between the original consultation and the diagnosis . A feature of relapsing fever is that it may last for weeks without appropriate treatment . Therefore , the patient might still need medical assistance even 2–3 weeks after the initial consultation . As for the other pathologies , before the beginning of our project , they were either unknown in Senegal ( infections caused by R . felis , R . africae , T . whipplei , and B . quintana ) or were rarely studied and never diagnosed ( Q fever ) . The current practice in Senegal is to use amoxicillin and/or cotrimoxazole in case of non-malarial fevers [32] . Discovering that all identified infections can be successfully treated with doxycycline and many of them are not sensitive to amoxicillin and/or cotrimoxazole should change the treatment strategy for acute unexplained fevers in West Africa . POC laboratory based on ICT and real-time PCR is a viable in locations with very limited infrastructure . However , the installation of this type of laboratory by a public health system may be limited by the finance and expertise . The costs of installation and running may be quite high for the developing countries . The running demands the qualified technicians , regular supervision and a good QC system that is not always available . The POC laboratory , however , may be used as a very effective tool for studying the epidemiology of infectious diseases . The diagnosis of virtually any infectious disease may be performed in a very short period of time . This rapid diagnosis may be very important for the global healthcare systems . The solutions proposed by POC may be the only tool for discovering the etiologies in cases of emerging and novel infectious diseases and for studying the epidemiology of poorly known or poorly studied infections . The example of infection caused by R . felis in Africa is very interesting . The emergence of this bacterium as a frequent cause of febrile disease in Senegal and Kenya [9] , [11] , [12] and the simultaneous absence of the bacterium in fleas [33] , which are considered the normal vectors of the bacterium worldwide , suggest that in Africa , the epidemiology of the febrile illness caused by R . felis may be different from that in other countries . We already possess some evidence that the clinical picture of primary infection may have some unique traits , such as a vesicular rash ( data not published ) . The POC laboratory may help to discover the enigmas of this infection in Africa . For example , it can help in the discovery of the vector and reservoirs of the bacterium and in defining the clinical picture of the illness . Moreover , POC laboratories may be easily arranged by broad-range diagnostic systems to perform large spectrum studies for potentially unknown pathogens . These laboratories may be very important for global healthcare and travel medicine . Finally , we can conclude that the establishment of the POC laboratory in rural Senegal may indicate the beginning of a new approach in the studies of emerging tropical infectious diseases . The close contact between the patient and the researcher may provide diagnosis of disease in a very short time . Additionally , the plasticity of the assays makes it possible to quickly change the specialization of the POC and to broaden the specificity of the molecular studies performed at the POC .
|
In tropical Africa , clinical laboratories capable of performing complicated diagnostic studies like PCR are rare and are almost always found in large cities . Moreover , a number of infectious diseases , many of them are emerging and neglected , may be quickly and reliably diagnosed only by molecular biology . This is one of the reasons why the repertoire of bacterial infectious diseases in tropical Africa is poorly known . The laboratory based on the Point-of-Care ( POC ) principle has been designed in order to resolve the time lag between test results and patient care , which is the critical point for the treatment . We report here the first successful experience of the installation of POC laboratory in rural Senegal . During the first year of the study ( February 2011 to January 2012 ) we identified DNA from at least one pathogenic bacterium in 80/440 ( 18 . 2% ) of the samples from febrile patients . In most of the cases it was relapsing fever and rickettsiosis agents . Malaria was diagnosed in 54 cases . In total , at least one pathogen ( bacterium or protozoa ) was identified in 127/440 ( 28 . 9% ) of studied samples . Discovering that at least 18 . 2% of acute infections can be successfully treated with doxycycline should change the treatment strategy for acute fevers in West Africa .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"borrelia",
"infection",
"public",
"health",
"and",
"epidemiology",
"infectious",
"disease",
"epidemiology",
"microbiology",
"rickettsia",
"relapsing",
"fever",
"bartonellosis",
"bacterial",
"diseases",
"emerging",
"infectious",
"diseases",
"neglected",
"tropical",
"diseases",
"trench",
"fever",
"bacterial",
"pathogens",
"infectious",
"diseases",
"q-fever",
"medical",
"microbiology",
"epidemiology",
"dengue",
"fever",
"biology",
"gastrointestinal",
"infections",
"bartonella",
"bacteremia",
"gram",
"negative",
"tropical",
"diseases",
"(non-neglected)",
"malaria"
] |
2013
|
Point-of-Care Laboratory of Pathogen Diagnosis in Rural Senegal
|
Genomic tools have revealed genetically diverse pathogens within some hosts . Within-host pathogen diversity , which we refer to as “complex infection” , is increasingly recognized as a determinant of treatment outcome for infections like tuberculosis . Complex infection arises through two mechanisms: within-host mutation ( which results in clonal heterogeneity ) and reinfection ( which results in mixed infections ) . Estimates of the frequency of within-host mutation and reinfection in populations are critical for understanding the natural history of disease . These estimates influence projections of disease trends and effects of interventions . The genotyping technique MLVA ( multiple loci variable-number tandem repeats analysis ) can identify complex infections , but the current method to distinguish clonal heterogeneity from mixed infections is based on a rather simple rule . Here we describe ClassTR , a method which leverages MLVA information from isolates collected in a population to distinguish mixed infections from clonal heterogeneity . We formulate the resolution of complex infections into their constituent strains as an optimization problem , and show its NP-completeness . We solve it efficiently by using mixed integer linear programming and graph decomposition . Once the complex infections are resolved into their constituent strains , ClassTR probabilistically classifies isolates as clonally heterogeneous or mixed by using a model of tandem repeat evolution . We first compare ClassTR with the standard rule-based classification on 100 simulated datasets . ClassTR outperforms the standard method , improving classification accuracy from 48% to 80% . We then apply ClassTR to a sample of 436 strains collected from tuberculosis patients in a South African community , of which 92 had complex infections . We find that ClassTR assigns an alternate classification to 18 of the 92 complex infections , suggesting important differences in practice . By explicitly modeling tandem repeat evolution , ClassTR helps to improve our understanding of the mechanisms driving within-host diversity of pathogens like Mycobacterium tuberculosis .
The genotyping technique known as MLVA ( multiple loci variable-number tandem repeats analysis ) , which identifies the number of copies of tandem repeat regions at specific pre-selected loci , has benefited the study of many bacteria . Data produced by MLVA can be used to glean information about bacterial lineage , pathogenicity and relation to other bacteria of the same species [1] . Our study focuses on a specific bacterium , Mycobacterium tuberculosis , but our methods are generally applicable to a variety of bacteria . Genetic and genomic approaches for interrogating the composition of Mycobacterium tuberculosis infections occurring within individuals has in some settings revealed an impressive degree of complexity , reflecting both within-host mutation and reinfection as distinct routes to complexity [2] . These complex infections , especially those comprising both drug-susceptible and drug-resistant isolates ( i . e . heteroresistance ) , can undermine the effective treatment of individual patients [3–5] , complicate laboratory testing and evaluation of treatment programs [2] , and affect the transmission dynamics of disease in communities [6 , 7] . While an individual’s clinical response to treatment may not depend on whether heteroresistance has arisen through within-host mutation or by reinfection , our ability to distinguish these mechanisms has profound implications for our understanding of the natural history of disease and for projections of disease trajectories . For example , high contributions of reinfection indicate limited immune protection associated with previous infection , and have implications for the impact of new and existing vaccines [8] and for the effectiveness of preventive therapy [9] . High contributions of within-host mutation would affect expected rates of acquired resistance and would have implications for optimal antibiotic dosing strategies [10 , 11] . Accordingly , accurate estimates of the prevalence of complex infections among tuberculosis patients and new methods for distinguishing the relative contributions of within-host mutation and reinfection to within-host diversity would be valuable . Mycobacterial interspersed repetitive unit-variable number tandem repeat ( MIRU-VNTR ) , the specific name of the MLVA technique for Mycobacterium tuberculosis , is a currently favored approach for genotyping strains and offers advantages for detecting within-host heterogeneity over other methods such as spacer oligonucleotide sequencing ( spoligotyping ) and restriction fragment length polymorphism analysis ( RFLP ) . These molecular genetic approaches for TB genotyping are reviewed in Mathema et al . [12] and an evaluation of their utility for detecting complex infections is described by Cohen et al [2] . MIRU-VNTR is a microsatellite typing system which produces a readout containing the number of copies of a repeat region at several pre-selected loci [13 , 14] . These copy number variants ( CNVs ) can then be used to compare the Mycobacterium tuberculosis strain to other similarly typed strains . If a patient harbors a complex infection , there will frequently be 2 ( and sometimes 3 ) different CNVs at a single locus , and these can result from a clonally heterogeneous or a mixed infection . This situation is illustrated in Fig 1 . Classifying complex MIRU-VNTR patterns as being due to either within-host mutation or reinfection is challenging . The current accepted approach for distinguishing clonal heterogeneity from mixed infection is a simple rule-based method: if two or more loci have multiple CNVs the infection is classified as “mixed” , whereas if only one locus has multiple CNVs the infection is classified as “clonally heterogeneous” [13–15] . This approach is sensible given that the more complexity observed within a particular genotype , the more likely it is to be due to reinfection with a distinct second strain . In addition , there are several sources of evidence which suggest that clonal evolution occurring over a relatively short period is unlikely to result in multiple complex loci [16–18] . Nonetheless , the rule this approach is based on suffers from several limitations . First , it does not take the context of the infection into account ( namely , whether the constituent strains are present in other members of the population ) . Second , it does not distinguish between copy numbers that are a small genetic distance apart ( such as 3 and 4 ) from ones that are far apart ( such as 3 and 15 ) , even though clonal heterogeneity is less plausible in the latter case . Third , this approach does not facilitate the resolution of mixed infections into their constituent strains . We propose a new method , which we call ClassTR , to classify complex infections using MLVA data . Our method is based on an established model of tandem repeat evolution that accounts for the stepwise character of mutations , which we extend by using differential rates of evolution for different loci . ClassTR leverages the entire set of isolate genotypes collected in a population in order to resolve complex strains into simple strains ( i . e . strains with only one CNV at each locus ) . Then , using a model of tandem repeat evolution it identifies the most likely sources of each simple strain to establish the probability of each patient having a mixed infection . We show that ClassTR outperforms the standard rule-based method , reducing its error rate by 61% on simulated data , and produces significantly different classifications than the standard method on a dataset collected from a community in KwaZulu-Natal , South Africa . ClassTR is implemented in the R Statistical Computing Language [19] in Supplementary Materials ( S1 Code ) .
We say that a patient harbors a complex infection if at least one of the MIRU-VNTR loci contains 2 different CNVs . We assume that there are always 1 or 2 CNV per locus , and this is indeed what we usually observe in practice . The ClassTR algorithm classifies these complex infections as resulting from either clonal heterogeneity or mixed infection ( when both clonal heterogeneity and mixed infection are present , ClassTR classifies the infection as mixed ) . It includes three steps , each of which is briefly discussed below and described in more detail in the Methods section . Briefly , ClassTR starts by creating an optimization problem to resolve the complex strains into their constituent simple strains . After solving this optimization problem , ClassTR uses the resulting simple strain representation of complex strains to infer the possible provenance of each of these complex strains . Finally , it computes the probability of clonal heterogeneity for each patient with a complex infection . We illustrate this process on a small example with 3 simple and 3 complex strains in Figs 2 and 3 . Following Aandahl et al [20] we define distances between strains based on explicit models of tandem repeat evolution: a constant model and a linear model . Both models assume that copy numbers evolve in a stepwise fashion , consistent with the process of slipped-strand mispairing [21] . The constant model assumes a Poisson process at each locus by which the copy number increases or decreases by 1 at a constant rate , while the linear model assumes a Poisson process at each copy , so that the rate of mutation is proportional to the current copy number . In both cases the distance between two strains represents the total number of mutation events required to go from one to the other . In addition to these basic models where different loci undergo mutations at the same rate , we also consider weighted models in which different loci mutate at different rates . In order to estimate these locus-specific mutation rates we use measures of locus diversity . We say that a set of simple strains exactly covers a complex strain if the set of CNVs at each locus of the simple strains is precisely the CNVs at the corresponding locus of the complex strain . In general , a strain with one or more complex loci having 2 CNVs each can be exactly covered by 2 simple strains . However , in the absence of additional constraints there can be as many as 2q−1 possible such covers of a strain with q complex loci , which is 2048 possibilities for q = 12 , the largest we observe in our data . We make a parsimony assumption and search for the covers of the complex strains that introduce the smallest possible number of additional simple strains ( i . e . ones not observed in the original dataset ) . This defines an optimization problem which may have multiple solutions , especially when the cases are not densely sampled from the population . We narrow down alternative possibilities by a system of rewards for using a strain frequently observed in the dataset and penalties for strains that are far removed from any other simple strains in the dataset . Once the optimization problem is solved , every complex strain is represented as a superposition of simple strains . These simple strains can be present in the original dataset or newly added . For each of the newly added simple strains we compute one or more predecessors among the original simple strains , defined as the closest among these strains according to the distance we chose . The final probabilities are then obtained by comparing the sets of predecessors of the two strains constituting a given complex strain; the more similar they are , the more likely the strain is to be the result of clonal heterogeneity . The South African dataset we work with consists of data collected during a prospective study of within-host diversity of M . tuberculosis . Briefly , 500 adult , sputum smear-positive TB patients in a geographic cluster of participating clinics in KwaZulu Natal were sequentially recruited for participation at the time of diagnosis and before treatment was initiated . Additional pre-treatment sputum was collected from each participant and cultured in solid and liquid media . Bacterial DNA was isolated from both media and genotyping was done by 24 loci MIRU-VNTR according to standardized protocols [14] . Out of the 500 study participants , the isolates of 436 ( 87% ) were successfully typed and included in this study . Of the 436 patients included in the study , 92 ( 21% ) had complex MIRU-VNTR patterns . We note that , like many other South African communities , the one in this study has a high HIV prevalence . The standard rule-based classification method designated 44 of 92 of the complex strains as clonal and the remaining 48 as mixed , whereas our method classified 50 of them as clonal and 42 as mixed . Only one patient got assigned a probability of 1/4 for clonality and 3/4 for mixed , and we ended up using the majority rule and classifying their infection as mixed . There were a total of 18 discrepant calls , 6 in which a strain was called clonal by the standard method but mixed by ours , and 12 in which the reverse occurred . Our results are summarized in Table 1 . In order to evaluate the performance of the 8 different variants of our method ( defined by the constant or the linear metric , as well as the commonly used Hamming metric and Goldstein metric , and weighted or unweighted loci ) , as well as the standard rule-based method based on the count of complex loci , we produced 100 simulated datasets with characteristics similar to our South African dataset . The details of our simulations are described in the Supplementary Materials ( S1 Text ) . We attempted to match the original dataset in terms of its strain clustering characteristics , distribution of the number of complex loci in strains , and distribution of the differences between the CNVs in a complex locus . To this end , we simulated the evolution of an initial population of strains with random but constrained mutation and reinfection events a large number of times , selected a number of subpopulations of appropriate size , and selected the final datasets according to their distance to the two target distributions . We selected 100 datasets of N = 415 strains each , n = 83 ( 20% ) of which were complex; 42 were the result of clonal heterogeneity and 41 were mixed infections . We applied the standard rule-based method and our method on each dataset . We evaluated the accuracy of each method as the average probability they assigned to the correct classification for the 83 complex strains; namely , if a method returned a probability p of clonality , we scored p if the complex strain was actually clonal and 1 − p if it was actually mixed . The accuracy of the standard rule-based method averaged 48% , not significantly different from the 50% that would be expected from a random classification . On the other hand , the accuracy of ClassTR using our metric of choice , the linear weighted metric , was 80% , for a 61% reduction in error . We also evaluated the correctness of the resolution of complex strains into their constituent simple ones , and found that ClassTR produced the correct resolution in 88% on clonal infections and 95% on mixed infections . The results of running different variants of our method on classification accuracy are shown in Table 2 . In addition , we created 9 groups of 100 datasets each , with similar characteristics but not constrained to resemble the original dataset as closely . Each group corresponded to a combination of one of three mutation rates ( low , medium and high ) and one of three reinfection rates ( low , medium and high ) . Our method outperformed the standard rule-based method on all of them except for the classification at the low mutation rate . The results of running ClassTR on those datasets are shown in Table 3 ( for the resolution , using randomly picked resolutions as the baseline ) and Table 4 ( for the classification , using the standard method as the baseline ) . They suggest that although the resolution performance of ClassTR deteriorates quite substantially on mixed infections at high mutation rates , its classification performance remains consistently good , both in absolute terms and in comparison with the standard method . This finding is further substantiated by Table 5 , which shows that ClassTR finds the correct classification when the correct resolution is given most of the time , with a slight deterioration at high mutation rates . However , it is also able to find the correct classification from an incorrect resolution quite frequently , as evidenced by a comparison of all three tables .
ClassTR is the first alternative method for classifying complex bacterial strains as either clonally heterogeneous or mixed infections . In contrast with the existing rule-based classification method , it includes an explicit model of tandem repeat evolution and utilizes information from other strains collected locally , provides probabilistic rather than deterministic classifications , and allows for the identification of individual strains within complex infections . There are two computational problems that ClassTR solves . The first one , which we called the Parsimonious Resolution Problem and showed to be NP-complete , is reminiscent of haplotyping problems in eukaryotic genomes [22] . The key difference is the haploid nature of the bacterial genomes we analyze; the observed complexity in the strains is the result of the genotyping technique we use rather than an actual allelic variation due to recombination . The second one , which is the classification problem for complex bacterial strains , is reminiscent of the type of problems that arise in cancer genomics [23] in deciding whether particular tumor genotypes are related to one another ( similar to clonal heterogeneity ) or have arisen independently ( similar to mixed infection ) . The key difference is the evolutionary model we adopt for tandem repeats , which may not apply to cancer . At the current time we lack a “gold standard” approach for determining which infections are actually complex , and whether these complex infections are due to within-host mutation or reinfection . This makes it challenging to evaluate the relative performance of the standard rule-based approach and ClassTR . While future advances in genome sequencing are likely to provide additional data to test ClassTR against the standard rule-based approach , our results on simulated data suggest that ClassTR provides more accurate classifications than the standard approach , and the additional computational resources are justified by the improvement in classification accuracy . While the performance of ClassTR and the standard method for classification of complex infections is similar if the true mutation rate of MIRU-VNTR loci is at the lowest end of the plausible range , ClassTR outperforms the standard method under scenarios with higher mutation rates within this range . Furthermore , applying ClassTR to our data from KwaZulu Natal generates results that are substantially different from the standard rule-based method , which demonstrates that the difference between these approaches is not just theoretical . In addition , ClassTR provides the only available approach to extract the constituent strains involved in a mixed infection from MIRU-VNTR data alone . There are several opportunities to modify the models we have used to even better reflect the evolutionary process driving copy number variation . First , our evolutionary model does not account for the fact that some evolutionary events may duplicate multiple segments in a single timestep . A more sophisticated model might allow for such duplication events to happen , albeit at a small rate , and this rate could be estimated from available data . Second , we model copy number increases and decreases symmetrically , whereas a more flexible model could allow these events to occur at different rates . Finally , an alternate model might be needed to account for the possibility that two strains may be simultaneously transmitted from one person to another or for the possibility of having more than two strains within a host , which may be relevant for certain types of infectious pathogens [24] . In addition , an intriguing opportunity for future work would be to investigate how accurate the classifications of complex infections as clonal or mixed could be at the time that each patient is admitted , rather than at the end of the study as we have done here , as well as to take into account the information about the strains found in a patient’s contacts , such as household contacts , perhaps by using these to constrain resolutions . In conclusion , ClassTR is a tool which we believe will advance our capacity to identify the mechanisms underlying within-host heterogeneity in TB and other bacteria . By distinguishing within-host mutation from reinfection , we anticipate that this method will improve our understanding of the natural history of pathogenic infection at the individual patient level , and will improve our ability to project transmission dynamics and the effects of interventions in communities .
In this section we define simple and complex strains and describe the principled way in which ClassTR separates complex strains into simple strains . We formally define a simple strain as a string of length L ( for MIRU-VNTR , L = 12 or L = 24 ) over the alphabet A consisting of all integers from 0 to some upper bound tmax . If s is a simple strain , we denote by sj its j-th symbol . We define a complex strain as a string of length L over the alphabet P ( A ) , the power set of A , so that each of its symbols is a subset of A . If s is a complex strain , we call sj the content of s at position j . A collection C of simple strains will be called a cover for a complex strain s if at each position 1 ≤ j ≤ L , we have sj ⊂ ∪c∈C cj . In other words , the content of the strain s at each locus is included in C . A collection C will be called an exact cover for a complex strain s if equality holds , i . e . sj = ∪c∈C cj ∀1 ≤ j ≤ L; in this case , C includes the content of s at each locus , and nothing else . A collection C will be called a minimal ( exact ) cover of s if C is an ( exact ) cover of s and no proper subset of C is . We always look for minimal exact covers for reasons of parsimony . When a complex strain s has all contents of size 1 or 2 , there exist minimal exact covers of size 2 , and the number of such covers is 2q−1 , where q is the number of positions with content of size 2 . The value of q attains a maximum of 12 in our dataset , meaning that a single complex strain can have up to 2048 different covers . Given the multiplicity of possible minimal covers for each complex strain , we use a global parsimony assumption to identify the ones that are actually present . Namely , we assume that , all other things being equal , the fewer simple strains we add to the ones in the dataset to cover all the complex strains , the better . Intuitively , this means that we attempt to explain complex infections in terms of strains we have observed as simple infections in the population . Thus we seek to cover all the complex strains by adding the smallest possible number of strains . Fig 2 presents a toy example of a dataset with its solution . In this section we formalize the problem of resolving the complex strains by introducing as few new simple strains as possible , which we call the parsimonious resolution problem . In the Supplementary Materials ( S1 Text ) we show that the decision version of this problem is NP-complete , even in the case of all copy number variants being 0 or 1 . As a corollary , our proof establishes that the parsimonious resolution problem for spoligotype data ( where a 0 indicates the absence and a 1 the presence of a particular region ) , a version of which was studied by Lazzarini et al [25] , is also NP-complete . The decision version of the parsimonious resolution problem ( PRP ) can be stated as follows . Given: an integer L; a finite alphabet A; a set of strings S = {s1 , s2 , … , sk} of length L over A ∪ A2 , but not entirely over A ( i . e . each position contains 1 or 2 elements of A , with at least one position containing 2 elements of A ) ; a set of “free” strings F = {f1 , … , fm} of length L over A; an integer K . Decide: whether there exists a collection C = {c1 , c2 , … , cK} of K strings of length L over A , such that , for each string s ∈ S there exist 2 strings c and c′ in C ∪ F , such that c ∪ c′ = s ( where the union is taken component-wise ) . The correspondence between the PRP and the problem we are actually solving is as follows: L is the number of loci , A is the set of possible CNVs , S are the complex infections present in the data and F are the simple infections present in the data . Finally , K is the number of additional ( new ) simple strains we are seeking to add to F in order to resolve all the complex infections . In this section we formulate the 0–1 integer linear program [26] for the parsimonious resolution problem . This integer linear program finds a set of simple strains that cover the complex strains in the dataset , paying for each simple strain that is not present in the dataset . It minimizes the total cost of these newly added simple strains . Its inputs are a set of simple strains that can be used “for free” and the set of complex strains to be covered . Its outputs are the variables corresponding to the new simple strains used in covering the complex strains . Let N be the number of simple strains , n be the number of complex strains , and qi be the number of complex loci in the ith complex strain . For simplicity we assume that there are exactly 2 copy number variants at each complex locus , which is the case for our dataset . Let Si be the set of all simple strains that may be used to cover the ith complex strain , so that |Si| = Qi = 2qi . Let us also define Q ≔ ∑ i = 1 n Q i . Let S ≔ ∪ i = 1 n S i and q := |S| . Note that q ≤ Q . We define two categories of variables , one to indicate usage , and the other to indicate coverage . The usage variables are denoted uj and are defined for every strain j in S . The value of uj is 1 if the simple strain j is used in the cover of at least one complex strain , and 0 otherwise . The coverage variables are denoted cij and are defined for every complex strain i and every simple strain j in Si . The value of cij is 1 if the simple strain j is used to cover the complex strain i , and 0 otherwise . For a complex strain i and a simple strain j in Si , we denote by i∖j the complement of j in i , namely , the simple strain that , together with j , covers i ( here we use the assumption that every complex locus has exactly 2 CNVs ) . The complement i∖j can be obtained by taking the CNV in each complex locus of i that was not used in j . For each simple strain j in S we also define the cost wj of adding it to the cover . The objective function is simply a linear combination of the usage variables uj with the costs wj as coefficients . We always take wj = 0 if the simple strain j is present in the dataset , because it is already available to be used in a cover . We also take wj = 1 for any simple strain j not present in the dataset , so the total cost ends up being the number of new strains used . The optimal solution is the one minimizing this total cost . This leads us to the following integer linear program formulation: Minimize ∑ j w j u j subject to ( 1 ) u j ∈ { 0 , 1 } ∀ j ( 2 ) c i j ∈ { 0 , 1 } ∀ i , j ( 3 ) c i j = c i ( i \ j ) ∀ i , j ( 4 ) c i j ≤ u j ∀ i , j ( 5 ) u j ≤ ∑ i c i j ∀ j ( 6 ) 1 ≤ ∑ j c i j ∀ i ( 7 ) The first two sets of constraints , Eqs ( 2 ) and ( 3 ) , ensure that all the variables take values 0 or 1 . The next set of constraints , Eq ( 4 ) , ensures that the simple strain j is used to cover the complex strain i if and only if its complement simple strain , i∖j , is also used to cover the complex strain i . The next two sets of constraints ensure that the simple strain j is marked as used if ( Eq ( 5 ) ) and only if ( Eq ( 6 ) ) it is used to cover at least one complex strain i . Finally , the last set of constraints , Eq ( 7 ) , ensure that the complex strain i is covered in at least one way by simple strains . The number of uj variables and constraints in Eq ( 2 ) is q ≤ Q; the number of cij variables and constraints in Eq ( 3 ) is Q; the number of constraints in Eq ( 4 ) is Q/2; the number of constraints in Eq ( 5 ) is Q; the number of constraints in Eq ( 6 ) is q ≤ Q; and the number of constraints in Eq ( 7 ) is n , for a total of Q + q ≤ 2Q variables and ( 5/2 ) Q + 2q + n ≤ ( 9/2 ) Q + n constraints . In particular , for our South African dataset , n = 92 and Q is roughly 8 , 000 , while for our simulated datasets , n = 83 and Q varies from 5 , 000 to 25 , 000 , so the total number of variables is always under 50 , 000 and the number of constraints under 100 , 000 . Integer linear programs of this size can typically be solved to optimality in seconds by CPLEX ( available from http://www-01 . ibm . com/software/integration/optimization/cplex-optimizer ) . The total time required by ClassTR is about 5 minutes for the South African dataset with N = 436 strains . We additionally tested our method on a much larger dataset containing N = 4075 strains with n = 364 of them complex . Its processing took less than an hour on a single CPU , suggesting that our algorithm scales well with input size in practice . In this section we define the four distances we use in ClassTR . These distances can be used to construct the predecessor sets which then allow us to calculate the probability of each complex strain being clonally heterogeneous or mixed . We define the constant metric dC between two simple strains as d C ( s , s ′ ) = ∑ j = 1 L | s j - s j ′ | . This corresponds to the minimum number of mutation events needed to get from one strain to the other in the constant model of tandem repeat evolution defined by Aandahl et al [20] . Indeed , since the constant model assumes a Poisson process at each locus , |i − j| is precisely the number of mutations required to get from i to j . We also define the linear metric dL between two simple strains as d L ( s , s ′ ) = ∑ j = 1 L ∑ k = min ( s j , s j ′ ) + 1 max ( s j , s j ′ ) 1 k . This corresponds to the expected number of timesteps needed to get from one strain to the other in the linear model of tandem repeat evolution defined by Aandahl et al [20] . Indeed , since the linear model assumes that a Poisson process takes place at each copy , it takes an expected 1/m timesteps to go from m to m − 1 copies . Two other standard metrics we use are the Goldstein metric dG and the Hamming ( categorical ) metric dH , respectively defined as d G ( s , s ′ ) = 1 L ∑ j = 1 L ( s j - s j ′ ) 2 and d H ( s , s ′ ) = ∑ j = 1 L [ s j ≠ s j ′ ] , where [I] is the Iverson bracket whose value is 1 if expression I is true and 0 otherwise . Note that the Goldstein metric is not a metric in the traditional sense because it does not respect the triangle inequality . In addition we define weighted analogs of all these metrics , which are obtained by multiplying the contribution of each locus by its weight . To estimate the weight of each locus ClassTR uses the Simpson index [27] , also known as the Hunter-Gaston index [28] , reasoning that the more diverse a locus is , the faster it evolves and the less weight it should carry . These weights then allow us to compute the corresponding weighted distances in the constant or linear models defined below , which we denote d c w and d L w , respectively . In our datasets these weights ranged from 0 . 16 to 1 . Given the set of simple strains generated by the optimization , we describe how to produce the final soft classification of complex strains along the clonally heterogeneous to mixed spectrum in this section . We start by choosing a distance function d on simple strains . Given a strain j , we define the predecessor set P ( j ) as the subset of the simple strains S present in the original dataset that are closest to s according to d . Formally , P ( j ) ≔ arg min s ∈ S d ( j , s ) . Of course , for any strain s ∈ S , the predecessor set only contains s itself ( we do not take the presence of duplicate strains into account ) . We also note that the more highly resolving the distance , the smaller the predecessor sets are going to be . Thus , the unweighted constant distance could give rise to ties for the closest strain , but the weighted constant distance or the linear distance is less likely to yield a tie . Intuitively , the more similar the predecessor sets of the constituent strains are to each other , the more likely the complex strain is to be clonally heterogeneous . For example , if two different covering strains are both very close to the same simple strain in the dataset , the complex strain composed of both of them is more likely to be clonal than if the two strains’ nearest matches in the data are two very different strains . We formalize this by using the Jaccard index [29] to evaluate the similarity of two sets A and B , defined as the size of their intersection divided by the size of their union: J ( A , B ) ≔ | A ∩ B | | A ∪ B | . Suppose that A and B are the predecessor sets of the constituent strains of a complex strain . Then we take the Jaccard index of A and B as the probability of the complex strains being clonally heterogeneous . Thus , a complex strain covered by two strains with identical predecessor sets will be classified as clonally heterogeneous , while one with two strains with non-overlapping predecessor sets ( for example one covered by two distinct simple strains present in the original dataset ) will be classified as mixed , with intermediate variants also possible as shown in Fig 3 . This probability is the value we report as our final classification .
|
Within-host heterogeneity of an infection can arise through two distinct mechanisms: within-host mutation and reinfection . While current genotyping techniques based on MLVA ( multiple loci variable-number tandem repeat analysis ) can identify within-host diversity , standard methods for classifying the mechanism driving this diversity have limitations . We present ClassTR , a novel approach for classifying these types of complex infections . ClassTR uses optimization to resolve complex strains into simple strains and explicit models of tandem repeat evolution to classify the infections as clonal ( due to within-host diversification ) or mixed ( due to reinfection ) . We illustrate ClassTR and validate its findings in the context of Mycobacterium tuberculosis infections . We construct simulated datasets to identify the best-performing variant of our method and find that it is significantly more accurate than the standard method of classification . We apply ClassTR to data from a study in South Africa and find substantial differences in the classifications produced by ClassTR and the standard method , demonstrating the real-world relevance of this approach . Our work suggests that an analysis of complex infections based on an evolutionary model improves our understanding of the drivers of within-host diversity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"microbiology",
"optimization",
"mathematics",
"fungal",
"evolution",
"molecular",
"biology",
"techniques",
"genotyping",
"bacteria",
"bacterial",
"pathogens",
"research",
"and",
"analysis",
"methods",
"mycology",
"genome",
"complexity",
"medical",
"microbiology",
"microbial",
"pathogens",
"actinobacteria",
"repeated",
"sequences",
"molecular",
"biology",
"tandem",
"repeats",
"mycobacterium",
"tuberculosis",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"computational",
"biology",
"organisms"
] |
2016
|
ClassTR: Classifying Within-Host Heterogeneity Based on Tandem Repeats with Application to Mycobacterium tuberculosis Infections
|
The genetics of lymphoma susceptibility reflect the marked heterogeneity of diseases that comprise this broad phenotype . However , multiple subtypes of lymphoma are observed in some families , suggesting shared pathways of genetic predisposition to these pathologically distinct entities . Using a two-stage GWAS , we tested 530 , 583 SNPs in 944 cases of lymphoma , including 282 familial cases , and 4 , 044 public shared controls , followed by genotyping of 50 SNPs in 1 , 245 cases and 2 , 596 controls . A novel region on 11q12 . 1 showed association with combined lymphoma ( LYM ) subtypes . SNPs in this region included rs12289961 near LPXN , ( PLYM = 3 . 89×10−8 , OR = 1 . 29 ) and rs948562 ( PLYM = 5 . 85×10−7 , OR = 1 . 29 ) . A SNP in a novel non-HLA region on 6p23 ( rs707824 , PNHL = 5 . 72×10−7 ) was suggestive of an association conferring susceptibility to lymphoma . Four SNPs , all in a previously reported HLA region , 6p21 . 32 , showed genome-wide significant associations with follicular lymphoma . The most significant association with follicular lymphoma was for rs4530903 ( PFL = 2 . 69×10−12 , OR = 1 . 93 ) . Three novel SNPs near the HLA locus , rs9268853 , rs2647046 , and rs2621416 , demonstrated additional variation contributing toward genetic susceptibility to FL associated with this region . Genes implicated by GWAS were also found to be cis-eQTLs in lymphoblastoid cell lines; candidate genes in these regions have been implicated in hematopoiesis and immune function . These results , showing novel susceptibility regions and allelic heterogeneity , point to the existence of pathways of susceptibility to both shared as well as specific subtypes of lymphoid malignancy .
Lymphoid malignancies represent clonal proliferations occurring at various stages of differentiation of B and T cells . B-cell differentiation is characterized by a canonical set of DNA modifications , including somatic hypermutation , class switching , and VDJ recombination . If aberrant , these result in lymphoid neoplasms ranging from less differentiated acute leukemia and lymphoma , to well-differentiated plasma cell malignancies [1] . Some genetic and environmental risk factors for lymphoma have been defined and antecedent autoimmune disorders increase risk for lymphoma several fold [2] . Familial clustering of lymphomas has been observed and may comprise mixed phenotypes of Hodgkin's lymphoma ( HD ) as well as the subsets of non-Hodgkin's ( NHL ) including follicular ( FL ) , diffuse large B-cell ( DLBCL ) , and chronic lymphocytic/small lymphocytic ( CLL/SLL ) [3] . While less common than B cell neoplasms , T cell malignancies are also part of the spectrum of familial lymphoma and may be seen alone or in combination with B cell neoplasms in kindreds with underlying immune deficiency or genomic instability [3] . The lack of genetic linkage to specific loci in such families has prompted the search for common susceptibility variants in the germline , which may provide evidence as to the etiology of these disorders . Genome wide association studies ( GWAS ) examining lymphoma susceptibility have focused on identifying risk loci associated with different subtypes of the disease , based on the a priori assumption that each of the subtypes have distinct biology and therefore , distinct pathogenesis . Thus far , a locus on 6p21 . 33 , near PSOR1 , and another region at 6p21 . 32 , near HLA-DRB1 have been associated with FL [4] , [5] , [6] and Hodgkin's disease [7] , [8] . A smaller study has described CDC42BPB at 14q32 to be associated with diffuse large cell lymphoma [9] . In order to test the paradigm that there are common and subtype specific germline susceptibility loci for lymphoma , we conducted a two-stage genome-wide association study ( GWAS ) . Our stage-1 consisted of 944 cases of lymphoma , including 282 familial cases , and 4044 public shared controls . Stage-2 consisted of 1245 cases and 2596 controls . We have used a higher ratio of controls to cases to enhance power to detect association , as the use of public shared controls comes at no cost [10] . We also analyzed published data for overlap of the GWAS hits to expression quantitative trait loci ( eQTL ) in lymphoblastoid cell lines . Secondary analyses , such as gene set enrichment were carried out to detect enrichment of biologically relevant candidates for further study .
In stage-1 , we analyzed 944 cases of lymphoma , including 275 FL , and 4044 controls and documented strong evidence of association between SNPs on Chr6 , with at least 9 SNPs showing PFL<1×10−7 at the HLA region ( chr6:32 . 17–32 . 89 Mb ) encompassing genes TNXB to HLA-DOB . The results of the stage-1 analysis for LYM , NHL , FL and DLBCL are shown as Manhattan plots ( Figure 1 ) and quantile-quantile ( QQ ) -plots ( Figure 2 ) . FL showed the strongest enrichment of association signals; particularly on Chr6 . We refrained from detailed analysis of smaller subsets , based on the power calculations performed using PGA [11] taking into account sample sizes , detectable relative risk and case to control ratios ( Figure S1 ) . Analysis of the major classifiers LYM and NHL and only the major subgroups FL , DLBCL were performed . In addition , a subset designated as NFD comprised any non-Hodgkin's lymphoma cases that were neither FL nor DLBCL . This subgroup was created to test if the associations in the larger LYM and NHL were driven primarily by the pre-dominant subgroups FL and DLBCL . Among all analyses , the lowest p-values in the FL subset were observed on chromosome 6p . The smallest p-value was for rs2621416 ( PFL = 8 . 69×10−9 , OR 1 . 82 ) ( Table S1 ) followed by rs9268853 ( PFL = 1 . 76×10−8 , OR = 1 . 74 ) . Imputation of the stage-1 data revealed strong associations with FL for the 6p21 . 32 SNP rs12194148 ( PFL = 1 . 18×10−16 , 14 . 5 kb from rs9268853; r2 = 0 . 62 , D′ = 1 . 0 ) , suggesting a subtype specific association with the HLA locus ( Figure 3C ) . In addition to the SNPs on chromosome 6p HLA region , we also found preliminary evidence of association of several SNPs at chromosome 3q25 . 2 with LYM , NHL and NFD . Another locus at 11q12 . 1 was defined by two SNPs with suggestive associations ( P<10−5 ) ( Table S1 , Figure 3B ) . Fifty SNPs were selected from stage 1 for genotyping in a larger set of 1245 lymphomas ( Table S1 ) . After adjusting for age and Jewish ancestry , nine of 50 SNPs had P-values below the nominal alpha level of 0 . 05 , while showing the same direction of effect as observed in stage 1 ( Table S2 ) . After adjusting for the 50 SNPs tested , rs4530903 , at the HLA locus , remained significantly associated with NHL , FL , and DLCBCL . This SNP also appears to be associated with LYM , but the p-value was marginally higher than the Bonferroni corrected threshold . Two other tests were significant after multiple test correction: rs707824 on chromosome 6p23 with NHL and rs12289961 on chromosome 11q12 . 1 with LYM . Thus , two novel susceptibility loci replicated in stage 2 . Notably , the SNPs at 11q12 . 1 also are nominally significant ( P<0 . 05 ) in the NFD subgroup , which is different from the observation for the SNPs at 6p21 . 32 . Based on this analysis , nine of these SNPs were advanced to a meta-analysis of both stage-1 and stage-2 data ( Table 1 ) .
The major finding of this study is the observation that some regions are most strongly associated with a particular subtype of lymphoma , e . g . 6p21 . 32 in FL , while others are most strongly associated with combined types of lymphoma , e . g . the novel regions on 11q12 . 1 . Evidence favoring a model of common susceptibility loci includes observations of familial clustering of multiple subtypes of lymphoma . Several studies have now discovered pre-disposing genetic loci at the HLA region for FL , DLBCL , CLL and HD [4] , [5] , [6] , [7] , [8] and some of these reports highlight the existence of shared susceptibility loci at the individual subtype levels that were studied . Etiologically , patients with HD have a higher risk of developing NHL as a secondary malignancy [14] . Similarly , patients with NHL have a higher risk of developing HD at a later stage [15] . At a molecular level , the model of common susceptibility pathways is supported by recent studies examining the coding sequences and genomes of non-Hodgkin's lymphomas , which have demonstrated increased mutation burden in shared genes [16] , [17] . In addition , recent tumor analysis has demonstrated that DLBCL and FL share somatic mutations in the same chromatin and histone modifying genes , MLL2 and MEF2B , respectively [16] . Such evidence notwithstanding , a direct test of subtype-specific association would require a very large number of cases per subtype , feasible as part of a combined consortium approach . However , as a first approximation of shared versus subtype specific susceptibilities to lymphoma , it is possible to determine if a putative locus shows heterogeneity . For the 11q12 . 1 region shown here to be a pan-lymphoma susceptibility locus , there was no evidence of such heterogeneity within the largest subtypes . Of the susceptibility markers reported here , the 6p21 . 32 HLA II region has been previously associated with FL and NHL [4] , [5] , [6] . In our report , the 6p21 . 32 region was implicated by three SNPS; rs4530903 upstream from HLA-DRB1 and HLA-DQA1 , rs2621416 upstream of HLA-DQB2 , and rs9268853 downstream of HLA-DRA , HLA-DRB5 and HLA-DRB1 , but upstream of BTLN2 . rs2621416 and rs9268853 have also been associated with risk for ulcerative colitis [18] and rheumatoid arthritis [19] respectively , both of which increases risk for certain types of lymphoma . Allelic heterogeneity at this same locus has also been demonstrated in FL , with both protective and risk alleles described [6] . rs2647012 , a previously reported SNP [6] is correlated ( r2 = 1 , D′ = 1 ) with rs2647046 in our results . None of the 6p21 . 32 SNPs are correlated with rs10484561 , the HLA-associated SNP previously described [4] . Our data support the earlier findings of allelic heterogeneity at this region , with a slightly stronger magnitude of the effect size . The novel regions reported here include 6p23 and 11q12 . 1 , represented by SNPs mapping near genes with biologically plausible ties to lymphoid development . The novel SNP at 6p23 , rs707824 , is upstream of JARID2 , encoding Jumonji , which co-localizes with the polycomb repressive complex 2 and H3K27me3 on chromatin and plays a role in self-renewal and differentiation of embryonic stem cells [20] . JARID2 is regulated by miR-155 where very high levels decrease endogenous JARID2 mRNA levels [21] . High levels of miR-155 are observed in different types of B-cell lymphomas ( DLBCL , HD and latency type III EBV-positive Burkitt lymphoma ) , and transgenic mice expressing miR155 at the late pro-B-cell stage of differentiation developed B-cell tumors . JARID2/Jumonji-deficient mice have widespread developmental defects including abnormalities of hematopoiesis [22] . rs707824 is located downstream of CD83 . CD83 antigen , also known as B-cell activation protein , is expressed on dendritic cells and is thought to have roles in the modulation of antigen presentation and CD4+ T cell generation [23] . The 11q12 . 1 region reported here was marked by two SNPs , rs948562 , located within the non-coding gene ZFP91 , and rs12289961 . rs12289961 at 11q12 . 1 is 230 kb upstream of the LPXN ( leupaxin ) locus , originally identified binding to alpha4 integrins and playing a role in integrin-mediated cell adhesion [24] . LPXN was found to be a member of a fusion protein with RUNX1 in human acute leukemia where wild-type LPXN was shown to transform NIH 3T3 cells [25] . Particularly relevant to its putative role suggested here in B-cell lymphomagenesis , LPXN is preferentially expressed in hematopoietic cells and plays an inhibitory role in B-cell antigen receptor signaling and B-cell function [26] . eQTL analysis showed that there was overlap between the most significant SNPs in the GWAS and lymphoblastoid cell lines cis-eQTL candidate genes , such as HLA-DQA2 and TAP2 . HLA-DQA2 plays a pivotal role in the immune system by presenting peptides derived from extracellular proteins . Gene set enrichment analysis showed interesting candidates related to lymphomagenesis and hematopoietic cell development in the top 20 significant genes . The one variant common in all gene enrichment analyses was RELN , which has been shown to be recurrently mutated in acute lymphocytic leukemia [27] . Based on patterns of inheritance of multiple subtypes of lymphoid neoplasms in families , as well as from the GWAS data reported here , there is evidence to suggest that multiple phenotypes of lymphoma may be associated with shared common genetic predispositions . The candidate genes uncovered in this GWAS suggest that in addition to the genes involved in immune regulation , such as HLA and JARID2 , those involved in B-cell development ( e . g . LPXN ) are logical targets for further studies . It is possible that the GWAS associations with multiple phenotypes reported here have resulted from the ascertainment utilized , since the study was enriched with a familial subset of samples . However , we included only one individual from each kindred , precluding a spurious association of a single SNP with multiple phenotypes in the same family . SNPs that show shared susceptibility , including some of those discovered here , may yet have strongest association with specific lymphoma subtypes . While this study reports associations within combined smaller subtypes , e . g . mantle cell and marginal zone lymphoma , larger sample sizes will be required to delineate whether these and other associations are shared or subtype specific . Thus , we have described two novel lymphoma-susceptibility regions , one at 11q12 . 1 and another putative susceptibility locus at 6p23 , and further characterized the 6p21 . 32 ( HLA class II ) association signal observed in a prior GWAS of FL . While genetic susceptibility to lymphoma has been viewed as subtype specific , here we propose an alternate model . Based on our analysis of the overlap between genotypes and phenotypes ( Figure S5 ) , we predict that the shared loci associated with multiple subtypes of lymphoma will be less frequent than subtype-specific susceptibilities . Finally , the effect sizes observed in this report ( 0 . 59–1 . 93 ) are somewhat higher than those previously reported , e . g . for breast and colon cancer , but well below thresholds required for clinical utility [28] . As in other cancer genome-wide association studies , the novel loci reported here harbor interesting genes in pathways that regulate hematopoiesis , offering potential new insights into the pathogenesis of lymphoid neoplasms .
All cases were ascertained through Memorial Sloan-Kettering Cancer Center IRB-approved protocols , or a protocol approved by the IRB at the Dana Farber Cancer Institute or Hadassah Hebrew University ( Table S4 ) . These protocols either required informed consent for identified use of specimens for research into the genetic basis of lymphoma , or allowed research use of specimens permanently de-identified prior to genotyping . The stage-1 of our study was comprised of 944 unrelated probands . This ascertainment was enriched to included 282 cases of familial lymphoproliferative syndrome , defined as two or more lymphoid cancers in the same lineage . These kindreds were characterized by mixed phenotypes of lymphoid malignancy ( Figure S6 ) , and kindreds contained from 2 to 5 affected relatives . In addition , stage-1 contained 107 cases of lymphoma with a first degree relative affected by a lymphoid malignancy , and 347 cases of early onset ( age of diagnosis <45 years ) lymphoma . Stage 2 was comprised of 1245 unrelated lymphoma probands from a prevalent ascertainment at MSKCC and unselected for specific histology or family history of lymphoma . Lymphomas were categorized according to a modification of the 2008 World Health Organization classification system; primary reports were obtained in all cases and reviewed by two of the authors ( KO and AZ ) . Because of the presence of multiple subtypes in kindreds with familial lymphoma , all subtypes of B and T cell lymphoma , including Hodgkin's disease and plasma cell neoplasms were included in both stage 1 and 2 , although it was recognized that sizes of these subgroups would be too small to allow subset analysis . The sample distribution of histologic subsets of lymphoma mirrors the prevalence of the disease subtypes in the US population . Genotyping of the cases was performed utilizing the Affymetrix 6 . 0 SNP array . For control data , Bipolar and GENEVA Diabetes Study ( NHS/HPFS ) data were downloaded from dbGAP ( accession phs000017 . v3 http://1 . usa . gov/xrXL1D and phs000091 . v2 http://1 . usa . gov/yevUOY ) . Affymetrix SNP 6 . 0 CEL files were arranged according to the batches in which data were originally genotyped . Data were initially quality checked for the gender and Mini-DM thresholds . Only CEL files that passed a Mini-DM >85% were used in the full Birdseed [29] genotyping of the 906 , 000 SNPs . The mean heterozygosity of each sample was computed ( 26 . 8 ) and samples with low or high heterozygosity were excluded . Samples that passed >95% Birdseed calls were further processed to generate PLINK [30] formatted files , using only calls that had copy number state two and a confidence score >0 . 9 . This was performed using the utility Birdsuite to PLINK from Broad Institute . Hapmap controls were removed . In addition , any sample that showed abnormal copy number profile states in Birdsuite were excluded ( CN0% , CN1% , CN2% , CN3% and CN4% ) . Particular attention was paid to any samples that had the CLL/SLL phenotype in the copy number variability screen , to exclude samples with somatic mosaicism caused by circulating tumor cells . Individuals from dbGaP marked as controls in the data-manifest were retained for further study . Samples with genetic or cryptic relatedness were excluded by using the relationship score-matrix ( PI_HAT<0 . 1 ) in the entire dataset . Data was filtered for multi-mapping , mitochondrial and monomorphic SNPs on the Affymetrix 6 . 0 SNP Chip . Individuals and SNPs were filtered for 95% genotyping rate and departures from Hardy-Weinberg equilibrium [31] . SNPs were also removed if they failed differential missing or haplotype-based differential missing tests as implemented in PLINK . Finally , the data was matched against previously called genotyping data from dbGAP for a subset of SNPs and their allele frequencies . Analyses were carried out on 944 cases and 4044 controls on 530 , 583 SNPs . Principal component analysis was carried out to test for population match in both cases and controls ( Figure S7 ) . Association was performed using case-control status with each phenotype specifically defined , along with age and the first four eigenvectors from the output of EIGENSTRAT [32] program using logistic regression . Controls for the replication were gathered from the New York Cancer Project ( NYCP ) , which is a study of 18 , 000 New York City residents that allows researchers to better understand how factors such as environment , lifestyle , diet , family health history , and genetics affect the development of cancer and an array of other life threatening diseases . The data include age , gender , history of cancers ( including lymphoma ) and ethnicity [33] . All subjects consented to use of samples to study the genetics of any disease state . Only samples with self-declared European ancestry were used for stage-2 . Since individuals of Ashkenazi Jewish ethnicity formed a subset of both ascertainments , ethnicity was used as one of the covariates in the analysis in stage-2 . Genotyping for stage-2 was carried out by designing multiplexed PCR using Sequenom iPLEX assays and analyzed using MassARRAY [34] . Genotypes were called using TYPER 4 . 0 . 2 software . The dataset ( BED , BIM , FAM ) was split to each chromosome , then subset using gtool [35] to create . gen and . sample files . Imputation was done using pre-phasing and best-guess imputing using IMPUTE2 [36] with the references used being 1000 genomes and Hapmap3 populations for genome build v36 . Best practices for imputation of the data were followed The dataset ( BED , BIM , FAM ) was split to each chromosome , then subset using gtool [35] to create . gen and . sample files . Imputation was done using pre-phasing and best-guess imputing using IMPUTE2 [36] with the references used being 1000 genomes and Hapmap3 populations for genome build v36 . Best practices for imputation of the data were followed [37] . The dosage output was filtered for confidence scores and analyzed using PLINK , filtered on INFO and plotted using locuszoom [38] . Haplotypes were viewed in Haploview [39] . The dosage output was filtered for confidence scores and analyzed using PLINK , filtered on INFO and plotted using locuszoom [38] . Haplotypes were viewed in Haploview [39] . SNPs were ranked on p-value in both major types and subtype specific analyses . Each index-ranked SNP ( within top 100 SNPs ) was graded based on a custom script used to generate scatterplots from Birdsuite , which were inspected and graded on the cluster separation and skew . In order to prioritize the SNPs that were to be replicated , SNPs were given a negative grade if they were singletons ( i . e . neighboring SNPs not showing low p-values ) . A positive grade was given if a given SNP showed low p-value ( P<5×10−4 ) in any other type or subtype . Only SNPs with good scatterplots were selected for the iPLEX design . Analysis was performed by logistic regression using the same criteria as stage-1 , however , instead of the PCA , self-reported ethnicity information was used . Only Caucasian samples were used in the replication study . A meta-analysis of the stage-1 and stage-2 data was performed using the results of the logistic regression . For test of heterogeneity specifically for the 6p21 . 32 locus , the combined dataset consisting of stage-1 and stage-2 was split into three major groups namely FL , DLBCL and any other NHL subgroup designated as NFD in this report . Since we have only one control set , the control samples were randomly assigned in a fixed ratio to match the percent cases per subset without replacement . The three clusters were joined together to perform Breslow-Day test using PLINK . We performed gene set enrichment analysis using the p-values from each of the subgroup and group analyses . The program VEGAS [40] was used to compute the gene enrichment analyses . It annotates SNPs to corresponding genes ( ±50 kb boundaries ) , produces a gene-based test statistic , and then uses simulation to calculate an empirical gene-based p-value . The Hapmap population was used as a reference . The top 10 percent of significant SNPs were chosen for the analysis with simulation performed 106 times . Venn diagram was created using Venny ( http://bioinfogp . cnb . csic . es/tools/venny/index . html ) . We analyzed available hapmap3 population data from lymphoblastoid cell lines [12] for eQTLs [12] using GENEVAR [13] . Two types of analyses were performed , ( 1 ) identifying cis-eQTLs in candidate genes discovered from the GWAS and ( 2 ) SNP-gene association analysis . Adjusted p-values ( Padj ) were derived from 10 , 000 permutations as implemented on the GENEVAR applet .
|
B-cell lymphomas comprise several diseases representing aberrant proliferations of immune cells at various stages of maturation . It might be expected that dissimilar subtypes of lymphoma will have different etiologic and pathogenic mechanisms , reflecting the distinct histologic and clinical characteristics of these diseases . This study aims to define both shared as well as specific genetic risk factors for lymphoma . Utilizing a genome-wide approach , we discovered novel locations in the genome associated with risk for lymphoid malignancies . Common variants in these regions , on chromosome 11q12 . 1 and 6p23 , were each associated with a modest modification of risk for lymphoma . These regions harbor several genes of biological importance in lymphoid maturation and function . We also further characterized the HLA region at 6p21 . 32 , previously associated with lymphoma risk and thought to be important in immune function . Some of the associated SNP markers were specific for one common subtype of lymphoma , e . g . follicular lymphoma . However , others were associated with combined subsets of disease , suggesting that there are both shared and subtype-specific associations between common genetic variants and human lymphoid cancer . Secondary analyses showed that the two novel regions harbor candidates that are biologically relevant and that regulate cell development and hematopoiesis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"cancer",
"genetics",
"genome",
"scans",
"population",
"genetics",
"genome",
"analysis",
"tools",
"trait",
"locus",
"analysis",
"population",
"biology",
"genetic",
"polymorphism",
"biology",
"genetics",
"genomics",
"genetics",
"of",
"disease",
"genetics",
"and",
"genomics",
"human",
"genetics"
] |
2013
|
Susceptibility Loci Associated with Specific and Shared Subtypes of Lymphoid Malignancies
|
Germline stem cell ( GSC ) self-renewal and differentiation into gametes is regulated by both intrinsic factors in the germ line as well as extrinsic factors from the surrounding somatic niche . dWnt4 , in the escort cells of the adult somatic niche promotes GSC differentiation using the canonical β-catenin-dependent transcriptional pathway to regulate escort cell survival , adhesion to the germ line and downregulation of self-renewal signaling . Here , we show that in addition to the β-catenin-dependent canonical pathway , dWnt4 also uses downstream components of the Wnt non-canonical pathway to promote escort cell function earlier in development . We find that the downstream non-canonical components , RhoA , Rac1 and cdc42 , are expressed at high levels and are active in escort cell precursors of the female larval gonad compared to the adult somatic niche . Consistent with this expression pattern , we find that the non-canonical pathway components function in the larval stages but not in adults to regulate GSC differentiation . In the larval gonad , dWnt4 , RhoA , Rac1 and cdc42 are required to promote intermingling of escort cell precursors , a function that then promotes proper escort cell function in the adults . We find that dWnt4 acts by modulating the activity of RhoA , Rac1 and cdc42 , but not their protein levels . Together , our results indicate that at different points of development , dWnt4 switches from using the non-canonical pathway components to using a β-catenin-dependent canonical pathway in the escort cells to facilitate the proper differentiation of GSCs .
Stem cell self-renewal and differentiation are critical for maintaining the organ systems of multicellular organisms . Loss of stem cell self-renewal leads to aging , due to an inability to replenish these organs; while loss of differentiation leads to tumors , which can progress toward diseases such as cancer [1–3] . Thus , identifying triggers of stem cell differentiation is pivotal for understanding the etiology of degenerative diseases and cancer . Germline stem cells ( GSCs ) self-renew and differentiate to produce gametes [4–7] . The balance between self-renewal and differentiation is critical for a steady supply of gametes for increased reproductive success . Dysregulation of GSC self-renewal and differentiation manifests itself as changes in fecundity without markedly altering the organism’s growth or survival . Additionally , processes regulating GSC differentiation are conserved in other stem cell systems [8–10] . Therefore , GSCs make an excellent system to identify triggers of stem cell differentiation . Drosophila GSCs are well characterized and genetically tractable [5] . GSC development starts during the larval stages . The female larval gonad is made up of the somatic niche and primordial germ cells ( PGCs ) [11 , 12] . The late larval somatic niche is comprised of intermingled cells ( ICs ) , the terminal filament and the cap cells [11–14] . The PGCs that give rise to the GSCs in the adults are interspersed with ICs , precursors of the adult escort cell [11 , 15 , 16] ( Fig 1A ) . Most of the PGCs remain undifferentiated during the larval stage [15 , 17 , 18] . As development progresses , the larval gonad transforms first into a pupal gonad and then into an adult ovary . During this transition , the PGCs closest to the cap cells acquire a stem cell fate and become GSCs while the rest directly differentiate [15 , 18] . Each adult ovary consists of 16–18 individual units called ovarioles . GSCs are located in the germarium , present at the anterior end of each ovariole [19] . GSCs divide to give rise to a self-renewed GSC and a stem cell daughter , the cystoblast ( CB ) ( Fig 1B ) . The CB expresses a differentiating factor called Bag of marbles ( Bam ) , undergoes differentiation and four incomplete divisions to form a sixteen-cell cyst . Of the sixteen cells , fifteen become nurse cells and one is specified as an oocyte [20–24] ( Fig 1B ) . Loss of GSC differentiation results in loss of or delayed progression towards becoming an oocyte . Soma-germ line interaction is critical for proper GSC self-renewal and differentiation [8 , 25 , 26] . Both PGCs in the larval gonad and GSCs in the adults are surrounded by somatic cells that constitute the somatic niche . Close contact and coordinated signaling between the surrounding somatic niche and the germ line is pivotal for self-renewal and differentiation [26 , 27] . In the larval stages , ICs regulate PGC proliferation [11] . Additionally , proper intermingling of ICs in the larval gonad promotes proper escort cell function in the adults [28 , 29] . In the adults , the somatic niche can be broadly divided into two regions; the self-renewal niche and the differentiation niche [26 , 30] . The terminal filament and cap cells comprise the self-renewal niche that is required for regulating GSC self-renewal ( Fig 1B ) [24 , 26 , 31–36] . Loss of adherens junction proteins , such as DE-Cadherin and Armadillo/β-catenin , in either germline or somatic cells within the self-renewal niche , leads to loss of contact and thus loss of GSC self-renewal [37 , 38] . The differentiation niche , formed by the escort cells , makes extensive contact with CBs by means of escort cell protrusions that encapsulate the CB and promote GSC differentiation ( Fig 1B ) [39–41] . Thus , soma-germline contact throughout development is a critical extrinsic cue that controls GSC differentiation . Signaling from the somatic niche coordinates the balance of self-renewal and differentiation . In the larval gonad , EGF/Spitz signaling from the PGCs is required for IC survival and intermingling [11] . In the adult germaria , decapentaplegic ( dpp ) signaling in the self-renewal niche regulates GSC self-renewal by phosphorylating the transcriptional regulator , mothers against dpp ( pMAD ) , to repress expression of the differentiation factor , Bam [20 , 24 , 31] . Ecdysone and dWnt4 , among others , have been shown to function in the differentiation niche to promote the differentiation of the CB . Ecdysone signaling in the escort cells regulates the formation of escort cell protrusions through an unknown mechanism [42 , 43] . dWnt4 autocrine signaling in the escort cells is required for escort cell survival , repression of dpp expression and for formation of escort cell protrusions to encapsulate the CB [44–46] . Canonical Wnt signaling plays a critical role in the adult differentiation niche . Wnts can signal through either a canonical or a non-canonical pathway to affect downstream changes [47] . In the escort cells , dWnt4 is known to regulate differentiation through the canonical pathway [44–46] . dWnt4 binds to the receptor , Frizzled2 and co-receptor , Arrow ( Arr ) , and relays a signal to the multidomain protein , Dishevelled ( Dsh ) , leading to stabilization of β-catenin [47–50] . β-catenin , a cytoskeletal protein that is also a transcription factor , translocates into the nucleus and initiates transcription of downstream targets including Frizzled3 ( Fz3 ) [51–57] . Depletion of any of the canonical pathway factors , such as β-catenin or co-receptor arr , specifically in the escort cells leads to loss of CB differentiation by regulating escort cell number [44–46] . Additionally , loss of dWnt4 leads to downregulation of a Wnt canonical reporter , Fz3 promoter fused to RFP ( Fz3RFP ) , and adhesion molecules , β-catenin , Innexin2 and DE-Cadherin [46] . These adhesion molecules promote escort cell encapsulation of the CB that is required for its differentiation [46] . Thus , dWnt4 uses components of the canonical pathway to control both the number and function of escort cells . Wnt ligands , such as dWnt4 , can modulate the Planar Cell Polarity ( PCP ) of cells [58] . Polarity within a plane of individual cells or of tissues is known as PCP . PCP can specify the proximal and distal end of a cell , as in the Drosophila wing , or the dorsal-ventral and anterior-posterior side of a cell , as in the Drosophila eye [58–64] . Two systems are known to independently regulate PCP–the core components and the Daschous ( Ds ) /Fat ( Ft ) system [65 , 66] . The core components consist of six proteins—Frizzled ( Fz ) , Dishevelled ( Dsh ) , Diego ( Dgo ) , Van Gogh ( Vang ) /Strabismus ( Stbm ) , Prickle ( Pk ) , and Flamingo ( Fmi ) . These proteins form two complexes that are asymmetrically arranged on the opposite ends of a cell–the Fz , Dsh and Dgo , and the Vang/Stbm , Pk complex . Fmi is enriched on both sides of the cell and stabilizes the complexes [59–61 , 67] . Small Rho GTPases such as Ras homolog gene family , member A ( RhoA ) , Ras-related C3 botulinum toxin substrate 1 ( Rac1 ) and Cell division control protein 42 ( cdc42 ) can function as the downstream effectors of the core complex [60 , 68–70] . The Ds/Ft system consists of two atypical transmembrane cadherin proteins–Ds and Ft . Ds and Ft are present on opposite ends of each cell and bind to each other at cell-cell junctions . A golgi associated kinase , Four Jointed ( Fj ) phosphorylates the extracellular domains of Ds and Ft and thereby promotes their binding [66 , 71] . The role of the core components and the Ds/Ft system has been well established in specifying PCP in the wing and the eye . Binding of Wnt to Frizzled2 , can also activate pathways downstream of Dsh: Dishevelled Associated Activator of Morphogenesis 1 ( DAAM1 ) , Ras homolog gene family member A ( RhoA ) , Rac1 and cdc42 pathway . This signaling cascade is also referred to as Wnt/Fz non-canonical pathway . Activated RhoA activates Rho associated kinase ( ROCK/dRok ) [47 , 61 , 72–75] . Together , these pathways regulate the cell division , cytoskeleton , cell movement and cell polarity [76–81] ( Fig 1C ) . It has been previously shown that , like dWnt4 mutants , flies carrying mutations in dsh1 , a PCP specific allele , show disrupted oogenesis [82] . In contrast , loss of other genes that regulate PCP such as fz , vang/stbm , fmi , dgo , and pk do not show such a defect , suggesting that PCP might not be operating during oogenesis [82] . However , it is not known if and how downstream components of the Wnt/Fz non-canonical pathway promote oogenesis . Here , we find that RhoA , Rac genes and cdc42 , the downstream Wnt non-canonical components , but not members of the core components that define PCP , such as vang , ds , fz and ft , regulate CB differentiation . We find that dWnt4 genetically interacts with the downstream non-canonical components: dsh1 , DAAM1 , RhoA , Rac1 and cdc42 . RhoA , Rac1 and cdc42 are expressed at higher levels and are more active in the ICs of the late larval gonad than in the adult escort cells . Conversely , the Wnt canonical reporter is not expressed in the late larval stages but is expressed in the adults . Consistent with this observation , we find that these downstream non-canonical pathway components are required in the late larval stages for regulating intermingling of ICs with the germ cells but their function is not essential in the adults to regulate differentiation . Additionally , we find that dWnt4 regulates the activity , but not the protein levels , of RhoA , Rac1 and cdc42 in the larval gonad . Thus , dWnt4 regulates assembly of the somatic differentiation niche of the GSCs by switching the mode of signaling during development .
To determine if the Wnt non-canonical pathway components regulate differentiation , we depleted these components in the escort cells . We made use of RNA interference ( RNAi ) in conjunction with c587-GAL4 , which is expressed in the escort cells , to deplete dsh , DAAM1 , RhoA , Rac1 and cdc42 [43 , 83 , 84] . Both control and mutant germaria were stained with 1B1 and Vasa . 1B1 marks the endoplasmic reticulum rich organelle , the spectrosome , in undifferentiated GSCs and CBs as well as in differentiating CBs; and it marks the branched structures , the fusomes , in differentiated cysts [85] . 1B1 also marks somatic cell membranes [86] . Vasa , an RNA helicase , marks the germ line [87] . We assayed for differentiation defects , defined as an accumulation of greater than 3 undifferentiated cells marked by spectrosomes . Depletion of dsh , DAAM1 , RhoA , Rac1 and cdc42 resulted in germline differentiation defects compared to control ( Fig 1D–1J ) ( S1A Fig ) ( Table 1 ) . We also depleted the mediators that define the PCP of a cell utilizing previously validated RNAi lines in conjunction with c587-GAL4 . We found that for vang , ds , ft and fz , we did not observe any significant differentiation defect [42 , 88–90] ( S1B–S1I Fig ) ( Table 1 ) . These results are consistent with Cohen et al’s observation that mutations in genes that regulate PCP of a cell , such as fz , stbm , fmi , dgo and pk , do not lead to defects in oogenesis [82] . Thus , we can conclude that downstream components of the Wnt non-canonical pathway such as dsh , DAAM1 , RhoA , Rac1 and cdc42 are required in the escort cells for promoting GSC differentiation . Additionally , these results also suggest that the core components that mediate PCP such as vang/stbm , ds , fz and ft do not play a significant role in the escort cells to regulate GSC differentiation . DAAM1 , RhoA , Rac1 and cdc42 are critical proteins required for various basic cellular processes , such as progression of the cell cycle and cell movement [76–80 , 91] . It is not surprising that they are required in the escort cells for proper function . We asked if these proteins act downstream of dWnt4 to promote differentiation . To answer this , we generated trans-heterozygous flies that contained one genetically reduced copy of dWnt4 and one reduced copy of dsh , DAAM , RhoA , Rac1 or cdc42 [92–96] . Dsh is a multi-domain protein that uses distinct domains to interact with either the Wnt canonical or the Wnt non-canonical proteins [97] . We used dsh1 , a PCP specific allele , to test if dWnt4 uses the downstream components of the PCP system [68 , 98] . If the downstream PCP components act through the same pathway , then these trans-heterozygotes could exhibit GSC differentiation defects , but if they act parallel or independent of dWnt4 , then these trans-heterozygotes should not exhibit differentiation defects . We stained these trans-heterozygote ovaries for 1B1 and Vasa to assay for differentiation defects . Compared to 17 . 5% ( n = 50 ) of heterozygous dWnt4 germaria , 98% ( n = 50 ) of dsh1;dWnt4 trans-heterozygote , 85% ( n = 50 ) of DAAM;dWnt4 trans-heterozygote , 80% ( n = 50 ) of RhoA , dWnt4 trans-heterozygote , 78% ( n = 50 ) of Rac1;dWnt4 trans-heterozygote , and 68% ( n = 50 ) of cdc423;dWnt4 trans-heterozygote germaria showed accumulation of spectrosomes ( Fig 1K–1Q ) ( S2A–S2E Fig ) ( Table 1 ) . As Dsh acts downstream of Wnt signaling , we also stained dsh1 , RhoA and Rac1 heterozygous , and trans-heterozygous flies with one mutant copy of dsh1 and one mutant copy of RhoA or Rac1 for 1B1 and Vasa . Compared to heterozygous dsh1 , RhoA and Rac1 germaria , we found significantly higher number of spectrosomes in dsh1 mutants , dsh1;RhoA trans-heterozygotes , and dsh1;Rac1 trans-heterozygotes ( S2D–S2H Fig ) ( Table 1 ) . Taken together , these results suggest that dWnt4 can function via dsh and the downstream Wnt non-canonical components to regulate GSC differentiation . If dWnt4 acts through the downstream Wnt non-canonical components to regulate GSC differentiation , then dWnt4 mutants and RhoA , Rac1 and cdc42 depleted escort cell mutants should phenocopy each other . Loss of dWnt4 leads to an accumulation of pre-CBs that do not express Bam due to loss of escort cell encapsulation and a reduction of escort cell number [44–46] . To test if depletion of the downstream non-canonical components in the escort cells also results in an accumulation of pre-CBs , we stained control and mutants for pMAD , which marks the GSCs , and used GFP under the control of bam promoter to analyze bam transcription and hence mark the differentiating progeny [99] . In addition , we also stained control and mutants for pMAD and BamC that marks proper translation of bam in the differentiating progeny . Similar to dWnt4 mutants , RhoA , Rac1 and cdc42 depleted escort cell mutants showed an accumulation of pMAD negative , and BamGFP and BamC negative cells , suggesting that these mutants accumulate pre-CBs ( S3A–S3E Fig ) ( S4A–S4D2 Fig ) . Interestingly , in addition to early pre-CBs , cdc42 mutants also showed an accumulation of BamGFP and BamC positive CBs , suggesting that cdc42 may have an additional role in promoting CB differentiation . Pre-CBs that accumulate in dWnt4 mutants are capable of differentiation upon ectopic expression of bam . To test if the undifferentiated germ cells that accumulate due to depletion of downstream components of PCP in the escort cells are also capable of differentiating , we ectopically expressed bam by using a transgene that expresses bam under the control of heat-shock promoter ( hs-bam ) [21] . We found , post heat-shock , these mutants showed loss of undifferentiated cells and an accumulation of cysts , as marked by the presence of fusomes comparable to the control ( 90% for c587-GAL4; 86% for c587-GAL4>RhoA RNAi , P-value = 0 . 5353; 76% for c587-GAL4>Rac1 RNAi , P-value = 0 . 0629 and 92% for c587-GAL4>cdc42 RNAi , P-value = 0 . 7263 {for all n = 50} ) ( S5A–S5E Fig ) . We also analyzed the expression of Bruno in RhoA , Rac1 and cdc42 depleted escort cell mutants that carry the hs-bam transgene , without heat-shock and post heat-shock . Bruno , a translational repressor , is expressed at very low levels in the undifferentiated cells but is expressed at high levels post-differentiation from 16-cell cysts onwards [100 , 101] . We found , post heat-shock , cysts in Rac1 and cdc42 mutants that carry the hs-bam transgene expressed Bruno , while cysts in RhoA mutants carrying the hs-bam transgene only weakly expressed Bruno ( S5F–S5M1 Fig ) . Altogether , these results suggest that the downstream Wnt non-canonical pathway components are required extrinsically in escort cells , to promote differentiation in the germ line . Loss of escort cell number and their encapsulation results in loss of pre-CB differentiation [39–41 , 102] . To determine if loss of the downstream Wnt non-canonical pathway components leads to loss of encapsulation , we visualized the cytoplasmic protrusions using FaxGFP , a somatic cell membrane marker [39] . It has been previously demonstrated that overexpression in the escort cells of dominant-negative Rho ( RhoDN ) , that disrupts Rho function , leads to loss of escort cell encapsulation [40] . We found that , similar to dWnt4 mutants and germaria where RhoDN was overexpressed in the escort cells , depletion of RhoA , Rac1 and cdc42 in the escort cells also resulted in loss of encapsulation ( Fig 2A–2D1 ) . To determine if this was due to decreased number of escort cells , we counted the number of Tj positive escort cells in these mutants . We found that these mutants exhibited significantly lower number of escort cells ( 16 . 5 ± 2 . 1 for c587-GAL4; 9 . 9 ± 3 . 7 for c587-GAL4>RhoA RNAi , P-value = 0 . 00013; 11 ± 1 . 5 for c587-GAL4>Rac1 RNAi , P-value = 4 . 1403E-06 and 9 . 3 ± 2 . 2 for c587-GAL4>cdc42 RNAi , P-value = 9 . 5478E-07 {for all n = 10} ) . However , the escort cells that were present in the germaria failed to extend cytoplasmic protrusions . These results demonstrate that like dWnt4 , downstream Wnt non-canonical pathway components in the escort cells regulate both encapsulation and their numbers , and thus CB differentiation [44–46] . These results taken together suggest that dWn4 acts through Dsh and the downstream Wnt non-canonical pathway components to regulate differentiation . Wnt proteins are known to regulate signaling through either the canonical or the non-canonical pathway [47] . Surprisingly , our data suggests that dWnt4 also acts through the non-canonical pathway components in the escort cells . There is little precedent for canonical and non-canonical Wnt pathways acting in the same cell at the same time . We therefore hypothesized that either dWnt4 signals via the canonical or non-canonical arm in distinct subsets of escort cells or at different stages of escort cell development . If there are subsets of cells that respond to canonical signaling but not to non-canonical signaling , then we could observe the canonical reporter to be present only in some escort cells . If the regulation is temporal , we hypothesized that we could see differences in either levels or activity of canonical and non-canonical reporters as a function of development . To test these hypotheses , we analyzed the expression patterns of downstream Wnt non-canonical pathway components and a Wnt canonical reporter during late larval stages , late pupal gonads and in the adult ovaries . Transgenic lines with GFP tagged RhoA , Rac1 or cdc42 that report endogenous expression were used for the Wnt non-canonical components [103] . These reporters were stained for their respective fluorescent proteins and Tj , which marks ICs in the larval gonad and all somatic cells , except for the terminal filament , in the adult gonad [29] . Vasa was used to mark the germ line . We found that RhoA , Rac1 and cdc42 showed high expression in the larval gonad ( Fig 3A–3B1 and 3D ) ( S6A–S6D1 Fig ) . In contrast , we found that RhoA , Rac1 and cdc42 were expressed at low levels in the escort cells of the pupal and adult germaria ( Fig 3C and 3D ) ( S6E–S6I1 Fig ) . This different expression pattern suggests that the Wnt non-canonical pathway members are regulated in a temporal manner in the escort cells and their precursors during oogenesis . RhoA , Rac1 and cdc42 are members of the Rho family of small GTPases . These proteins are inactive when bound to GDP and active when bound to GTP [104 , 105] . The active forms of these proteins regulate various downstream signaling cascades [76 , 77 , 106 , 107] . We observed that downstream components of the Wnt non-canonical pathway , RhoA , Rac1 and cdc42 are expressed at high levels in the larval gonad . We asked if these proteins are present in their activated form in the larval gonad . To test this , we performed live-imaging of larval gonads of transgenic lines that report the active form of RhoA/Rac and cdc42 [108] . In these biosensors , GFP is fused to the Rho family GTPase Binding Domain ( RBD ) of downstream effector proteins . Expression of GFP suggests the presence of active forms of these proteins . For RhoA/Rac activity , the downstream effector protein , Protein Kinase N ( Pkn ) was fused to GFP and placed under squash ( sqh ) promoter [108] . For cdc42 activity , the cdc42 binding domain of Wiskott-Aldrich Syndrome protein ( WASp ) was fused with GFP and this was placed under the control of sqh promoter [108] . In order to observe the germ cells in these biosensors , we used kusabira-orange ( KO ) fused germline marker , Vasa-KO [109 , 110] . Similar to the expression pattern of the GFP tagged RhoA , Rac1 and cdc42 we observed the presence of active form of RhoA , Rac1 and cdc42 in the ICs of the larval gonad ( Fig 3E–3F2 ) . We also asked if these proteins are present in their active forms in the escort cells of the adult germaria . We found that RhoA/Rac activity was attenuated and cdc42 was present in its activated state in the escort cells , albeit at lower levels compared to the larval gonad ( S7A–S7C Fig ) . Together , these results demonstrate that the Wnt non-canonical components are not only present , but also active in the ICs . To determine the activity of the canonical pathway in the larval gonad , we used Fz3RFP [57] . We found that while Fz3RFP showed no expression or was expressed at background levels in the ICs of the larval gonad , it was expressed at higher levels in the escort cells of pupal and adult stages [46 , 111 , 112] ( Fig 3G–3I ) ( S7D–S7D1 Fig ) . Additionally , all the escort cells in the adult germarium expressed the canonical reporter ( n = 10 , 99 . 5% Tj positive cells expressed Fz3RFP ) . Thus , we concluded that Wnt canonical and Wnt non-canonical pathway components , and their activity , are temporally regulated in escort cells and their precursors during oogenesis . To determine if the different expression patterns of the downstream Wnt non-canonical pathway components mirrored their role in the two developmental stages , we used the UAS-GAL4 system to deplete genes using RNAi in the escort cells at specific developmental time points . The UAS-GAL4 system is temperature dependent and therefore , different temperatures can be used to attenuate genes at various developmental time points . Maximal RNAi activity is attained at 29 oC whereas minimal RNAi activity is attained at 18 oC [113] . c587 is expressed in the larval gonad , including the ICs , but is only expressed in the escort cells of the adult inner germarium ( S8A–S8C1 Fig ) . To elucidate if the downstream Wnt non-canonical pathway components are required in the adults to regulate CB differentiation , the flies were kept at 18 oC until they eclosed . Once eclosed , the flies were shifted to 29 oC and kept at this temperature for 7 days [114] ( Fig 4A ) . The germaria were then stained for 1B1 and Vasa . This strategy has been previously used to show that the dWnt4 canonical pathway is required in the adults to regulate CB differentiation [45] . We depleted dWnt4 , RhoA , Rac1 , cdc42 and also expressed cdc42DN in the escort cells to assay for differentiation defects . We found that compared to control , 18 oC-29 oC temperature shift dWnt4 mutants exhibited a differentiation defect ( Fig 4B and 4C and Fig 4G ) . In contrast , compared to control , 18 oC-29 oC temperature shift RhoA , Rac1 , cdc42 and cdc42DN mutants did not exhibit a significant accumulation of undifferentiated cells , suggesting that the downstream Wnt non-canonical pathway components do not play a critical role in the adult to regulate CB differentiation ( 3 ± 1 for c587-GAL4; 4 ± 1 for c587-GAL4>dWnt4 RNAi , P-value = 0 . 002958; 3 ± 2 for c587-GAL4>RhoA RNAi , P-value = 0 . 431434; 3 ± 1 for c587-GAL4>Rac1 RNAi , P-value = 0 . 91346 , 3 ± 1 for c587-GAL4>cdc42 RNAi , P-value = 0 . 54991 and 3 ± 1 for c587-GAL4>cdc42DN , P-value = 0 . 242545 {for all n = 50} ) ( Fig 4B and Fig 4D–4G ) ( S9A and S9B Fig , S9E–S9G Fig ) . To elucidate if the downstream Wnt non-canonical pathway components are required in the larval gonad to regulate CB differentiation , the flies were kept at 29 oC until they eclosed . Once eclosed , the flies were shifted to 18 oC and kept at this temperature for 7 days [114] ( Fig 4H ) . Compared to control , 29 oC-18 oC temperature shift dWnt4 , RhoA , Rac1 , cdc42 and cdc42DN mutants resulted in an accumulation of undifferentiated cells ( 2 . 5 ± 1 for c587-GAL4; 5 . 6 ± 2 for c587-GAL4>dWnt4 RNAi , P-value = 1 . 04811E-17; 5 ± 1 . 5 for c587-GAL4>RhoA RNAi , P-value = 1 . 43965E-15; 5 ± 1 . 5 for c587-GAL4>Rac1 RNAi , P-value = 1 . 53586E-18; 5 ± 2 . 5 for c587-GAL4>cdc42 RNAi , P-value = 2 . 34120E-18 and 5 ± 1 for c587-GAL4>cdc42DN , P-value = 1 . 49752E-17{for all n = 50} ) ( Fig 4I–4N ) ( S9C–S9F and S9H Fig ) . Taken together , these results demonstrate that the Wnt non-canonical pathway components act primarily prior to the adult stage to regulate CB differentiation . As RhoA , Rac1 and cdc42 are highly expressed and active in the larval gonad and were not required in the adult germaria for CB differentiation , we asked if these proteins are required earlier in development . To test this we used an IC driver traffic jam-GAL4 ( tj-GAL4 ) . Tj is expressed only in ICs in the late larval gonad , therefore tj-GAL4 is a more restricted driver for ICs [11 , 29] . We depleted dsh , DAAM1 , RhoA , Rac1 , cdc42 and expressed RhoADN and cdc42DN versions in the ICs , in the late larval stages using tj-GAL4 . Because we posited that RhoA , Rac1 and cdc42 act downstream of dWnt4 , we also monitored the late larval gonads of dWnt4 mutants for defects , by staining for Vasa , Tj and 1B1 . We found that compared to dWnt4 heterozygotes , dWnt4 mutants and dsh , DAAM1 , RhoA , RhoADN , Rac1 , cdc42 and cdc42DN mutants showed loss of intermingling of ICs with PGCs ( n = 25 ) ( Fig 5A–5H ) ( S10A–S10C Fig ) . Although loss of Rac1 in ICs led to loss of intermingling , the phenotype was weaker than dWnt4 mutants , RhoA and cdc42 depleted ICs . Additionally , depletion of Rac1 in the ICs resulted in a decrease in the size of the larval gonad . It has been shown Rac1 , Ras-related C3 botulinum toxin substrate 2 ( Rac2 ) and Mig-2-like ( Mtl ) act redundantly to regulate PCP [115] . For this reason , we also depleted Mtl and Rac2 specifically in the ICs and found that Mtl depleted ICs resulted in a strong loss of ICs intermingling but Rac2 depleted ICs did not exhibit any intermingling defect ( Fig 5I and 5J ) . Although we find that depletion of Rac2 does not lead to a differentiation defect , we do not know if this is due to a feeble RNAi mediated depletion , or in fact Rac2 does not play a role . These results suggest that dsh , DAAM1 , RhoA , Rac1 , Mtl and cdc42 regulate intermingling of ICs in the larval gonad . In order to further validate our results obtained by RNAi and expression of dominant negative forms of RhoA and cdc42 , we also analyzed larval gonads of dsh1 mutants , cdc425 , a hypomorphic allele that has been previously shown to exhibit mild PCP defects , cdc42 ( cdc422/cdc425 ) hypomorphic mutant and heterozygous larval gonads of flies having one genetically reduced copy of all three Rac genes [115] . We found that these mutants and the heterozygous larval gonad for flies that remove all three Rac genes also exhibited loss of ICs intermingling ( 69% dsh1 mutants , n = 20; 55% cdc425 mutants , n = 20; 80% cdc42 ( cdc422/cdc425 ) mutants , n = 20 and 60% heterozygous Rac1 , Rac2 and Mtl larval gonads , n = 10 ) ( S10D–S10H Fig ) . Together , these results suggest that RhoA , Rac genes ( Rac1 and Mtl ) and cdc42 regulate intermingling of ICs . As RhoA , Rac1 and cdc42 are known to play a role in cell division , we wondered if there were additional defects apart from loss of intermingling in these mutants . We found that compared to dWnt4 heterozygotes and tj-GAL4 , dWnt4 mutants and RhoA , Rac1 , Mtl and cdc42 mutant larval gonads had significantly fewer ICs ( 151 ± 18 for dWnt4/CyO; 84 ± 8 for dWnt4 mutants , P-value = 6 . 74402E-05; 213 ± 36 for tj-GAL4; 36 ± 13 for tj-GAL4>RhoA RNAi , P-value = 5 . 16854E-06; 67 ± 14 for tj-GAL4>Rac1 RNAi , P-value = 9 . 83637E-05; 111 ± 10 for tj-GAL4>Mtl RNAi , P-value = 0 . 000231 and 61 ± 17 for tj-GAL4>cdc42 RNAi , P-value = 2 . 07541E-05 {for all n = 5} ) . It is possible that the decrease in the number of ICs is either due to decrease in their division rate or due to death of ICs . To test if loss of dWnt4 , RhoA , Rac1 , Mtl and cdc42 affects the division rate of the ICs , we stained these mutants with a mitotic marker , Phospho-Histone 3 ( PH3 ) , Tj and 1B1 . We observed that RhoA , Rac1 , Mtl and cdc42 depleted IC mutants show a decrease in the number of ICs expressing PH3 , suggesting that RhoA , Rac1 , Mtl and cdc42 also regulate division of the ICs . However , dWnt4 mutants did not exhibit this defect compared to heterozygous , but displayed a defect compared to wild type ( WT ) control ( 12 . 5 ± 2 for WT control; 4 . 5 ± 2 for dWnt4/CyO , P-value = 1 . 72943E-08; 3 ± 2 for dWnt4 mutants , P-value ( WT ) = 7 . 42348E-09 , P-value ( dWnt4/CyO ) = 0 . 104558; 9 ± 3 for tj-GAL4 , n = 10; 4 ± 2 for tj-GAL4>RhoA RNAi , n = 10 , P-value = 0 . 00010; 3 ± 2 for tj-GAL4>Rac1 RNAi , n = 10 , P-value = 2 . 56584E-05; 7 ± 2 for tj-GAL4>Mtl RNAi , n = 10 , P-value = 0 . 035212 , and 5 ± 3 for tj-GAL4>cdc42 RNAi , n = 10 , P-value = 0 . 003773 ) ( S11A–S11H Fig ) . To determine if the ICs in these mutants show increased cell death , we stained them for a death marker , cleaved Caspase3 , along with Tj and 1B1 . We observed that only ICs of RhoA mutants showed Caspase3 staining . Compared to 0% of Tj positive cells in control , 3% of Tj positive cells in RhoA depleted IC mutants exhibited Caspase3 staining indicating that in addition to its role in cell division , RhoA also regulates IC survival ( n = 5 ) ( S11I–S11O Fig ) . Together , these results show that dWnt4 and the downstream components of the Wnt non-canonical pathway , RhoA , Rac1 , Mtl and cdc42 play a critical role to regulate intermingling and cell division of ICs in the larval gonad . Additionally , as RhoA depletion exhibited a stronger defect than dWnt4 mutants it is likely that it has roles independent of dWnt4 signaling . dWnt4 mutants and RhoA , Rac1 and cdc42 depleted escort cell germaria exhibit reduced number of escort cells and loss of CB encapsulation [46] . Additionally , depletion of RhoA , Rac1 and cdc42 in the ICs resulted in reduced number of ICs . We asked if reduction in the escort cell number and loss of encapsulation of CB in the adults is a consequence of reduced number of ICs in the larval gonad . To answer this question , we counted the number of Tj positive escort cells in the adult flies , of 29 oC-18 oC temperature shift escort cell depleted RhoA , Rac1 and cdc42 mutants . We found that these mutants exhibited significantly lower number of escort cells ( 15 . 4 ± 2 . 6 for c587-GAL4; 10 . 4 ± 2 . 6 for c587-GAL4>RhoA RNAi , P-value = 0 . 00048; 10 . 6 ± 2 for c587-GAL4>Rac1 RNAi , P-value = 0 . 00068 and 6 . 9 ± 1 . 6 for c587-GAL4>cdc42 RNAi , P-value = 8 . 1045E-08 {for all n = 10} ) . These results suggest that RhoA , Rac1 and cdc42 act in the ICs to regulate the escort cell number and therefore , proper GSC differentiation in the adults . To determine if PCP core components regulate intermingling in the larval gonad we depleted vang/stbm , ds , fz and ft in ICs and found that depletion of these genes did not exhibit any intermingling defects ( S12A–S12E Fig ) . To ascertain that the canonical pathway is not required in the larval gonad , we depleted the canonical specific co-receptor , arr in ICs and found that these mutants also did not exhibit any intermingling defect ( S12A and S12F Fig ) . We used arr as β-catenin has roles outside of Wnt signaling and loss of arr in the adult escort cells results in CB differentiation defects [45 , 116] . These results suggest that the Wnt canonical pathway and vang/stbm , ds , fz and ft , the core components and Ds/Ft systems that define PCP , do not play a critical role in the ICs . Similar to dWnt4 mutants , RhoA , Rac1 and cdc42 depletion in the ICs showed a defect in intermingling of ICs . To determine if RhoA , Rac1 and cdc42 act downstream of dWnt4 , we analyzed the expression and activity of RhoA , Rac1 and cdc42 with the help of the GFP tagged lines , in larval gonads of both control and dWnt4 mutants . We found that compared to the control , expression of RhoA , Rac1 and cdc42 was not altered in the ICs of dWnt4 mutants , suggesting that dWnt4 does not regulate the protein levels of RhoA , Rac1 and cdc42 ( P-value = 0 . 64182 for RhoAGFP , P-value = 0 . 79908 for Rac1GFP , and P-value = 0 . 63675 for cdc42GFP in dWnt4 mutants {for all n = 3} ) ( S13A–S13G Fig ) . The activity level of RhoA/Rac and cdc42 was determined using biosensors in dWnt4 mutants . To ensure that there is no background , we also imaged larval gonads , without the biosensors ( S13H–S13I1 Fig ) . We found that RhoA/Rac and cdc42 activity was significantly downregulated in ICs ( P-value = 0 . 02067 , n = 25 for RhoA/Rac activity and P-value = 0 . 01369 , n = 25 for cdc42 activity ) ( Fig 6A–6F ) . We also observed cdc42 activity in the terminal filaments and found that it was not altered in dWnt4 mutants ( P-value = 0 . 35901 , n = 3 ) ( S13J Fig ) . These results suggest that dWnt4 regulates the activity of RhoA , Rac and cdc42 in the ICs to regulate intermingling of ICs with PGCs .
Here , we find that dWnt4 regulates CB differentiation through the downstream non-canonical pathway components . We show that RhoA , Rac1 and cdc42 are expressed at high levels in the ICs of the larval gonad and are active while the expression of Fz3RFP , a β-catenin-dependent transcriptional reporter , is not detectable . Conversely , we find low levels of RhoA , Rac1 and cdc42 in the adults but the β-catenin-dependent canonical reporter is expressed at high levels . Consistent with this , we find a role for RhoA , Rac1 and cdc42 but not the β-catenin-dependent canonical pathway in the larval gonad for promoting intermingling of ICs ( Fig 7 ) . Our results , in conjunction with Mottier-Pavie et al . and Wang et al . , that show the canonical pathway is required in the adult , point towards a switch from utilizing Wnt non-canonical components to utilizing a canonical pathway to regulate the formation of the differentiation niche [44 , 45] . Our findings that the depletion of vang/stbm , ds , fz and ft in either ICs or escort cells does not lead to differentiation defects , suggests that PCP does not play a critical role in regulating CB differentiation . These results are consistent with previous literature , where Cohen et al . showed that mutations in , fz , fmi , dgo , vang/stbm and pk , genes required for PCP of a cell , do not lead to significant disruption of ovarian morphology [82] . We show that the other PCP system components , ds and ft , are also not required to regulate differentiation . Together , Cohen et al . and our results suggest that it is not the PCP system , but the downstream Wnt non-canonical components that play a critical role in the differentiation niche to regulate differentiation . We propose that dWnt4 uses the downstream Wnt/Fz non-canonical pathway to regulate formation of the differentiation niche . We find that RhoA , Rac1 and cdc42 regulate intermingling of ICs in the larval gonad . In addition , we find that loss of RhoA leads to death of ICs and loss of Rac1 leads to decrease in the larval gonad size . dWnt4 mutants do not show either of these outcomes . Together , these results suggest that RhoA and Rac1 have roles in the larval gonad independent of Wnt signaling . It has been extensively shown that Rac1 , Rac2 and Mtl have redundant roles in axon growth , guidance and PCP in the eye and the wing [93 , 95 , 115] . Indeed , we found that Rac1 depleted ICs resulted in loss of intermingling but this phenotype was not as strong as dWnt4 , RhoA and cdc42 depleted ICs . Rac2 depleted ICs resulted in no intermingling defect . However , depletion of Mtl in the ICs resulted in a strong loss of intermingling phenotype . This suggests that small GTPases play a critical role in ICs function , both independently and in coordination with Wnt signaling . In the adult , escort cells extend dynamic cytoplasmic protrusions to encapsulate the GSCs , the differentiating daughter and differentiating cysts , a function necessary for proper differentiation [28 , 39–41 , 102 , 117] . It was recently shown that in adult female germaria , stat regulates the formation of these protrusions through cdc42 . Loss of stat or woc , a component of the stat pathway , leads to formation of smaller protrusions instead of longer stable protrusions . Loss of woc also leads to an accumulation of undifferentiated cells . Overexpression of cdc42 in woc mutants rescued the small protrusion defect and partially rescued the phenotype . However , Banisch et al report that expression of dominant active or dominant negative cdc42 in the escort cells only mildly affects protrusions and does not lead to an accumulation of undifferentiated cells . Our temperature shift experiments also suggest that expression of dominant negative cdc42 exclusively in the adults does not result in an accumulation of undifferentiated cells , while its expression prior to eclosion does ( S9C and S9D and S9E and S9F Fig ) . Moreover , Banisch et al . also show that active cdc42 is expressed in the escort cells and active RhoA could not be detected in most of the escort cells using Dia-RBD:GFP but could be detected mainly in the escort cell body and weakly in the escort cell protrusion using Capu-RBD:GFP [117] . We think that cdc42 may have a role in the adults by controlling protrusive activity of escort cells , but it does not affect CB differentiation in a biologically meaningful way . This suggests that cdc42 and Rho activity in the adult is attenuated and their activity is independent of dWnt4 . Cell movement requires dynamic actin-myosin polymerization at the leading edge , driven by Rac1 and cdc42 , and actin-myosin contraction at the lagging end , driven by RhoA [118 , 119] . In our study , we find that both RhoA/Rac and cdc42 are present and active in the ICs . Loss of RhoA , Rac1 and cdc42 leads to loss of intermingling of ICs , suggesting that these proteins drive ICs cell movement and intermingling in the larval gonad . In contrast , only cdc42 , but not RhoA , is active in the adult escort cells , consistent with the fact that these cells are stationary [41 , 117] . Thus the switch in the module used by Wnt signaling to regulate the activity of RhoA/Rac and cdc42 parallels the developmental demands of these cells to first move between germ cells in the larval gonad , then to stop and create long , stable protrusions that promote differentiation in the pupal and adult gonad . Cancer metastasis requires cell migration . The invasion of metastatic cancer cells requires cells to lose epithelial properties and gain mesenchymal properties . During this process , known as Epithelial-mesenchymal transition ( EMT ) , the epithelial cells , which are otherwise polarized , non-motile and have strong cell-cell interaction , subsequently lose polarity , cell-cell adhesion and become motile [120 , 121] . RhoA , Rac1 and cdc42 are conserved from lower eukaryotes to mammals and are key players that affect cell division , cytoskeletal rearrangement , cell polarity , and cell motility [76 , 77 , 79 , 80 , 91 , 106 , 118] . It has been previously discovered that during cancer , these proteins are upregulated and therefore help initiate metastasis [122] . We have discovered that RhoA , Rac1 and cdc42 are turned on in the ICs of the larval gonad and switched off in the adult escort cells . This is fascinating because the ICs and the escort cells are essentially the same cells at different developmental time points . A better understanding of how this switch is mediated may give us an insight into cancer and aid in developing mechanisms to block metastasis .
The following fly stocks were used in the study: c587-Gal4 , trafficJam ( tj ) -Gal4 , dWnt4C1/CyO ( 6651 ) , bamGFP , arrow RNAi ( 31313 ) , Rho1 RNAi ( Bloomington 32383 ) , UASRhoDN ( Bloomington 7327 ) , Rac1 RNAi ( Bloomington 28985 and 34910 ) , Mtl RNAi ( Bloomington 51932 ) , Rac2 RNAi ( v28926 ) , cdc42 RNAi ( Bloomington 35756 and 37477 ) , UAScdc42DN ( Bloomington 6288 ) , dDAAM1 RNAi ( Bloomington 39058 , V24885 ) , dsh RNAi ( v101525 ) , frizzled RNAi ( Bloomington 34321 ) , van gogh RNAi ( Bloomington v7376 ) , daschous RNAi ( Bloomington 28008 ) , fat RNAi ( Bloomington 29566 ) , y1 w*; Rho172F/CyO ( Bloomington 7326 ) , y1 w67c23; P{EPgy2}Rac1EY05848/TM6B , Tb1 ( Bloomington 15461 ) , y[1] w[*]; Rac1[J10] Rac2[Delta] P{w[+mW . hs] = FRT ( w[hs] ) }2A Mtl[Delta]/TM6B , Tb[1] ( Bloomington 6679 ) , y1 w* Cdc423/FM6 ( Bloomington 7337 ) , y1 w* Cdc422 P{neoFRT}19A ( Bloomington 9105 ) , y1 w* Cdc425 P{neoFRT}19A ( Bloomington 52237 ) , y1 Mi{MIC}DAAMMI04569 w1118/FM7h ( Bloomington 38567 ) , w1 dsh1 ( Bloomington 5298 ) , RhoAGFP ( V318439 ) , Rac1GFP ( Bloomington 52285 ) , cdc42GFP ( V218151 ) , w*;P{sqh-Pkn . RBD . G58A-eGFP}312a P{sqh-Pkn . RBD . G58A-eGFP}312b ( Bloomington 52298 ) , w*;P{sqh-WASp . RBD-GFP}378a P{sqh-WASp . RBD-GFP}378b ( Bloomington 56746 ) , UASmCD8GFP ( Bloomington 32184 ) , faxGFP , Sco/CyO;MKRS/TM6 ( Lehmann Lab ) , Sco/CyO;Nos-Gal4::VP16 , Vasa-KO/TM6 ( Lehmann Lab ) ; Sco/Cyo;Fz3RFP ( Bach Lab ) . 3–4 day old fly ovaries were dissected in PBS , fixed for 30 min in PBS plus 5% formaldehyde , incubated for 1 h in PBST ( 0 . 2% Tween 20 ( Sigma ) in PBS ) supplemented with 1% Triton X-100 ( Sigma ) , followed by incubation for 2 h in BBT ( PBST supplemented with 1% ( w/v ) bovine serum albumin ( BSA; Sigma ) ) . Primary antibodies were added in BBT and incubation was carried out overnight at 4 oC . The following day , ovaries were washed four times for 10 min , 20 min , 30 min in BBT and for 30 min in BBT supplemented with 2% ( w/v ) donkey serum ( Sigma ) . Secondary antibodies were added in BBT supplemented with 4% ( w/v ) donkey serum ( Sigma ) and incubated for 2 h followed by five washes , 10 min each in PBST . VECTASHIELD ( Vector Laboratories ) with DAPI was added prior to mounting . Fixation and staining of larval gonads was carried out as previously described [123] . In order to express a specific gene only in the adults , the flies were kept at 18 oC until they eclosed . Once eclosed , the young flies were shifted to 29 oC and kept at this temperature for 7 days . Staining was followed to determine any differentiation defects . In order to express a specific transgene only in the larval stage , the flies were kept at 29 oC until they eclosed . Once eclosed , the young flies were shifted to 18 oC and kept at this temperature for 7 days . These flies were then dissected and stained for observation . Unless specified , all experiments that utilized the UAS-GAL4 system were performed constitutively at 29 oC . Immunostaining of the ovaries and larval gonad was carried out with the following Primary antibodies: Mo 1B1 ( 1:20 , DSHB ) , Rb Vasa ( 1:5000 , Rangan Lab ) , Ch Vasa ( 1:500 , Rangan Lab ) , GP Traffic Jam ( 1:5000 , Godt Lab ) , Rb GFP ( 1:2000 , ab6556 ) , Rb pMAD ( 1:200 , abcam AB52903 ) , Mo BamC ( 1:200 , DSHB ) , Rb Bruno ( 1:500 , Lehmann Lab ) , Rb PH3 ( 1:200 , Cell Signaling 97015 ) , and Rb Caspase 3 ( 1:300 , Cell Signaling 96615 ) . Alexa 488 ( Molecular Probes ) , Cy3 and Cy5 ( Jackson Labs ) conjugated secondary antibodies were used at a concentration of 1:500 . The tissues were visualized under 10X , 20X , 40X and 63x objective lenses . The images were acquired using a Zeiss LSM-710 confocal microscope under 20x , 40X and 63x objective . The tissues were dissected in Schneider’s media and mounted . These were then visualized under 20X objective lens . The images were acquired using a Zeiss LSM-710 confocal microscope under 20x objective . P-values were determined by two-tailed equal variance t test in mutants vs . wild type strains . Z-score was determined by a two-tailed test for data represented in percentage . Fly food was created using the procedures from the Ruth Lehmann lab at NYU ( summer/winter mix ) , and used to fill narrow vials to approximately 12mL .
|
Germ line association with the somatic cells is critical for various aspects of germ cell biology , including migration , self-renewal and differentiation . In Drosophila females , soma–germ line association begins during embryogenesis and continues until the mature egg is formed . In the adult , the somatic escort cells promote differentiation of the germline stem cell daughter using Wnt signaling . dWnt4 , a Wnt ligand , acts in an autocrine manner in these escort cells , using the canonical pathway to regulate survival , division and encapsulation of the stem cell daughter , a function critical for differentiation . Here , we show at an earlier stage , in the larvae , the same ligand uses components of Wnt non-canonical pathway , RhoA , Rac1 and cdc42 , to regulate proper mingling of escort cell precursors between the germ cells . Thus , dWnt4 uses different modules of signaling at different points in development to promote cell movement and control cytoplasmic protrusions . As Wnts have been associated with cancers , understanding how Wnts modulate cell movement by switching on and off different modules may lead to insights into the etiology and progression of cancers .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"reproductive",
"system",
"rna",
"interference",
"gonads",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"cell",
"differentiation",
"developmental",
"biology",
"stem",
"cells",
"epigenetics",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"genetic",
"interference",
"animal",
"cells",
"stem",
"cell",
"niche",
"gene",
"expression",
"biochemistry",
"rna",
"signal",
"transduction",
"cell",
"staining",
"anatomy",
"nucleic",
"acids",
"cell",
"biology",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"wnt",
"signaling",
"cascade",
"cell",
"signaling",
"signaling",
"cascades",
"genital",
"anatomy"
] |
2018
|
A switch in the mode of Wnt signaling orchestrates the formation of germline stem cell differentiation niche in Drosophila
|
The apical complex is the definitive cell structure of phylum Apicomplexa , and is the focus of the events of host cell penetration and the establishment of intracellular parasitism . Despite the importance of this structure , its molecular composition is relatively poorly known and few studies have experimentally tested its functions . We have characterized a novel Toxoplasma gondii protein , RNG2 , that is located at the apical polar ring—the common structural element of apical complexes . During cell division , RNG2 is first recruited to centrosomes immediately after their duplication , confirming that assembly of the new apical complex commences as one of the earliest events of cell replication . RNG2 subsequently forms a ring , with the carboxy- and amino-termini anchored to the apical polar ring and mobile conoid , respectively , linking these two structures . Super-resolution microscopy resolves these two termini , and reveals that RNG2 orientation flips during invasion when the conoid is extruded . Inducible knockdown of RNG2 strongly inhibits host cell invasion . Consistent with this , secretion of micronemes is prevented in the absence of RNG2 . This block , however , can be fully or partially overcome by exogenous stimulation of calcium or cGMP signaling pathways , respectively , implicating the apical complex directly in these signaling events . RNG2 demonstrates for the first time a role for the apical complex in controlling secretion of invasion factors in this important group of parasites .
Apicomplexans are obligate parasites of metazoans that non-destructively enter their host cells . Here they feed and replicate before destructively escaping in search of further cells to invade . Apicomplexa comprises over 6000 species that parasitize virtually every animal group [1] . The malaria-causing parasites , Plasmodium spp . , are best known for their pattern of invasion and release from human red blood cells , causing cyclic fevers and the symptoms of malaria that annually result in 0 . 6 to 1 million deaths per year and morbidity in up to 220 million people [2] . Toxoplasma gondii can infect most nucleated mammalian cell types and infects approximately one third of the human population . Human infections are typically relatively asymptomatic , however T . gondii causes acute and even fatal disease in immuno-compromised individuals ( encephalitis and ocular disease ) , severe or lethal developmental defects in unborn fetuses , and significant agricultural losses through miscarriage in livestock [3] . Early-diverging apicomplexans ( gregarines ) are limited to invertebrate hosts and their invasion is incomplete , with feeding often achieved through the apical tip of the parasite being intimately buried within the host cell [4] , [5] . The defining feature of Apicomplexa is a complex assemblage of structural and secretory elements at the apical point of the cell , forming the namesake of the group—the apical complex . The apical complex is instrumental in the host cell invasion processes [6] , [7] . It provides both a semi-rigid framework to these apically pointed cells , and a focal point for secretory organelles that release various invasion factors that mediate interaction with , and invasion of , the host cell . The apical complex is organized around an apical polar ring that serves as a microtubule organizing center that nucleates an array of subpellicular microtubules that descend toward the posterior of the cell ( Figure 1A ) [8]–[10] . These microtubules subtend flattened membrane sacs , or alveoli , that line most of the plasma membrane [11] . A fibrous proteinaceous membrane skeleton supports the alveolar sacs against the microtubules [12] . The alveoli and proteinaceous skeleton form a structure called the inner membrane complex ( IMC ) , which , together with the subpellicular microtubules , provides the shape and stability of the cell . The apical polar ring marks the apical extremity of the IMC . A mobile conoid , consisting of tightly bent tubulin filaments fused to form a tapered hollow barrel , sits within the apical polar ring [10] , [11] , [13] . The conoid can either be recessed in the cell , so that its tip is flush with the apical polar ring , or , during invasion , be extruded from the apical polar ring to form an extended point to the cell ( Figure 1A ) . At the tip of the conoid are two preconoidal rings , and a pair of short microtubules sit eccentrically within the conoid . These preconoidal rings and interconoidal microtubules move together with the conoid during extrusion [8] . The structural elements of the apical complex provide orientation to the cell , and are the focal point for arrays of secretory organelles—micronemes and rhoptries—that cluster towards the base of the conoid in readiness for a staged sequence of release ( Figure 1A ) [14] . Microneme contents are secreted first , prior to invasion , and coat the parasite with proteins that facilitate host cell adhesion , gliding motility , and contribute to formation of an annular moving junction with the host plasma membrane through which the parasite enters the host . During invasion rhoptries secrete further elements of the moving junction , as well as proteins that establish the properties of the parasitophorous vacuole within which the parasite typically resides . The elements of the apical complex are highly conserved throughout Apicomplexa , although secondary reduction is evident . For example , the conoid is only intermittently present within various members and life stages of haemosporins ( including Plasmodium spp . ) and likely completely lost from piroplasms such as Babesia and Theileria [15] , [16] . The presence of the apical polar ring , however , is seemingly universal . Apicomplexans also have a distinctive model of cell division , whereby daughter cells form within a mother cell . The pellicle , consisting of the alveolar sacs and protein skeleton ( and plasma membrane in mature cells ) , with the associated subpellicular microtubules , is amongst the first structures formed in the new daughter cells , and this lays down the scaffold for nuclei and organelles to correctly partition into these daughters [17]–[22] . Markers for the conoid in T . gondii also appear early in daughter cell formation [23] , suggesting the apical complex is also formed early in this process . The apical complex , therefore , likely plays pivotal roles in both cell division and host cell invasion . Despite extensive characterization of the apical complex through ultrastructural studies , there is relatively limited knowledge of the molecular composition of its structural elements , and even less experimental illumination of its function . T . gondii provides the best studied system to date , with several proteins that associate with the apical complex structures identified either through proteomics or reporter protein tagging . The conoid itself is composed of tubulin , and is known to be decorated with several proteins ( TgCentrin3; calcium-binding domain proteins CAM1 and CAM2; dynein light chain , TgDLC ) , although the functions of these proteins remain untested [13] , [23] . The preconoidal rings are associated with TgCentrin2 and SAS6L , two proteins typically implicated with centriolar function [23]–[25] . SAS6L knockout cells showed a subtle negative growth phenotype , however the basis of this phenotype has not been determined [24] . The novel protein TgICMAP1 decorates the intraconoidal microtubules but its function is unknown [26] . Even the curious behavior of conoid extrusion , while hypothesized to provide some mechanical role in invasion , has evaded clear insight into its function [27] . A small number of proteins associated with the apical cap of the IMC are closely associated with the apical complex , but these appear to serve more general pellicle functions in gliding motility ( GAP70 ) , coordination of assembly and spatial organization of the pellicle ( IMC15 , ISP1 , MORN1 ) or are of unknown function ( PhIL1 , TgDLC ) [18] . The composition of the apical polar ring is the most poorly characterized of these structures , despite its central role and universality in the phylum . A single protein , RNG1 , has been localized to this structure [28] . RNG1 associates with the apical polar ring only as daughter cells reach maturity , so presumably is not responsible for the early formation of this ring . Attempts to generate a knockout were unsuccessful , suggesting an essential but undetermined function . As part of a broader study of conserved pellicle proteins of Infrakingdom Alveolata ( Apicomplexa , Dinoflagellata , Ciliophora ) , we previously identified a novel T . gondii protein that localized as a ring in the region of the apical complex [29] . To investigate the function of the apical complex we have investigated the localization and behavior of this ring protein ( we now call RNG2 ) during cell replication and invasion , and examined its role by inducible knockdown of RNG2 expression . These results confirm that new components of the apical complex are first assembled at the centrosomes at the earliest stages of cell replication , and that the apical complex acts as a gatekeeper that regulates secretion during parasite invasion .
To investigate the fine localization and interactions of RNG2 with other apical structures we used 3D-structured illumination microscopy ( SIM ) , which provides an 8-fold increase in volume resolution over conventional light microscopy [33] , [34] . Given the large size of RNG2 and its potential to fill a volume larger than SIM resolution , we separately tagged the N-terminus and C-terminus , with epitope tags HA and cMyc , respectively ( HA-RNG2-cMyc , see knockdown construct below ) . Immuno-detection of these tags in intracellular parasites shows that each forms a continuous apical ring in mature cells measuring 380 +/− 20 nm ( standard deviation ( SD ) ) in diameter ( Figure 1C , D ) . These two rings are consistently displaced , with the C-terminus ( red ) adjacent to the apical extremity of the GAP45-labeled IMC ( blue ) , and the N-terminus ( green ) occurring below the IMC in the region of the retracted conoid ( Figure 1Di and ii ) . This implies that RNG2 forms a tube or collar , with a consistent orientation of the protein termini . The conoid can also be extruded relative to the apical polar ring , a state that typically occurs during invasion . Conoid extrusion can be artificially achieved by exposure to the calcium ionophore A23187 [35] , [36] . When extracellular parasites were treated with A23187 to effect conoid extrusion , the orientation of the RNG2 N- and C-terminus was completely reversed , with the N-terminus now extended anterior to the C-terminus ( Figure 1Diii and iv ) . To test that the RNG2 protein termini could be reproducibly detected , we also created a RNG2 fusion with fluorescent proteins ( mCherry-cMyc-RNG2-GFP ) . Observation of these live cell markers resulted in the same pattern of a polarized orientation of the RNG2 collar that inverted upon conoid extrusion ( Figure S1 ) . To better understand the RNG2 flip during conoid extrusion , the two termini were co-labeled with markers for the apical polar ring , RNG1 [28] , and the conoid , CAM1 [23] . RNG1 was transiently expressed as a RNG1-GFP fusion in the HA-RNG2-cMyc cells . When the conoid was in the retracted position , the C-terminus of RNG2 ( red ) collocates with RNG1 ( blue ) , whereas the RNG2 N-terminus ( green ) is posterior ( Figure 2A and B ) . This indicates that the C-terminus of RNG2 is in close association with the apical polar ring . When the conoid is extruded by A23187 , the N-terminus of RNG2 ( green ) is now extended anterior of RNG1 ( blue ) , suggesting that it has passed through the apical polar ring ( Figure 2C ) . CAM1 is a known conoid marker , although its precise location on the conoid remains undetermined [18] , [23] . We transiently expressed CAM1-GFP in the HA-RNG2-cMyc cells to observe the position and behavior of the RNG2 termini with respect to this conoid marker . Using SIM the CAM1-GFP ( blue ) resolved as a narrow ring of smaller diameter than the apical polar ring , but of similar apparent depth ( Figure 2D-F ) . The N-terminus of RNG2 was a consistent distance posterior to CAM1 , maintaining this position when the conoid was extruded and RNG2 flipped . These data suggest that the C-terminus of RNG2 is attached to the apical polar ring , and the N-terminus is attached to the conoid at a position posterior to CAM1 . During conoid extrusion and retraction the RNG2 N-terminus moves with the mobile conoid , while the C-terminus is apparently anchored to the apical polar ring . CAM1 itself is confined to a narrow band on the conoid , likely towards the conoid tip as when the conoid was retracted the CAM1 ring was approximately level with RNG2 C-terminus ( Figure 2E ) . Our first report of RNG2 indicated that this protein associates with the apical complex early during cell division , whereas RNG1 is a late marker of daughter cell formation [28] , [29] . To correlate RNG2 behavior with other known early events in daughter cell formation , we have used markers of the centrosome ( centrin 1 antibodies ) and centrocone ( MORN1-cMyc expression ) with the 3′ endogenous tagged RNG2-HA cells where RNG2 maintains its native promoter [29] . Duplication of the centrosome is one of the earliest events of daughter cell formation [37] , and we see RNG2 association with daughter cells only after this event ( Figure 3A–C ) . RNG2 first appears as two diffuse dots apparently in contact with each centrosome ( Figure 3C ) . Subsequently , RNG2 resolves into a small ring that is separate and anterior to the centrosome ( Figure 3D ) . This indicates a separation of these nascent structures , but we often see a small residual amount of RNG2 that persists with the centrosome ( Figure 3D , arrowheads ) , suggesting a direct interaction with the centrosome occurs . The centrocone is an elaboration of the nuclear envelope that serves as the connection point between centromeres of chromosomes in the nucleus to the extranuclear centrosome [38] . Following duplication of the centrosome and development of the mitotic spindle , the centrocone elongates before separating into two resolved centrocones . MORN1 labels the centrocone , as well as the basal complex of the IMC ( seen as a ring at the base of the mother cell as well as forming daughter cells: Figure 4 ) [23] , [39]–[41] . The nascent dots of RNG2 are seen before the centrocone separates into two structures , and before any MORN1 is associated with the IMC of the daughter buds ( Figure 4B ) . RNG2 resolves into rings before the inner membrane complex protein IMC1 is associated with daughter cells ( Figure 4D , see also Figure 5E ) . Together these data suggest that RNG2 is recruited to centrosomes immediately after their duplication , and then resolves into rings during initiation of daughter cell pellicle buds . To examine the function of RNG2 we generated an inducible knockdown cell line ( iΔHA-RNG2 ) by 5′-replacement of the native RNG2 promoter with the tetracycline regulatable promoter ( t7s4 ) in a Δku80/TATi background [42] , [43] . An HA coding sequence was also appended to the 5′-terminus of the gene to follow RNG2 expression , and subsequently a cMyc tag at the 3-terminus ( Figure 5A ) . Correct integration into the single rng2 locus was verified by PCR , Western blot , and immunofluorescence microscopy . By PCR , the RNG2 coding region of the iΔHA-RNG2 mutant occurred downstream of the t7s4 promoter ( P2 , 3; expected fragment size 2281 bp ) and not the native promoter region ( P1 , 3; expected fragment size 2134 bp ) ( Figure 5A and B ) . Western blots show that the N-terminal HA tag labels a >188 kDa protein in iΔHA-RNG2 cells , consistent with the predicted size of RNG2 ( Figure 5C ) . This is of identical size to RNG2-HA that we previously generated by 3′ endogenous gene tagging ( Figure 5C ) . Further , Western blotting against cMyc in the iΔHA-RNG2-cMyc cell line confirms this correct subsequent gene-tagging event , including some common minor presumed C-terminal processed or degradation products seen in both RNG2-HA and RNG2-cMyc cell lines ( Figure 5C ) . Finally , immuno-localization of HA in iΔHA-RNG2 cells ( and cMyc in iΔHA-RNG2-cMyc cells ) ( Figure 5E ) shows the protein at apical rings that are first observed early in daughter cell formation , consistent with the RNG2-HA localization and behavior ( Figures 1-4 ) . Together , these data indicate that we have successfully replaced the native promoter of RNG2 with the regulatable t7s4 promoter . The expression levels of HA-RNG2 in the iΔHA-RNG2 cells grown in the absence of the tetracycline analogue , anhydrotetracycline ( ATc ) is equivalent to RNG2-HA that is driven by the native promoter ( Figure 5C ) . When iΔHA-RNG2 parasites were incubated in the presence of ATc ( 0 . 5 μg ml−1 ) , HA-RNG2 expression was reduced to below detectable levels by Western blot within one day of ATc incubation ( Figure 5D ) . Similarly , we were unable to detect HA-RNG2 protein by immunofluorescence assay ( IFA ) after one day of ATc incubation ( Figure 5E ) . These data indicate effective knockdown of RNG2 . To test for a growth phenotype associated with RNG2 knockdown , we measured growth across eight days using a plaque assay . In the absence of ATc , iΔHA-RNG2 cells developed plaques in host cell monolayers that were indistinguishable from plaques in the unmodified parental cell line ( Figure 5F ) . When ATc was present , plaque sizes were dramatically reduced in iΔHA-RNG2 cells but not the parental cells , indicating that RNG2 plays a significant role in parasite proliferation . The early presence of RNG2 during daughter pellicle bud assembly suggested that knockdown of RNG2 might perturb this important developmental stage and this could be responsible for the growth phenotype observed . We tested this by measuring parasite replication rates in host cells with or without RNG2 knocked down . iΔHA-RNG2 cells were cultured for three days , with or without ATc , then allowed to invade fresh host cells . After 24 hours of further incubation with or without ATc , infected monolayers were fixed and parasites immuno-stained with surface antigen SAG1 to count the number of parasites per vacuole . There was no significant difference in the replication rate of cells with or without RNG2 expression . Most vacuoles in both treatments contained eight cells ( three division cycles ) , with equivalent distributions of vacuoles either lagging ( two or four cells ) or ahead ( 16 or 32 cells ) of this replication state ( Figure 6 ) . The morphology of the +ATc replicating cells also appeared normal by SAG1 labeling , with the only notable difference being a reduction in the total number of vacuoles ( see below ) . We also tested for structural phenotypes associated with RNG2 knockdown . Formation of daughter inner membrane complexes , as measured by IFA , appeared unimpaired in parasites depleted of RNG2 ( Figure 5E , bottom panel ) . Additionally , we examined the ultrastructure of cells grown with and without ATc by transmission electron microscopy and observed no differences in the apical complexes , including the apical polar ring , conoid , cell pellicle and position of rhoptries and micronemes ( Figure 7A ) . The apical polar ring marker RNG1 was examined in cells replicating in the presence of ATc to determine if recruitment of this late ring marker requires RNG2 . RNG1 localization to the apical polar ring was unchanged from wild type in the absence of RNG2 ( Figure 7B ) , consistent with the loss of RNG2 having no effect on the structure of the apical polar ring . To test if the integrity of the sub-pellicular microtubular basket was perturbed by RNG2 knockdown , we performed detergent extraction of parasite pellicles . Without ATc , extracted pellicles were typical , with all microtubules anchored to the apical polar ring and conoid ( Figure 7C ) . RNG2 rings remained stably associated with the extracted pellicles , indicating a detergent-resistant , strong association . Extracted pellicles from ATc-treated cells were identical with the exception of lacking RNG2 staining ( Figure 7D ) . The ability of the conoid to be extruded from below the apical polar ring was also examined in knockdown cells . Extracellular parasites were stimulated using A23187 and the number of cells with everted conoids scored . No significant difference was seen in the rates of extrusion between iΔHA-RNG2 cells treated without ATc ( 81 . 4% , +/− 10 . 4% SD ) or with ATc ( 84 . 6% , +/− 6 . 0% SD ) . We conclude that loss of RNG2 does not impair intracellular growth or daughter cell development . In the absence of a replication defect to explain the growth phenotype of RNG2 knockdown , we investigated host cell invasion . We assayed invasion efficiency using a red/green invasion assay [44] , [45] . iΔHA-RNG2 parasites were cultured for two days with or without ATc , mechanically harvested from the host cells in potassium-rich Endo buffer , and then activated for invasion in low potassium buffer and allowed to invade host cells for 10 minutes . Invasion was then arrested by chemical fixation and parasites differentially immuno-stained according to whether they were outside or inside the host cells . The RNG2 knockdown parasites showed a strong invasion defect , with 66% reduction in invasion efficiency ( Figure 8A ) . Gliding motility is integral to invasion , and we tested the ability of iΔHA-RNG2 parasites to glide with or without ATc . Without ATc , parasites showed typical extracellular gliding activity , indicated by trails of surface protein SAG1 left on coated coverslips ( Figure 8B ) . These trails were ablated by treatment with the actin inhibitor cytochalasin D . In RNG2 knocked down parasites ( +ATc ) , we could see no such trails or evidence of gliding motility . Gliding , however , could be restored in RNG2 knockdown cells by treatment with the calcium ionophore A23187 ( Figure 8B ) . To further dissect the events of parasite invasion , we monitored evacuole formation , which is the release of rhoptry contents into the host cell [46] . To do this we allowed parasites to settle onto host cells in the presence of cytochalasin D . This enables parasites to apically attach to host cells , and secrete their rhoptry contents as evacuoles , but prevents them from further invasion . After 15 minutes , we fixed parasites and visualized evacuoles by immuno-staining with the rhoptry marker ROP1 . Evacuoles can be seen as extended ROP1-positive emissions within the host cell . For cells grown without ATc , 44% of parasites formed evacuoles , which is consistent with controls in previous experiments ( e . g . [47] ) . For iΔHA-RNG2 cells grown with ATc , only 27% of parasites generated evacuoles , which is a 39% reduction compared to controls ( Figure 8C ) . Staged release of invasion factors from secretory organelles—micronemes , rhoptries and dense granules—facilitates the coordinated events of parasite invasion and establishment of the parasitophorous vacuole [14] . In extracellular parasites , some secreted micronemal proteins translocate to the parasite surface , where they function in attachment of parasites to the host cells , a critical component of processes such as motility and invasion . Subsequent to their function , these micronemal proteins are cleaved by parasite proteases and released into the extracellular medium [48] . We assayed the release of two species of micronemal proteins , MIC2 and AMA1 , from extracellular parasites . In the absence of ATc , iΔHA-RNG2 constitutively secreted both MIC2 and AMA1 ( Figure 9 ) . In ATc-treated iΔHA-RNG2 cells , we observed a marked microneme secretion defect for both MIC2 and AMA1 , with little or no detectable protein in the extracellular medium ( Figure 9 ) . To check that MIC2 and AMA1 were synthesized to equivalent levels in ATc-treated cells we assayed for these proteins in intact parasites , and found no reduction in either microneme protein type ( Figure S2A ) . IFAs against AMA1 and MIC2 similarly showed no difference in microneme distribution in the cells , predominantly towards the cell apex , when cells are depleted of RNG2 ( Figure S2B , C ) . This indicates that the reduction in microneme proteins in the supernatant was caused by reduction in secretion rather than available protein . Microneme secretion is regulated by two main signaling pathways: activation of calcium-dependent protein kinases ( CDPKs ) by calcium release from the ER; and activation of Protein Kinase G ( PKG ) by cGMP [49]–[51] . These two pathways can be experimentally manipulated by either inducing calcium store release into the cytoplasm with the ionophore A23187 , or by elevating internal cGMP concentration by treatment with cGMP-specific phospho-diesterase inhibitor Zaprinast . Microneme secretion is upregulated by both A23187 and Zaprinast , and we saw equally robust secretion measured by MIC2 and AMA1 markers with both these stimuli in cells where RNG2 was present ( -ATc ) ( Figure 9 ) . In RNG2 knockdown parasites ( +ATc ) we also observed upregulation in microneme secretion ( Figure 9 ) . Interestingly , however , we observed distinct responses to calcium and cGMP stimuli . Calcium stimulation produced upregulation of secretion of both MIC2 and AMA1 in the knockdown cells to levels equivalent to the RNG2-expressing parasites ( Figure 9C , F ) . Although cGMP stimulation did elevate MIC2 and AMA1 secretion , this was clearly less in RNG2 knockdown cells than when RNG2 was present ( Figure 9C , F ) . To test this further , we stimulated microneme secretion with 8-Br-cGMP , an alternative activator of the cGMP pathway . We observed a similar muted response in microneme secretion when RNG2 was absent ( Figure S3 ) . These results suggest that RNG2 is critical for the constitutive secretion of micronemes , and demonstrates a relative insensitivity of microneme secretion to the cGMP pathway in parasites lacking RNG2 .
The apical polar ring is the unifying structural feature of the apical complex , and arguably of all of Apicomplexa [10] . It provides a structural basis for the assembly of new daughter cells , and a focal point for invasion events into host cells . However , its molecular composition is poorly characterized , and its specific functions are experimentally untested . The T . gondii protein RNG2 provides insight into the early formation of this structure , its molecular interactions with other apical complex structures , and the molecular function of this gateway into the host cell . The earliest events of cell division in T . gondii are extension and fission of the Golgi apparatus and the duplication of the centrosomes [20] , [37] , [52] . Immediately following centrosome duplication , the first recruitment of molecules associated with the new daughter cell pellicles is seen . Rab11B and IMC15 appear closely associated with the centrosomes shortly after their duplication [18] , [19] . Rab11B is implicated in trafficking Golgi-derived vesicles to the developing alveolar sacs of the daughter cell pellicles , and IMC15 is one of the earliest pellicle scaffolding proteins that likely contributes to the coordination of alveolar sac assembly . We show that the earliest nascent components of the apical complex are also recruited to the centrosomes just after their division . RNG2 appears initially as a dot in contact with each centrosome before the mitotic spindle is evident . RNG2 then resolves into a ring associated with the apical polar ring as the nascent apical complex separates from its centrosome . Some residual RNG2 remains associated with the centrosome during daughter cell development suggesting that RNG2 interacts directly with the centrosome . Centrin2 is permanently associated with the centrosomes , but it also locates to the preconoidal rings , basal complex and peripheral annuli during daughter cell assembly [23] , [25] , [39] . It is unclear if centrosomal centrin2 contributes to these latter structures . The apical polar ring acts as the microtubule organizing center that gives rise to the subpellicular array of 22 microtubules [8] , [10] . Knockdown of RNG2 has no effect on daughter cell formation or stability of the microtubule array , so it is evidently neither necessary for ring formation or recruitment of other proteins required to form this ring . Rather , RNG2 is presumably itself recruited to the new ring . It therefore presents a conundrum as to why such early association of RNG2 within the apical complex occurs , if no early phenotype of its loss is evident . It is possible that the order of assembly of the apical complex might simply require an early recruitment of RNG2 . The T . gondii pellicle is a remarkably robust and stable structure , including unusually static associated microtublues [11] , and even after detergent extraction RNG2 remains in place . Perhaps correct integration of RNG2 in association with the apical polar ring and conoid necessitates early addition , despite the functional role awaiting daughter maturation . A growing theme in apicomplexan biology is the key role of the centrosome in coordinating the essential structures of new daughter cells , including the Golgi apparatus , apicoplast , nucleus and pellicle buds [18] , [37] , [53]–[59] . Even after the daughter buds separate from the centrosomes , a striated fiber , homologous to rootlets of the flagellar apparatus found in flagellated cells , provides a tether between the centrosomes and the daughter pellicles as they continue to develop [57] . RNG2 suggests an additional , direct role for the centrosome in recruiting proteins that are then assembled into the new apical complexes . This behavior is similar to that seen in the basal bodies of ciliates that also appear to provide recruitment surfaces for repetitive , charged residue-containing proteins that are subsequently deployed to other cell pellicle structures [60] . Basal bodies are homologous structures to the centrioles of centrosomes , and RNG2 was identified by broad similarity to such pellicle proteins of ciliates [29] . The resolvable displacement of the N- and C- termini of RNG2 in mature parasites , and their non-identical behavior during conoid extrusion , suggests that RNG2 interacts both with the apical polar ring and the conoid . The C-terminus forms a ring consistent with position , diameter and static behavior of the apical polar ring [8] . Predicted acylation sites ( at least two palmitoylation sites by CSS-Palm 3 . 0 , [61] using the highest stringency threshold , and up to eight ) might facilitate association with the apical portion of the alveolar sacs , as is known to occur for some IMC-bound proteins [62]–[66] . The N-terminus , on the other hand , associates with the mobile conoid at a position posterior to the conoid protein CAM1 . When the conoid is extruded the N-terminus is clearly anterior to the C-terminus , suggesting that it is not attached to the very base of the conoid which is aligned with the apical polar ring in this state [8] . The simplest model for RNG2 , therefore , is that it forms a collar between the apical polar ring and the conoid ( Figure 10 ) . During conoid extrusion , a substantial reorientation of the RNG2 collar occurs , with the ring created by the N-terminus passing through that of the RNG2 C-terminus , and the collar turning inside-out ( Figure 10B ) . It is notable that RNG2 knockdown does not effect conoid extrusion , indicating that RNG2 itself does not function in conoid extrusion . Instead , we speculate that conoid extrusion may mediate an altered RNG2 conformation , and changed presentation of RNG2 surfaces to other apical organelles or molecules . It is possible that this enables changed RNG2 function . Given the role of RNG2 in invasion—and more specifically in microneme secretion—conoid extrusion could function as an invasion-ready switch to enable RNG2 to perform its role . The dramatic reduction in parasite growth with RNG2 knockdown provides the first experimental insight into the role of the apical polar ring in apicomplexan biology . Loss of RNG2 was associated with reduced parasite proliferation , motility and invasion . Secretion of both the micronemes and the rhoptries was markedly reduced . A reduction in microneme secretion alone could explain all of these phenotypes given that motility and evacuole formation are necessary for invasion , and are dependent on proteins secreted from micronemes such as the adhesin MIC2 for motility , MIC8 for rhoptry secretion and AMA1 for orientating the cell for moving junction formation [47] , [67]–[70] . Reduction in microneme secretion was not due to any obvious perturbation of microneme formation or maturation as RNG2 depletion did not reduce microneme protein content or presentation of micronemes to the apical portion of the cells . The loss of microneme secretion in RNG2 knockdown cells was not absolute , with some secretion still occurring . Inhibition of secretion appeared to be more pronounced for MIC2 compared to AMA1 . This might reflect different sensitivities of respective immuno-detections , or could indicate alternative secretion pathways for these two micronemal proteins . Recently , pools of different micronemal proteins have been shown to rely on distinct Rab-GTPases [71] . While both MIC2 and AMA1 were in the Rab5A/C-independent pools , it is unknown if there is further division amongst these different proteins . Despite a reduction in invasion , approximately one third of RNG2 knockdown parasites were still able to invade , suggesting that they are able to overcome the microneme secretion defect . It is technically possible that , upon knockdown , there remains a residual amount of RNG2 protein that allows a low level of microneme secretion and subsequent invasion . Notably , however , recent knockout studies of the motility associated proteins MyosinA , MIC2 and Actin1 showed that parasites lacking these proteins were still able to invade host cells , albeit with dramatically reduced efficiency [72] . Similarly , approximately 15% of AMA1 knockdown cells remain invasion competent , suggesting this protein is also not completely essential to parasite invasion [47] . It is conceivable , therefore , that a RNG2-independent invasion pathway exists , mirroring the presumed “alternative” invasion pathways seen in the MyosinA , MIC2 and Actin1 knockouts [72] . The inhibition of microneme secretion by RNG2 knockdown was not due to a mechanistic block in secretion , as agonists of either the calcium- or cGMP-signaling pathways were able to reverse this inhibition . Indeed , the restoration of gliding motility in RNG2 knockdown cells by calcium stimulation ( A23187 ) is consistent with this motility phenotype being due to lack of microneme secretion . Calcium and cGMP signaling have been implicated in regulation of key events in parasite invasion cycles [51] , [73]–[75] . These molecules are believed to be second messengers for various extracellular stimuli , and in T . gondii act on up to twelve calcium-dependent protein kinases ( CDPKs ) and a cGMP-dependent protein kinase G ( PKG ) , respectively [51] . While there is evidence of a division of labour amongst some of these different protein kinases for discrete events of invasion ( e . g . TgCDPK1 ) and egress ( e . g . TgCDPK1 and 3 ) , there is also some level of redundancy and/or codependency on calcium and cGMP signaling [50] , [64] , [65] . For instance , while TgCDPK1 is required for microneme secretion during invasion [50] , the PKG inhibitor compound 1 will also block this secretion [49] . We found that the response of microneme secretion in RNG2 knockdown cells to calcium stimulation was the same as that of cells retaining RNG2 . However , for cGMP stimulation the response in the knockdowns was markedly less than in RNG2-expressing parasites ( using either Zaprinast or 8-Br-cGMP ) . This indicates that parasites depleted in RNG2 are less sensitive to cGMP with respect to microneme secretion , and that RNG2 has some role either in cGMP sensing , or is downstream of PKG . If RNG2 has a role in the calcium signaling pathway then it is presumably upstream of calcium sensing , given that overriding calcium release provided full secretion . These data also indicate that activation of the calcium-dependent pathway can complement this cGMP pathway defect , whereas it cannot rescue a stronger inhibitor of the cGMP pathway such as compound 1 [49] . This suggests subtle interplay of multiple layers of regulation of these important cell processes , consistent with the observations of others [49] , [64] . While the role of signaling molecules in the control of invasion events has been recognized for some time , this is the first insight into a location-specific function of the apical complex in these processes . RNG2 implicates the apical polar ring in control of microneme release , creating a regulated gateway for secretion at the apical complex . The precise site of microneme secretion is controversial , with some arguing that release occurs between the apical polar ring and the base of the extruded conoid [76] , rather than at the apical aperture of the conoid . The location of RNG2 between the apical polar ring and conoid base is therefore intriguing , however , we cannot currently say whether this supports basal release , or whether RNG2 controls the onward traffic of micronemes to the conoid apex . RNG2 activity in these processes may be activated directly by CDPK and/or PKG phosphorylation , and phosphoproteomic studies have identified numerous phosphorylated sites in RNG2 , including one calcium-dependent event [30] , [77] , [78] . Alternatively , it is possible that RNG2 may recruit other proteins to the apical polar ring that provide a link in these phospho-signaling events . The repetitive and high charged amino acid content of the class of proteins to which RNG2 belong [29] , and prediction of coiled-coil domains , is consistent with facilitating protein-protein interactions . The sequence of RNG2 is apparently fast evolving , with the homologue in closely related Neospora caninum sharing only 58% amino acid identity . Consequently , homologues are difficult to identify in more distantly related apicomplexans . Nevertheless , we predict that the function of secretion regulation that RNG2 confers on the polar ring of the apical complex is likely conserved throughout Apicomplexa . Possible homologues of the apical complex are also evident in the early ancestor lineages of apicomplexans ( Chromerids and Colpodelids ) as well as some some early-diverging members of the neighboring dinoflagellate lineage ( Perkinsids and Psammosa ) [79]–[83] . These organisms include symbionts , micropredators and parasites , and are believed to use their apical complex for myzocytoic feeding or entry into their metazoan partners . In all these lineages , conspicuous putative secretory organelles are clustered around structural elements of an apical complex , which are intimately associated with , and perhaps derived from , the flagellar apparatus of these flagellate organisms . Recent reports of flagellar-associated proteins ( striated fiber assemblins and SAS6L ) contributing to T . gondii apical complex assembly corroborate this ancestral state of the apical complex [24] , [57] . Moreover , the cGMP signaling pathway is known to correlate specifically with presence of flagella , where it contributes to the important flagellum function of environmental sensing and signal transduction [84] . Together these observations suggest that a key feature of the evolution of the apical complex was likely specialization of flagellar-associated structures in the regulated delivery of secreted factors that facilitate predation and ultimately parasitism .
T . gondii tachyzoites were grown by serial passage in human foreskin fibroblast ( HFF ) cells as previously described [85] . RNG2-HA parasites were previously generated by 3′ endogenous tagging with 3XHA coding sequence of the rng2 gene ( toxodb . org gene ID:TgME49_244470 ) [29] . To generate the conditional RNG2 knockdown parasite strain ( iΔHA-RNG2 ) , we began by amplifying 2015 bp upstream of the RNG2 start codon ( 5′ flank ) using the primers 5′-CTGACATATGGAGACTGCCACAAAGGAAGGTACAC and 5′-GATCATCCATCGAAACGCTCCGTGACGGAAGTA . We digested the product with NsiI and NdeI and ligated this into the equivalent sites of the vector pPR2-HA3 ( Chris Tonkin and GvD , unpublished ) , a modified version of the vector pPR ( a kind gift from Lilach Sheiner , U . Georgia; [42] ) . We next amplified a 2042 bp fragment beginning at the start codon of RNG2 ( 3′ flank ) with the primers 5′-GATCCCCGGGATGCACCCCCACCTTTCTTCCGCAG and 5′-CGATGCGGCCGCGACGGTGGTGTTATTGATTGGTTGC . We digested this with XmaI and NotI and ligated this into equivalent sites of the pPR2-HA3 vector that already contained the RNG2 5′ flank . The resulting vector positions the first RNG2 codon downstream of the ATc-regulatable t7s4 promoter and a 3xHA tag . We linearized the resulting vector with NotI and transfected this into TATiΔku80 parasites ( a kind gift from Lilach Sheiner and Boris Striepen , U . Georgia; [42] ) . Parasites were selected with pyrimethamine and cloned by limiting dilution . To identify parasite clones where the t7s4 promoter had successfully replaced the native RNG2 promoter , we utilized the primers P1 ( 5′- CAGATTCCGAATTCTTTGG ) , P2 ( 5′-TGTAGAGCTGGTGCGTGAG ) and P3 ( 5′-AAGGGGACGCAGTTCTCGGA ) in the combinations described in Figure 5A . For RNG2 cMyc tagging , we PCR amplified a 3′ fragment of rng2 gene using the primers 5′- GATCAGATCTGCAGCTGACACACTCCTGACG and 5′- GCATTCTAGAGTTTGTTGATGCGTCCGAGACAAC , digested this with BglII and XbaI and ligated into the BglII and AvrII sites of the vector pgCM3 , a vector that fuses the 3′ region of a gene-of-interest with a 3× cmyc tag ( NK and GvD , unpublished ) . This vector was linearized with AvrII , transfected into the iΔHA-RNG2 strain , selected on chloramphenicol and cloned by limiting dilution . RNG2 knockdown was induced by culturing with 0 . 5 μg ml −1 of anhydrotetracycline ( ATc ) . All PCRs were performed with Phusion polymerase ( Thermo Scientific ) . To tag the C-terminus of RNG2 with GFP , we digested the pCTG vector [86] with AvrII and BamHI and ligated this into the equivalent sites of the pHA3-LIC-DHFR ( RNG2 ) vector [29] . This exchanged the 3xHA tag with a GFP tag in a RNG2 3′ replacement vector that we termed pGFP-LIC-DHFR ( RNG2 ) . We linearized the resultant vector with NsiI , transfected this into the TATi/ΔKu80 parasites [42] and selected on pyrimethamine . We cloned the resulting drug resistant parasites and confirmed expression of RNG2-GFP by microscopy . We next tagged the N-terminus of RNG2-GFP with a mCherry-3× c-myc tag through a promoter replacement strategy . First , we digested the pPR2-HA3 ( RNG2 ) vector described above with NheI and XmaI to excise the 3x-HA tag . We then digested the mCherry-3× c-myc tag from the vector pCTChM3 ( a kind gift from Chris Tonkin , Walter and Eliza Hall Institute ) with NheI and XmaI and ligated the resulting fragment into the pPR2 ( RNG2 ) vector to generate the vector we termed pPR2-GFP ( RNG2 ) . We next replaced the pyrimethamine-resistance cassette in this vector with a chloramphenicol-resistance cassette . We PCR amplified the chloramphenicol-resistance cassette from the vector pgCM3 using the primers 5′-GATCATGCATAAAACCCTCGAAGGCTGCTAGTAC and 5′-GATCACTAGTGGATCCCCCTCGGG . The resulting PCR product was digested with SpeI and NsiI and ligated into the equivalent sites of the pPR2-GFP ( RNG2 ) vector to generate a vector we termed pPR2-CAT-GFP ( RNG2 ) . We linearized this vector with NotI and transfected this into the RNG2-GFP cell line , selecting on chloramphenicol . We cloned drug-resistant parasites , and confirmed integration through Western blotting . For C-terminal 3XcMyc-tagged RNG1 and CAM1 ( via 3′ endogenous gene replacement ) the coding sequence of each gene was amplified and cloned into pBTM3 ( GvD , unpublished ) . The primers used to amplify the coding sequence of RNG1 were 5′- GATCAGATCTAAAATGGCGCTAATTCCCTCGC and 5′- GATCCCTAGGCGCCAGGTAGTAGACAGGTGGA , and CAM1 were 5′-GATCCCTAGGTTTATTCGCGGAAGGCAGAGAC and 5′-TGGACTGTGGTCGACGCAGAAG . For transient expression of RNG1-GFP , the 3XcMyc tag was removed from the RNG1-cMyc vector by digestion with AvrII and Not1 and GFP coding sequence ligated in its place . The CAM1-GFP vector was a kind gift by Martin Blume , Bio21 , Australia . To label MORN1 we transiently transfected parasites with a cMyc-tagged MORN1 plasmid ( a kind gift from MJ Gubbels , Boston College ) . Immunofluorescence assays ( IFAs ) were performed as previously described [87] using antibodies and their concentrations listed in Supplemental Table S1 . 3D-SIM was implemented on a DeltaVision OMX V4 Blaze ( Applied Precision ) with samples prepared as described [88] , excited using 488 and 568 nm lasers and imaged using band pass filters at 528 and 608 with a 60× oil immersion lens ( 1 . 42 NA ) . Parasite pellicles were extracted in deoxycholate as previously described [66] . Briefly , we filtered parasites through a 3 μm filter and resuspended them in phosphate-buffered saline . We attached parasites to coverslips with 0 . 1% polyethyleneimine ( PEI ) and extracted in 10 mM deoxycholate for 10 min at room temperature . We fixed parasites in 4% paraformaldehyde for 10 min and then proceeded as for IFAs . Cells or cell extracts were analyzed on a Leica TCS SP2 confocal laser-scanning microscope . Only the brightness/contrast ratio of the images was modified , using Adobe Photoshop CS4 . For live cell imaging , parasites were incubated in glass-bottomed dishes ( MatTek ) in phenol-red free Dulbecco's modified Eagle's medium supplemented with 1% foetal calf serum and antibiotics . During imaging , parasites were incubated in a 5% CO2/air atmosphere in a humidified 37°C chamber . Imaging was performed using a DeltaVision set-up with an inverted Olympus IX71 microscope , an Olympus objective lens ( UPlanSApo , 100×/1 . 40 oil ) , and a Photometrics CoolSNAP HQ2 camera . Images were acquired using 2×2 binning , and deconvolved prior to linear adjustment of contrast and brightness . For transmission electron microscopy , parasites were cultured for three days on 0 . 5 μg ml−1 ATc , fixed in PBS with 2 . 5% paraformaldehyde and 1% glutaraldehyde , post fixed in 1% OsO4 , and pellet-embedded in 1% low-melting agarose . The agarose block was ethanol dehydrated , embedded in LR White resin and polymerized . Ultrathin sections were cut on a Leica Ultracut R microtome , lead and uranium stained and visualized with a Philips CM120 BioTWIN transmission electron microscope at 120 kV . Antibodies and their concentrations used for Western blots are listed in Table S1 . Parasites were filtered through a 3 μm filter , counted by haemocytometer and solubilized in sample buffer ( Invitrogen ) at equivalent cell densities . Standard Western blot detection was performed , with Horse Radish Peroxidase conjugated secondary antibodies detected using SuperSignal West Pico Chemiluminescent Substrate ( Pierce ) . Signal strength was quantified using a BioRad Chemidoc imager . For growth assays extracellular parasites were filtered , counted by haemocytometer , and 500 parasites added to 25 cm2 tissue culture flasks containing a confluent monolayer of HFF cells . ATc ( 0 . 5 μg ml−1 ) was added from the outset of the experiment . To visualize plaque sizes , flasks were aspirated , fixed with 5 ml 100% ethanol ( 5 minutes ) , stained with 5 ml crystal violet solution ( 15 minutes ) then washed once with 1× phosphate-buffered saline ( PBS ) and dried before imaging . For replication assays , parasites grown for three days with or without ATc ( 0 . 5 μg ml−1 ) , harvested and filtered . Equal numbers were allowed to invade HFF cells on coverslips for two hours . Coverslips were washed three times with Dulbecco's modified Eagle's medium ( DMEM ) ( supplemented with 1% FCS , 0 . 2 mM L-glutamine ) to remove uninvaded parasites , and cultured for 24 hours with ongoing +/− ATc regimens . Cells were then fixed and processed for SAG1 IFA and parasite number per parasitophorous vacuole scored . To assess conoid extrusion ability , parasites were grown for three days with or without ATc , harvested , filtered and resuspened in DMEM to 2 . 5×107 parasites ml−1 . A23187 was added to samples at a final concentration of 5 μM ( or equivalent volume of DMSO as a control ) , and parasites incubated for 30 seconds at 37 °C , then fixed with 1 . 25% glutaraldehyde and settled on PEI coated coverslips . Conoid extrusion was scored by phase microscopy with >200 cells counted per replicate ( n = 3 ) . Red/green invasion assays were performed as described previously [44] , [45] . Briefly , parasites were grown for two days with or without ATc , harvested within HFF cells by trypsinisation , then mechanically released by passage through a 26 gauge needle in Endo Buffer ( 44 . 7 mM K2SO4 , 10 mM MgSO4 , 106 mM sucrose , 5 mM glucose , 20 mM Tris-H2SO4 , 3 . 5 mg/ml BSA , pH 8 . 2 ) . Cells were counted and resuspended to 2 . 5×107 parasites ml−1 , and 200 μl allowed to settle onto HFF cells on coverslips in Endo buffer for 20 minutes . Endo Buffer was then aspirated , and replaced with 200 μl of Invasion Buffer ( DMEM supplemented with 3% FCS , 10 mM HEPES , pH 7 . 4 ) . After 10 minutes at 37 °C , cells were fixed with 2 . 5% Paraformaldehyde and 0 . 02% glutaraldehyde in PBS , blocked , then probed with anti-SAG1 ( Abcam ) to label uninvaded cells . Samples were then permeabilized ( 0 . 25%TX100 in 1xPBS ) for 10 minutes and probed with anti-GAP45 to label all cells . IFAs were completed with secondary antibodies as normal , and then imaged using a Leica SP2 confocal microscope . Fields of view were selected observing the green ( anti-GAP45 ) channel only to eliminate biased selection of parasites . Images were processed using the Leica SP2 software , and labeled cells scored according to invaded ( red ) or uninvaded ( red and green ) . A minimum of 200 parasites were scored for each of three biological replicates , and invasion percentage calculated as invaded over total parasites . For motility assays , PEI coated coverslips were incubated with fetal calf serum for two hours . Parasite cultures grown for two days with or without ATc were needle passed , filtered and resuspended to 107 cells ml−1 . 1 ml of parasites was placed on coverslip with either no drug ( plus DMSO to equivalent of drug samples ) , 1 μm cytochalasin D , or 5 μM A23187 , and incubated at 37 °C for 90 minutes . Samples were then fixed and SAG1 IFAs performed . Evacuole assays were performed as previously described [47] . Parasite cultures grown for two days with or without ATc were harvested , pelleted , washed with Invasion Buffer and resuspended to 2 . 5×108 cells ml−1 . Cells were maintained at 20°C in all steps post harvesting . 50 μl of parasite suspensions were mixed separately with an equal volume of either Invasion Buffer alone ( plus DMSO to equivalent of drug samples ) , 1 . 0 mM μM Zaprinast , 10 μM A23187 or 10 mM 8-Br-cGMP ( final concentrations 0 . 5 mM , 5 μm , and 5 mM , respectively ) . Cells were incubated at 37°C for 20 minutes to allow secretion , then arrested on ice for 2 minutes before parasites were separated from secreted soluble proteins by centrifugation at 8000rpm at 4°C for 2 minutes . 85 μl of supernatant was removed , centrifuged at 8000rpm at 4°C for 2 minutes to remove any remaining cells , and 75 μl removed and boiled with Sample Buffer as the secreted protein fraction . The pelleted cells were washed with PBS and then boiled with Sample Buffer . Secreted protein samples were analyzed for MIC2 and AMA1 by Western blot , and cell pellets analyzed for Tom40 to verify equal cell numbers used for the different assay conditions .
|
Apicomplexan parasites comprise major human pathogens , including the malaria-causing parasites Plasmodium spp . , and Toxoplasma gondii that causes birth defects and neurological disorders . Key to the success of this group was the evolution of the apical complex , a structure at the focus of the events of host cell invasion . This structure was recently shown to derive from elements of the flagellar apparatus , and rudiments of an apical complex are used for feeding in related protists . Evolution of host cell invasion in Apicomplexa has entailed development of a coordinated secretion of invasion factors from the cell apex . Little is known , however , of the behaviour or function of the components of the apical complex during invasion . We have characterized a new protein , RNG2 , that forms a ring at the heart of the apical complex in T . gondii . This is a dynamic ring that links the mobile conoid with the apical polar ring , and is assembled as one of the first structures in replicating parasites . When RNG2 is artificially depleted , cells become insensitive to the molecular cues for secretion , and invasion of host cells is blocked . This reveals that the apical complex participates directly in regulating secretion , and controlling the events of invasion .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"signal",
"transduction",
"cell",
"biology",
"cell",
"cycle",
"and",
"cell",
"division",
"biology",
"and",
"life",
"sciences",
"cell",
"processes",
"molecular",
"cell",
"biology",
"cellular",
"structures",
"and",
"organelles",
"cell",
"signaling",
"cytoskeleton",
"parasitology"
] |
2014
|
The Apical Complex Provides a Regulated Gateway for Secretion of Invasion Factors in Toxoplasma
|
The Global Trachoma Mapping Project ( GTMP ) was implemented with the aim of completing the baseline map of trachoma globally . Over 2 . 6 million people were examined in 1 , 546 districts across 29 countries between December 2012 and January 2016 . The aim of the analysis was to estimate the unit cost and to identify the key cost drivers of trachoma prevalence surveys conducted as part of GTMP . In-country and global support costs were obtained using GTMP financial records . In-country expenditure was analysed for 1 , 164 districts across 17 countries . The mean survey cost was $13 , 113 per district [median: $11 , 675; IQR = $8 , 365-$14 , 618] , $17 , 566 per evaluation unit [median: $15 , 839; IQR = $10 , 773-$19 , 915] , $692 per cluster [median: $625; IQR = $452-$847] and $6 . 0 per person screened [median: $4 . 9; IQR = $3 . 7-$7 . 9] . Survey unit costs varied substantially across settings , and were driven by parameters such as geographic location , demographic characteristics , seasonal effects , and local operational constraints . Analysis by activities showed that fieldwork constituted the largest share of in-country survey costs ( 74% ) , followed by training of survey teams ( 11% ) . The main drivers of in-country survey costs were personnel ( 49% ) and transportation ( 44% ) . Global support expenditure for all surveyed districts amounted to $5 . 1m , which included grant management , epidemiological support , and data stewardship . This study provides the most extensive analysis of the cost of conducting trachoma prevalence surveys to date . The findings can aid planning and budgeting for future trachoma surveys required to measure the impact of trachoma elimination activities . Furthermore , the results of this study can also be used as a cost basis for other disease mapping programmes , where disease or context-specific survey cost data are not available .
An estimated 285 million people worldwide live with visual impairment , including 39 million who are blind [1] . Trachoma is the leading infectious cause of avoidable blindness , responsible for the visual impairment of about 1 . 9 million people , of whom 0 . 5 million are irreversibly blind [2] . The disease is a public health problem in 42 countries , with just over 200 million people being at risk , the majority of whom are in Sub-Saharan Africa [3] . The World Health Organization ( WHO ) has called for global elimination of trachoma as a public health problem by 2020 . To achieve it , WHO recommends the use of a strategy known as SAFE . This includes ( i ) Surgery to correct the advanced stage of the disease , known as trachomatous trichiasis ( TT ) , when one or more eyelashes rub on the eyeball; ( ii ) Antibiotics to clear ocular Chlamydia trachomatis infection , and ( iii ) Facial cleanliness and ( iv ) Environmental improvement to reduce infection transmission . The “S” component targets individuals , while “A” , “F” and “E” are community-based interventions . WHO recommends that “A” , “F” and “E” are delivered for at least three years in districts in which ≥10% of children aged 1–9 years are estimated to have the sign trachomatous inflammation-follicular ( TF ) , increasing to 5 years in districts in which ≥30% of children have TF . Data on the prevalence of TF and TT at district level ( where “districts” generally contain populations of 100 , 000–250 , 000 people ) are therefore vital for planning interventions and judging the subsequent success of those efforts [4] . The Global Trachoma Mapping Project ( GTMP ) was launched in 2012 , with the aim of completing the baseline map of trachoma worldwide by estimating the prevalence of TF and TT in all suspected trachoma-endemic areas for which prevalence data were unavailable . Between December 2012 and January 2016 the GTMP examined more than 2 . 6 million people across 29 countries ( Fig 1 ) . In addition to providing the evidence base for commencing district-wide implementation of the SAFE strategy , prevalence surveys are also used to evaluate the impact of trachoma elimination interventions . The need for such surveys will increase in the near future , as elimination efforts are ramped up globally . The cost of population-based surveys has been cited as a reason for delays in conducting impact and surveillance surveys , so providing data on actual costs is important to inform debate [5] . The existing literature on the economics of trachoma elimination is sparse , with only one published study addressing the incremental costs of prevalence surveys [5] , and several other papers focusing on socioeconomic burden , and the cost-effectiveness of interventions [6–9] .
Surveys were carried out in areas defined as evaluation units ( EUs ) . Evaluation units were used to address two issues: ( 1 ) local definitions of a “district” vary greatly in terms of population size and geographical area , and ( 2 ) WHO encourages baseline trachoma prevalence surveys to be conducted at larger-than-district level , where trachoma is expected to be highly and widely endemic [10] . EUs ( rather than districts ) have been used as the principal unit of analysis in order to allow comparability of data on survey costs in this paper . However , cost per districts and clusters are also reported in the tables . Each EU-level survey followed a multi-stage cluster random sampling methodology . First-level clusters were broadly synonymous with “villages” , or represented the lowest administrative unit covered by census data . In each first-level cluster , teams randomly selected a number of households for inclusion , corresponding to the number of households that a team could screen in one day . To estimate a 10% prevalence of TF with absolute precision of ±3% at the 95% confidence level , 1 , 019 children aged 1–9 years needed to be sampled , including a design effect of 2 . 65 to account for the clustered design; this was inflated by a factor of 1 . 2 to allow for non-response . The number of first-level clusters required varied based on the local demographical situation , with 20 set as a minimum . Field teams comprised one trachoma grader and one data recorder , and in most cases a driver and a local guide . ( In a small number of subprojects , due to local norms , one male grader plus one female grader were needed per team to grade male and female subjects , respectively . ) Graders and recorders were recruited by the local health ministry and trained by GTMP-certified trainers , using a standardised training protocol and materials , over five days , consisting of intensive classroom sessions , field based training and testing [4] . In the field , the teams were supervised by an ophthalmologist or senior ophthalmic nurse , who was appointed for every seven to ten teams on average . In selected households , all residents over the age of one year were invited to participate . Consenting individuals were examined for the signs of TT , TF and trachomatous inflammation-intense ( TI ) . Additional questions were asked about households’ access to water and sanitation ( WASH ) , and observation was undertaken of latrine facilities , if present . In some surveys , data were also collected on other diseases of local importance . All data were captured electronically , using a purpose-built Open Data Kit-based Android smartphone application , and uploaded to a cloud-based server for quality assurance checks ( including Global Positioning System ( GPS ) localisation of households surveyed ) and subsequent data cleaning and analysis . The GTMP was implemented as a series of subprojects . A subproject was a group of surveys conducted in a defined geographical area , for which a unique budget was established and expenditure reports prepared . To ensure comparability across countries and surveys , a number of inclusion criteria were introduced for this study . Mapping subprojects were only included in the costing analysis if they 1 ) followed the standardised GTMP methodologies; 2 ) only collected data on trachoma and WASH , or were integrated surveys of trachoma and other diseases in which the trachoma- and WASH-data-related expenditure could be explicitly identified . In this study , the cost of conducting baseline trachoma surveys included only financial expenditure covered by the GTMP . An activity-based costing accounting model was used to analyse survey expenditure . Data on in-kind contributions and expenditure from project partners or health ministries that would have been incurred in the absence of the mapping project were not collected , with the exception of vehicles . For example , any per diems received by health ministry staff during field work were included , but their base salaries were not . When vehicle time was donated by the health ministry or its partners , a cost estimate was included , based on the duration of mapping and the average daily cost of vehicle hire in previous programmes ( mean cost of daily vehicle hire: Malawi MKW90 , 000 , Nigeria NGN15 , 000 , United Republic of Tanzania TZS150 , 000 , Zambia ZMW341 , Zimbabwe USD$180 ) . Finally , administrative costs charged by implementing partners were excluded from the expenditure analysis on the basis that rates were agreed contractually and did not necessarily reflect the actual overhead costs of implementing partners . All data were collected retrospectively between June 2015 and March 2016 . Financial data were obtained via country-specific budgets and expenditure claim forms submitted by in-country implementing partners . Financial data were extracted from the GTMP financial reporting system ( CLAIMS ) or collected via Excel spreadsheets . Other data , such as the list of EUs , districts , and clusters mapped for each subproject , were provided by the GTMP data managers who cleaned and analysed all field data . Specific information on in-country surveys was also collected through narrative reports and email or phone queries to implementing partners . GTMP expenditure was allocated to two categories of activity ( in-country survey and global support expenditures ) and further divided into five sub-categories as indicated in Table 1 . Global support expenditure was broken down into three activities: grant management , epidemiological support , and data stewardship and processing . In-country survey expenditure was allocated to either training or survey implementation sub-categories . Since it was not possible to link training expenditure to specific subprojects ( because training activities often involved groups of trainees destined to work in different subprojects ) , this expenditure has been allocated to each subproject in proportion to the number of EUs surveyed . Details of items included or excluded for each activity and sub-activity are provided in Table 2 . Data were collected and analyzed in Excel and Stata 12 . 0 ( Stata Corp , TX , USA ) . Descriptive statistics were calculated including means , medians and interquartile ranges ( IQRs ) since in-country survey expenditure by country or subproject did not follow a normal distribution . The IQR includes all values between the 25th and 75th percentile , while outliers are defined as any values above the 95th percentile . It was not possible to allocate survey expenditure to specific EUs , districts or clusters , given that financial data were aggregated and reported by partners at subproject level . The mean unit cost of mapping was hence obtained by dividing the total expenditure by the number of districts , EUs and clusters surveyed in each subproject . In-country survey and global operational expenditure were converted in US Dollars ( USD ) using the mean annual market exchange rates over the mapping period [11] . All USD expenses were then inflated by applying the United States Consumer Price Index , using 2015 as the base year [12] . In order to facilitate cross-country comparisons , the mean survey expenditure per cluster was also converted to International Dollars . An international dollar is a hypothetical currency adjusted by purchasing power parity conversion factors to compensate for price level differences between countries [13] . Therefore , each country’s mean expenditure has been divided by the latest Purchasing Power Parity conversion factors available for each country ( PPP , private consumption ) [14] . Figures in International Dollars ( PPP ) are provided in the results section . Ethics clearance for the costing study was granted by the London School of Hygiene & Tropical Medicine ethics committee ( reference 11195 ) . The research used routine financial data from the project and did not require additional primary data collection . The underlying epidemiological data were collected during the GTMP and ethics clearance was obtained from the London School of Hygiene & Tropical Medicine ( references 6319 and 8355 ) and from relevant ethical review boards identified by the ministry of health in each country [4] . Informed verbal consent was sought for each survey participant or from the parent or legal guardian when the participant was minor . Verbal consent was preferred over written consent considering that surveys were generally conducted in rural areas with high illiteracy rates . Verbal consent was requested and recorded using the LINKS application ( purpose-built application on Android smartphones used for data collection ) .
The total cost of the global GTMP support package was $5 , 129 , 348 ( Table 3 ) . This covered centralised support functions for rolling out the GTMP programme in 29 countries , as well as discussions with a further 20 countries in which mapping , ultimately , was not taken forward for one of a variety of reasons , including excessive security risks or insufficient evidence to justify conducting formal population-based prevalence surveys . Global support expenditure included grant management ( 40% ) , epidemiological support ( 31% ) , and data stewardship and processing ( 29% ) . Overall , 89% of in-country survey expenditure was spent on survey implementation and 11% on training . The mean training expenditure per mapping team was $2 , 357 , ranging from $792 per team in Egypt to $8 , 513 in Pakistan . Survey implementation expenditures included fieldwork ( 74% ) , coordination & planning ( 9% ) , and supervision ( 6% ) . When breaking down in-country survey expenditure by inputs , personnel represented the highest share , accounting for 49% of total spend , followed by transportation ( 44% ) , supplies ( 4% ) , and other expenditure ( 3% ) ( Table 4 ) . The median in-country survey cost was $15 , 839 per EU , ranging from $7 , 550 in Niger State , Nigeria to $43 , 580 in Papua New Guinea . The mean cost per EU was $17 , 566 ( IQR $10 , 773–$19 , 915; SD = $9 , 470 ) . As shown in Fig 3 on the “All regions” box plot , the mean unit cost was higher than the median value because of four outliers: Papua New Guinea ( $43 , 580 ) , Vanuatu ( $42 , 795 ) and two regions in Ethiopia ( Ethiopia Somali region $41 , 719; Gambella region , $40 , 767 ) . Fig 3 also shows that subprojects in the Middle East and North Africa had the smallest variance in expenditures per EU , as demonstrated by their relatively low standard deviation ( US$ 4 , 309 ) compared to the Sub-Saharan African countries ( US$ 9 , 277 ) and the Pacific Islands and South Asia ( US$ 12 , 777 ) . At cluster level , the median survey cost was $625 , ranging from $290 in Zanzibar , United Republic of Tanzania , to $1 , 687 in Papua New Guinea ( mean of unit costs $692; IQR $452–$847; SD = $350 ) . Outlier values were Papua New Guinea ( $1 , 687 ) , Ethiopia Gambella region ( $1 , 631 ) and Somali region ( $1 , 558 ) . A summary of in-country survey expenditure by country and subproject is presented in Table 5 . More detailed information is also available in S1 Dataset .
This study presents the costs of conducting trachoma prevalence surveys in suspected endemic populations , with support from the GTMP . It is important to have setting-specific unit costs for conducting trachoma prevalence surveys against which future surveys can be benchmarked . Indeed , the evaluation of trachoma elimination efforts requires re-estimation of TF and TT prevalence following implementation of the SAFE strategy . It is thus anticipated that the need for surveys will increase in the near future as work is ramped up to meet the 2020 trachoma elimination target . We believe that our work’s value is not limited to future efforts to plan and budget for trachoma surveys . As one of the largest infectious disease mapping exercises ever conducted , the GTMP offers important information for mapping other diseases too . We find that the mean cost for trachoma prevalence surveys across our sample was $17 , 566 per EU . However , our analysis also shows that the level of expenditure for trachoma prevalence surveys was substantially different between world regions , countries , and at sub-national level ( Table 5 ) . These variations can be attributed to a number of parameters , including i ) geographic location ( topography , transport infrastructure , etc . ) ; ii ) demographic characteristics of the population ( population density , age groups , etc . ) ; iii ) seasonal effects ( dry or wet season ) and iv ) operational constraints ( timeframe , security , logistics , unexpected events , NGOs's travel policies etc . ) . These parameters had an influence on the means of transportation , travel distance , per diem rates and the duration of mapping , and consequently on transport and personnel expenditure , which represented nearly 85% of the total cost of the surveys . When comparing different parts of the world , these factors were key to accounting for unit cost differences observed between the groups of subprojects in Sub-Saharan Africa , Middle-East and North Africa , and Pacific Islands and South Asia . In the Pacific Islands and South Asia , the survey unit cost and variance ( mean = US$ 24 , 681 per EU , standard deviation = US$ 12 , 777 ) can be explained by low density of population , security risks and relatively high price levels for commodities and services in some countries . In Vanuatu , due to the island configuration and the low density of population , a high number of clusters had to be surveyed across 17 islands in order to reach the required sample size . In addition , in both Vanuatu and Papua New Guinea , survey teams had to travel by aircraft or boat , which incurred high transport costs and increased the vulnerability of the work to weather-related delay , such as tropical storms . For security reasons , training for the Pakistan teams had to be conducted by international staff in Nigeria , increasing the cost of transportation for training participants . Finally , the high prices of commodities and services in island nations such as Solomon Islands and Vanuatu further inflated survey costs there . In Sub-Saharan Africa , unit cost variations between sub-projects can also be explained by local conditions ( mean = $16 , 734 per EU , SD = $9 , 277 ) . The lowest survey costs were in Zanzibar , United Republic of Tanzania ( US$ 7 , 551 per EU ) and Nigeria ( US$ 8 , 577 per EU ) because of relatively high population densities and hence shorter mapping duration . On the other hand , in some regions of Ethiopia ( Gambella , Beneshangul Gumuz , Ethiopia Somali and Afar ) , more resources were required to overcome implementation challenges caused by low population densities , unpredictable environments and harsh climatic conditions , such as extreme temperatures , volcanic activity and flooding . In Afar , for example , temperatures exceeded 45°C in the afternoon , and fieldwork was only possible for half of each day . In addition , remoteness and poor road infrastructure in some districts increased the time and resources required to reach scattered clusters . As a result , the time required to map one cluster increased substantially; resulting in high survey costs for the Gambella ( US$ 40 , 767 per EU ) and Ethiopia Somali ( US$ 41 , 719 per EU ) regions compared to other subprojects . In contrast , with the notable exception of Darfur , Sudan , the contexts in which we worked in the Middle East and North Africa group shared quite similar characteristics , in terms of geography and population density . This underlies both the low mean unit cost ( $15 , 252 per EU ) and low variance ( SD = $4 , 309 ) observed there . There is only one previously published study on the cost of trachoma mapping: by Chen et al . [5] . Their analysis is based on data from eight national programmes in Sub-Saharan Africa , covering a total of 165 districts and 1 , 203 clusters . Their quoted unit costs were lower than ours , with a median cost per cluster of US$344 compared to US$625 for the GTMP , when using the same base year ( 2015 ) [12] . However , caution is advised when comparing the overall cost estimates between these two studies since the survey methodology , countries and districts mapped , and the activities or expenditure included in the analysis , all diverged; in particular , Chen et al . excluded the cost of international support in their district- and cluster-level analyses . Despite these differences , both the previous work and ours found that transport and per diems during field work were the main cost drivers for trachoma prevalence surveys . Chen et al . also found large variations in survey unit costs; ranging from an average cost of US$ 1 , 511 per district in Ethiopia to US$ 25 , 409 for Ayod in Southern Sudan . Moreover , similar local factors affecting activity costs have also been observed on a published costing study concerning lymphatic filariasis [42] . Hence , it is important for funders and program managers to recognize the significant role that local factors play when planning and budgeting . A specific and noteworthy example of this is that the level of supervision and epidemiological support required can vary greatly from one setting to another . Supervision and coordination activities can increase the overall survey expenditure in situations where local capacity is more limited; but are likely to reduce the risk of having to remap due to poor data quality . Similarly , international experts are essential for training where local expertise is not available . In the GTMP , the massive scale , the short timeline for implementation , as well as the drive to produce objectively gold-standard data , required a significant level of investment in global support systems including international coordination , centralised epidemiological support , and single-source provision of data management services though a small , dedicated team . The global support expenditures amounted to US$3 , 318 on average per district ( US$ 5 , 668 per EU ) . Although a global support expenditure of US$ 5 , 668 per EU is substantial when compared to in-country survey expenditure , some of these costs would otherwise have been covered in-country in duplicative fashion , for example , through local costs for grant management , data collection platform set-up , epidemiologist salaries , and creation of data processing systems . Undertaking the GTMP this way would have increased its overall cost . Moreover , the GTMP was designed to be carried out quickly so that trachoma elimination activities could start , where needed , in time to meet the global 2020 elimination target [4] . The consequence is that 1 , 546 districts were surveyed over three years with full GTMP support , compared to 1 , 115 districts mapped in 12 years prior to 2012 . It is possible that survey expenditure would have been lower had survey activities been implemented only when conditions in some areas were felt to be ideal: outside the rainy season , for example , or completely separated from periods of risk for political or civil unrest . However , the higher cost of mapping to the GTMP’s tight deadline is likely to be more than offset by accelerating the elimination of trachoma[43] . Our analysis is based on rigorous financial data obtained from the GTMP for 17 countries , but had some limitations . First , the unit costs presented do not fully encapsulate the total economic cost of conducting trachoma surveys , since they do not include donations and expenditure that were not covered by the GTMP itself , particularly base salaries for health ministry employees ( outside of per diems covered by GTMP ) , apart from vehicles which costs have been estimated and added to the appropriate survey cost ( in instances when vehicles were made available by the health ministry or partners in Malawi , Nigeria , United Republic of Tanzania , Zambia and Zimbabwe ) . Second , budgeting and financial reporting in the GTMP was done by subproject , and it was therefore not possible to allocate expenditure items to specific district-level surveys . Unit costs presented in this paper are hence the average expenditures for EUs , districts , and clusters within a subproject . Finally , in-country survey costs were estimated based on subprojects that met our study inclusion criteria; 382 of 1 , 546 districts surveyed with GTMP support were excluded . Despite these limitations , this study provides data on the setting-specific costs of conducting trachoma prevalence surveys across 52 mapping subprojects and 17 countries , and adds considerably to previous work [5] . The information we provide here can be used by endemic countries and partners to help budget for future surveys . Data on the global distribution of infectious diseases is increasingly available [44]; but significant knowledge gaps remain , even for global priorities such as malaria , HIV and tuberculosis [45–47] . We hope that our work can be used to guide studies that will start to close those gaps .
|
There are currently few data sets available to aid programmes in planning and budgeting for population-based surveys in low- and middle- income countries . With the objective of identifying cost drivers and key variables influencing prevalence survey costs , the authors collected expenses incurred during the Global Trachoma Mapping Project ( GTMP ) which surveyed 2 . 6 million people across 29 countries . Expenditure from surveying 1 , 164 districts in 17 countries was analysed . Our results showed that the majority of in-country expenditure was spent on personnel ( per diems , accommodation , meals and beverages ) ( 49% ) and local transportation ( 44% ) and that the median survey expenditure was US$11 , 675 per district ( or US$15 , 839 per evaluation unit , US$625 per cluster and US$4 . 9 per person examined ) . There were large variations in survey unit costs across settings , based on local geographic , demographic , seasonal effects and local operational characteristics . In addition , the resources required for the global support and coordination of the GTMP were analysed and amounted to US$5 . 1m ( US$3 , 318 per district or US$5 , 668 per EU ) . Global support expenses can be substantial for a large multi-country mapping exercise conducted in a limited period of time such as the GTMP . Findings from this study can be used to inform other disease mapping projects , and to inform planning and budgeting for the prevalence surveys that will assess the impact of trachoma elimination interventions .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"united",
"states",
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"transportation",
"tropical",
"diseases",
"geographical",
"locations",
"social",
"sciences",
"north",
"america",
"bacterial",
"diseases",
"research",
"design",
"data",
"management",
"eye",
"diseases",
"neglected",
"tropical",
"diseases",
"surveys",
"ethiopia",
"africa",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"economics",
"people",
"and",
"places",
"finance",
"survey",
"research",
"ophthalmology",
"trachoma"
] |
2017
|
The cost of mapping trachoma: Data from the Global Trachoma Mapping Project
|
Like many organisms the fungal pathogen Candida albicans senses changes in the environmental CO2 concentration . This response involves two major proteins: adenylyl cyclase and carbonic anhydrase ( CA ) . Here , we demonstrate that CA expression is tightly controlled by the availability of CO2 and identify the bZIP transcription factor Rca1p as the first CO2 regulator of CA expression in yeast . We show that Rca1p upregulates CA expression during contact with mammalian phagocytes and demonstrate that serine 124 is critical for Rca1p signaling , which occurs independently of adenylyl cyclase . ChIP-chip analysis and the identification of Rca1p orthologs in the model yeast Saccharomyces cerevisiae ( Cst6p ) point to the broad significance of this novel pathway in fungi . By using advanced microscopy we visualize for the first time the impact of CO2 build-up on gene expression in entire fungal populations with an exceptional level of detail . Our results present the bZIP protein Rca1p as the first fungal regulator of carbonic anhydrase , and reveal the existence of an adenylyl cyclase independent CO2 sensing pathway in yeast . Rca1p appears to regulate cellular metabolism in response to CO2 availability in environments as diverse as the phagosome , yeast communities or liquid culture .
Atmospheric carbon dioxide ( CO2 ) with a concentration of 0 . 039% is not only central to the Earth's biogeochemical carbon cycle but is also sensed as a signal by many organisms . The nematode and parasite of insects Neoaplectana carpocapsae localizes its prey via a CO2 gradient [1] , while avoidance behaviour in another nematode , Caenorhabditis elegans [2] , or the model organism Drosophila melanogaster is provoked by elevated CO2 [3] . C . elegans detects CO2 via a cGMP-gated ion channel [2] whereas in D . melanogaster CO2 is sensed by a pair of 7 transmembrane domains chemoreceptors localized on specialized sensory neurons [4] . In the fungal kingdom CO2 , under its hydrated form bicarbonate ( HCO3− ) , is critical for cellular metabolism . Although hydration of CO2 to HCO3− and a proton occurs spontaneously , this reaction is greatly enhanced by the metalloenzyme Carbonic Anhydrase ( CA ) , which operates at a rate of up to 106 reactions per second [5] . Fungal CAs fix the membrane permeable gas CO2 as HCO3− inside the cell , which is subsequently used as substrate for fundamental carboxylation reactions including the conversion of acetyl-CoA to malonyl-CoA ( EC 6 . 4 . 1 . 2 ) , or pyruvate to oxaloacetate ( EC 6 . 4 . 1 . 1 ) . The direct relevance of HCO3− synthesis for fungal survival is reflected by the fact that the CA deletion mutants of Candida albicans , Cryptococcus neoformans , Saccharomyces cerevisiae , Sordaria macrospora , Aspergillus fumigatus or Aspergillus nidulans fail to grow in ambient air [6] , [7] , [8] , [9] , [10] . However , when cultured in a CO2−enriched atmosphere , where sufficient HCO3− is spontaneously formed to meet the metabolic requirements , CAs are optional . In fungi CO2 is also sensed as a signal to regulate the expression of virulence factors . In the pathogenic yeast C . albicans , high level of CO2 triggers filamentous growth and the white-opaque switch [7] , [11] . Recently we have shown that in C . albicans CO2/HCO3− is detected by the enzyme adenylyl cyclase Cyr1p which regulates most processes considered essential in C . albicans virulence [7] , [12] . Here , Cyr1p senses CO2/HCO3− by a lysine residue ( position 1373 ) of the C-terminal catalytic-site [13] potentially linking HCO3− , generated by CA , and cAMP signaling . In humans CAs are involved in medically relevant processes including bone calcification , or renal clear-cell-carcinoma progression; consequently , understanding their regulation and use of inhibitors has attracted considerable interest [14] . This led to the identification of the first regulator of CA , the bHLH transcription factor HIF-1α , which controls the expression of major hypoxia-induced genes including CA IX [15] . Another recently identified CA regulator is AphB from Vibrio cholera [16] . This LysR-type transcription factor also activates the ToxR virulence cascade via the tcpPH operon which ultimately induces the production of cholera toxin . Notably the CAs of S . cerevisiae [17] , [18] , S . macrospora [8] , A . fumigatus and A . nidulans [9] are expressed in response to the availability of environmental CO2 . However , fungal genomes do not posses orthologs of either HIF-1α or AphB-type CA regulators . This suggests the existence of an , as yet , undiscovered CO2 signaling mechanism controlling fungal CA expression . In this report we investigate the existence of such a pathway in fungi by using , as a model , the well characterized CO2 sensing system of the pathogenic yeast C . albicans . We have shown that C . albicans posses a single β-CA , required for growth under CO2 limiting atmosphere [7] . We now demonstrate that the expression of both transcript and protein of this CA is controlled by the level of environmental CO2 and that CA is further induced in an ex vivo model of phagocytosis by mammalian phagocytes , suggesting that CO2 might be limiting even in the relatively high CO2 conditions in the host . We find that such regulation in C . albicans is independent from the already known sensor adenylyl cyclase , described above , suggesting the existence of a cAMP-independent CO2 signaling pathway in fungi . By implementing a systematic functional screen we identify the bZIP transcription factor Rca1p as the C . albicans regulator of CA expression in response to CO2 availability . Furthermore , by using Chromatin Immuno Precipitation ( ChIP ) on chip and ChIP-qPCR experiments we confirm that Rca1p binds to the CA promoter , and to 84 additional genes . The broad significance of our findings is further underlined by our data revealing the existence of a conserved CO2 sensing pathway controlled by an Rca1p ortholog in the model organism S . cerevisiae . Finally , using advanced microscopy , we contribute to the understanding of CO2 flux and metabolic adaptation inside yeast populations on a hitherto unprecedented level of resolution .
The CA , Nce103p , from C . albicans and S . cerevisiae are known to be required for growth in ambient air ( [6] , [7] and Figure 1A , B ) . However , to allow an in-depth analysis of yeast CA expression we developed an antibody directed against C . albicans Nce103p , and additionally constructed a strain expressing a functional tagged CA in S . cerevisiae: Scnce103Δ+ScNCE103-GFP ( Figure 1B ) . Since the CAs from C . albicans and S . cerevisiae are required for growth in a low but not high CO2 atmosphere we asked whether expression of the enzyme itself is regulated by the availability of this gas . To address this question we performed Western-blot analysis using appropriate antibodies to detect C . albicans CA and CA chimera of S . cerevisiae . Single bands with the predicted molecular weights for CaNce103p ( 32kDa ) , and ScNce103-GFPp ( 54kDa ) were detected and we now show for the first time that CA protein expression is highly regulated in both yeast species ( Figure 1C ) . In fact , CA is strongly expressed when yeasts are grown in ambient air but non-detectable when cultured in air enriched with 5 . 5% CO2 , precisely mirroring the requirements for growth of the NCE103 gene in yeast ( Figure 1A , B ) . Furthermore , quantitative real time polymerase chain reaction ( qRT-PCR ) analysis with reverse transcripts of total RNA extracted from C . albicans and S . cerevisiae grown at low and high CO2 concentrations show a similar regulation of CA transcript when expressions were normalized to ACT1 ( Figure 1D ) , confirming previous reports made in S . cerevisiae [17] , [18] . We have previously shown that the adenylyl cyclase Cyr1p from C . albicans functions as a major CO2-sensor , promoting the yeast-to-hyphae switch in response to high levels of CO2 [7] , [13] . We now asked if CO2 regulation of CA expression was similarly coordinated by Cyr1p or cAMP signaling examining CA protein and transcript levels in a strain where both alleles of CYR1 have been deleted ( cyr1Δ ) . Notably , CO2 regulation of both protein and transcript levels of NCE103 remain unaltered in the cyr1Δ strain displaying a pattern of expression identical to the control strain CAI4+pSM2 ( Figure 1E ) . Furthermore , Western-blot or qRT-PCR analysis revealed that supplementation of culture media with exogenous cAMP at concentrations known to mimic Cyr1p activity [19] ( 10 mM ) did not affect CA expression in CAI4+pSM2 grown in ambient air ( Figure S1 ) . Similarly to C . albicans , addition to the growth media of 10 mM cAMP did not affect the expression of CA protein or transcript in S . cerevisiae ( Figure S1 ) . Taken together our data demonstrate that CO2 regulation of CA in C . albicans is independent of the known CO2 sensor Cyr1p and its product cAMP . Furthermore , they strongly support the existence of a novel , cAMP-independent , CO2 signaling pathway in yeast . To identify the key components of this novel CO2 sensing pathway we systematically screened a C . albicans knock-out library searching for strains with an altered expression of their CA in response to CO2 . The library consisted of 158 C . albicans non-essential transcription factor mutants ( provided by D . Sanglard ) . CA protein expression was investigated in each mutant grown in either ambient air or air enriched with 5 . 5% CO2 . Repeated screening identified a single candidate ( HZY7-1 ) that failed to induce CA protein when grown in ambient air . HZY7-1 harbors a mutation in the C . albicans orf19 . 6102 gene . To confirm the HZY7-1 phenotype , we independently inactivated the two orf19 . 6102 alleles in a CAI4 background , using the URA-blaster approach [20] , and re-introduced URA3 at its native locus to generate strain rca1Δ . Subsequent to validating gene inactivation by Southern blot and qRT-PCR ( Figure S2 and S3 ) , we confirmed a striking loss of Nce103p protein induction in rca1Δ in ambient air ( Figure 2A ) . We also validated that during the inactivation process , we did not alter the expression of 2 genes partially overlapped by RCA1: orf19 . 6103 and MVD ( Figure S3 ) . In light of these findings we named the gene encoded by orf19 . 6102: Regulator of Carbonic Anhydrase 1 . RCA1 encodes a 283 amino acid ( aa ) hypothetical protein which contains a conserved basic leucine zipper ( bZIP ) domain in its C-terminus , required for DNA interaction ( Figure S4 ) . Reintroduction of RCA1 , either on its own ( rca1Δ+RCA1 ) or tagged at its C-terminus with Haemagglutinin ( rca1Δ+RCA1−HA3 ) , into the rca1Δ strain restored CA protein induction in C . albicans cells exposed to low CO2 level ( Figure 2A ) . These observations were also confirmed by qRT-PCR ( Figure 2B ) . In a previous study of transcriptional variation that follow phagocytosis of C . albicans by murine macrophages , NCE103 was found to be mildly induced ( ∼1 . 9-fold after two hours of co-culture [21] , while in S . cerevisiae , NCE103 was one of the genes most highly induced by phagocytosis ( 13 . 8-fold; [22] ) . We assayed expression of CaNCE103 in phagocytosed C . albicans cells after one hour of co-culture by qRT-PCR and found an induction of 2 . 1-fold relative to cells in media alone , even though both populations were exposed to a high-CO2 environment ( 5 . 0% in a tissue culture incubator ) . This change is of similar magnitude , but slightly faster , than observed by microarray . This induction was completely absent in an rca1Δ strain ( Figure 2C ) . These results indicate that Rca1p regulates CaNCE103 in a physiological environment which could be correlated to a CO2 concentration scarcer within the immune cell due to a limited penetration across multiple membranes , the sequestering activity of the mammalian CAs , or the reduced metabolic production of CO2 in the fungal cell as a result of a shift to slower , and gluconeogenic growth . Since CA expression in S . cerevisiae is also controlled by ambient CO2 levels we investigated the existence of Rca1p orthologs in this yeast . In S . cerevisiae , we identified Cst6p ( BLAST; Score: 117; E value: 1e−16 ) as a potential Rca1p ortholog . Cst6p encodes for a putative 587 aa protein with a bZIP domain in the C-terminus ( Figure S4 ) . In order to prove that Cst6p is a yeast CA regulator we constructed the mutants in the S . cerevisiae ScNCE103-GFP background ( ScNCE103−GFP+cst6Δ ) . Successful gene inactivations were confirmed by diagnostic PCR and qRT-PCR ( Figure S2 and S3 ) . Using anti-GFP antibodies for ScNCE103−GFP+cst6Δ we found that its CA , similar to the C . albicans rca1Δ strain ( Figure 2A ) , was not induced in low ambient CO2 when compared to the controls ( Figure 3A ) . This regulation in S . cerevisiae mutant was also confirmed at transcript level by qRT-PCR ( Figure 3B ) . In the mutant , introduction of a plasmid expressing CST6 restores the expression of NCE103 in air ( Figure 3A ) . Taken together these data show that CO2 regulation by Rca1p orthologs is conserved in yeast . S . cerevisiae Cst6p is a transcription factor previously described to bind a specific DNA motif: TGACGTCA [23] . We identified this motif in the NCE103 promoters of S . cerevisiae ( position -285 bp to ATG ) , but not of C . albicans . To assess the role of this motif in CO2 regulation of CA expression we neutralized it by removing 7 and 4 bases pairs of the TGACGTCA sequence in the promoters controlling Nce103p expression in S . cerevisiae . Notably the resulting strains ( Scnce103Δ+ScNCE103−GFP−MUT ) failed to induce CA when exposed to low environmental CO2 ( Figure 3B ) , exactly mirroring the expression pattern displayed by the S . cerevisiae cst6Δ mutants ( Figure 3B ) . In summary , our data show that CO2 regulation of CAs expression in yeast is controlled by a conserved transcriptional factor , but involves divergent DNA motifs between S . cerevisiae and C . albicans . To confirm that Rca1p directly binds to the CA promoter , and identify any additional genes it controls , we performed Chromatin Immuno Precipitation on Chip ( ChIP-Chip ) in air and air enriched with 5 . 5% CO2 . We introduced the HA-tagged RCA1 allele , described above , into the heterozygous RCA1 mutant ( rca1Δ/RCA1 ) . The resulting strain ( rca1Δ/RCA1+RCA1−HA3 ) expressed one wild-type and the HA-tagged RCA1 copy . Next , we confirmed that CA levels in rca1Δ/RCA1+RCA1−HA3 and in control strain rca1Δ/RCA1+RCA1 were fully responsive to CO2 by Western blotting , using anti-Nce103 antibodies ( Figure S5 ) . Subsequently , genome-wide location profiling of Rca1-HA3p in low and high CO2 using C . albicans whole-genome oligonucleotide tiling arrays [24] produced a total of 182 binding peaks , when the experiment was carried out in air , and 140 in air enriched with 5 . 5% CO2 ( log2-transformed pseudomedian signal intensity cutoff: 0 . 5; P≤0 . 01 ) ( Table S1 and S2 ) including 61 common “hits” between the two conditions . In depth analysis revealed that 85 of the hits could directly be associated with ORFs ( hits located within 1kbp before an ATG start codon ) ( Figure 4A ) . Notably , we identified significant enrichments of several consecutive probes localized in the promoter of NCE103 ( between position −654 and −479bp before ATG ) in samples extracted from cells grown in air but not in those supplemented with 5 . 5% CO2 ( Figure 4B ) . In addition to the promoter region , Rca1p binding was also enriched in the coding region of NCE103 ( Figure 4B ) . This binding profile has been previously reported for another bZIP transcriptional factor , Cap1p , and suggests binding of the protein to the transcriptional machinery [25] . To confirm the association of Rca1p to the CA promoter , we performed ChIP in tagged and untagged strains grown in low and high CO2 , followed by a qPCR with primers specifically designed to amplify the predicted binding region of Rca1p on the NCE103 promoter . As expected , we observed a 2 . 13 fold enrichment of this sequence in the tagged strain compared to the untagged strain in air , compared to only 1 . 12 fold in 5 . 5% CO2 ( Figure 4C ) . These results show a significant association of Rca1p to the promoter of NCE103 in air compared to the high CO2 environment . With respect to the other Rca1p associated genes , forty four of the 85 hits were specific to ambient air samples , 19 to enriched CO2 and 22 shared between the two conditions ( Figure 4A ) . Rca1p binding peaks were directly associated with 4 other putative transcription factor encoding genes ( CTA24 , TFB3 , ZCF4 and ZCF22 ) and 2 genes involved in cell wall biosynthesis ( CHT2 encoding a chitinase and OCH1 coding for a α-1 , 6-mannosyltransferase ) . Since both CHT2 and OCH1 are involved in C . albicans virulence we selected them to examine the predicted role of Rca1p on their expression by qRT-PCR ( Figure 4D ) [26] , [27] . Transcript levels of CHT2 and OCH1 were significantly higher in rca1Δ in air when compared to the control strain ( Figure 4D ) . These data show that in addition to Rca1p's function as activator of CA expression in low CO2 , this regulator can also operate as a repressor . Remarkably , 46 of the 85 ( 54% ) Rca1p associated genes are presently uncharacterized ( Table S1 and S2 ) . Although this observation precludes assigning a significant enrichment of genes to any cellular function , process , component ( GO Term Finder , http://www . candidagenome . org/cgi-bin/GO/goTermMapper ) or protein families ( pfam ) , it suggests a broader involvement of Rca1p in CO2 sensing . A similar conclusion can be made following database searches with the TGACGTCA sequence involved in Cst6p binding which was retrieved in 49 promoters of S . cerevisiae genes . Analysis of both lists with GO Slim Mapper coupled to a chi-square test revealed a significant under-representation of genes in the process of RNA metabolic process ( P-value: 0 . 0066 ) in C . albicans as well as in the response to chemical stimulus process for both C . albicans and S . cerevisiae ( respectively P-value 0 . 0436 and 0 . 0322 ) while the latter was over-represented in the budding yeast contrary to C . albicans . However , it is important to note that the number of genes involved was relatively low ( respectively 2 , 4 and 6 ) . Altogether , these results show that , except for NCE103 , no apparent commonality of putative Rca1p targets or pathways can be identified and the large number of uncharacterized genes in the two lists of genes poses limitations to the full elucidation of the impact of these transcriptional factors on yeast cell biology . At the same time , these data could point to the existence of yet undiscovered pathways and underline the intrinsic differences between the two fungal organisms . In summary , our data establish Rca1p as the first regulator of a fungal CA and imply a wider role of this transcription factor in a new fungal CO2 sensing pathway . Since CA is critical for yeast growth in air ( Figure 1 ) , and its induction depends on Rca1p , it can be predicted that inactivation of RCA1 should also result in a growth deficiency . Indeed we observed that rca1Δ has a 77% increase of its generation time compared to the control strain ( Figure 5A ) . This phenotype is not restored in high CO2 pointing to a wider role of Rca1p in cell growth which could be set downstream of the CA . The enhanced growth rate of rca1Δ compared to nce103Δ is explained by residual expression of the highly effective carbonic anhydrase . We also confirmed that inactivation of RCA1 does not lead to significant morphological alterations ( Figure S6 ) . Our ChIP-chip data suggest a connection of Rca1p to filamentous growth and cell wall biogenesis , an observation that we confirmed by showing a strong decrease in the morphological response of rca1Δ to serum ( Figure 5B and S6 ) and an increased sensitivity of rca1Δ to Congo red , caffeine and SDS ( Figure 5C ) . These results set Rca1p as an important player of C . albicans key biological functions . In S . cerevisiae , we were not able to reach identical conclusions as inactivation of CST6 did not result in enhanced sensitivity to cell-wall perturbing agents . Additionally , only a 20% increase in generation time was observed for the cst6Δ mutant . Notably this phenotype was complemented by growing the strain in elevated CO2 ( Figure 5A ) . These data confirm that the orthologs of RCA1 are involved in the regulation of different cell functions further underlining their intrinsic difference emerging from the ChIP-chip and bioinformatic analysis . Interestingly expression levels of the Rca1p orthologues is also variable between the two species ( Figure 5D ) . Using specific primers for each species ( C . albicans and S . cerevisiae ) , we investigated the level of RCA1 and CST6 transcript in low and high CO2 environment . In C . albicans , RCA1 expression is 2 . 5 fold higher in hypercapnia compared to normal atmosphere ( Figure 5D ) . In contrast , the CST6 transcript in S . cerevisiae did not display any significant variation of the expression between the two conditions ( Figure 5D ) . While the function as a regulator of carbonic anhydrase is shared among Rca1p orthologs , their regulation in response to environmental CO2 differs . Sequence comparison of the Rca1p orthologs from C . albicans and S . cerevisiae identified three putative sites of phosphorylation ( Figure S4 ) . We investigated the role of these residues in the function of C . albicans Rca1p by complementation of rca1Δ with constructs expressing Rca1p with a replacement of serine to alanine in position 124 and 126 ( rca1Δ+RCA1−S124A and rca1Δ+RCA1−S126A respectively ) or serine to glycine in position 222 ( rca1Δ+RCA1−S222G ) . Loss of serine in position 126 or 222 only partially impact on the CO2 regulation of Nce103p expression; however mutating serine 124 lead to a striking unresponsiveness to ambient CO2 resulting in enhanced expression of Nce103p in both air and air enriched with 5 . 5% CO2 ( Figure 6 ) . Our results point to a critical role of serine 124 for Rca1p activity in response to CO2 concentrations . We have previously shown that in C . albicans colonies metabolically-generated CO2 accumulates and is subsequently used to activate the adenylyl cyclase Cyr1p promoting the switch from yeast to filamentous growth essential for pathology [13] . We now substantially expand these results to entire populations of S . cerevisiae taking advantage of the regulation of expression of ScNce103-GFPp by CO2 . Matching CA protein expression detected by Western blots ( Figure 2A ) , a strong fluorescent signal was recorded in ScNCE103-GFP cells grown in ambient air but absent in air enriched with 5 . 5% CO2 ( Figure 7A ) . Next we visualized Nce103p expression not only in individual cells but an entire fungal colony , monitoring for the first time the flux of CO2 in a fungal population . Using high resolution two-photon excitation confocal microscopy [28] we examined a cross-section of a ScNCE103-GFP colony grown for 4 days on solid nutrient agar . We observed that cells in the superficial layers , exposed to the low CO2 concentrations found in ambient air , strongly express the Nce103-GFPp construct; while the internal layers of the colony do not show any significant fluorescence ( Figure 7B ) . Strikingly when grown in a 5 . 5% CO2 atmosphere , this gradient was absent , and no fluorescence was observed at any position in the colony . Similarly , no fluorescence was seen in cst6Δ , regardless of the CO2 concentration or the position in the colony ( Figure 7B ) , indicating that the abscence of GFP expression in the center of the ScNCE103-GFP colony grown in air was unlikely due to a lack of viability or metabolic activity of the corresponding cells . By contrast , our positive control constitutively expressing GFP displays homogenous fluorescence through the cross-section ( Figure 7B ) . In conclusion , our data visualizing the flux of CO2 inside yeast populations are in full agreement with those generated by Western blot or qRT-PCR in single cells ( Figure 1C , D ) . Furthermore , they illustrate , with a high level of detail , the capacity of yeast to generate CO2 enriched micro-environments and adjust metabolic expression in a population .
Carbon dioxide is a major signal in all organisms ranging from humans to fungi [7] , [29] . CO2 regulates numerous phenotypes including virulence in the fungal pathogens of humans C . albicans or C . neoformans [7] , [12] . Here we demonstrate for the first time a novel , cAMP independent , CO2 sensing pathway in the yeast species C . albicans and S . cerevisiae . We report that at the core of this new sensing pathway lies a bZIP transcription factor , Rca1p in C . albicans and its ortholog Cst6p in S . cerevisiae . We show that Rca1p and its orthologs regulate the expression of a major enzyme involved in fungal metabolism , CA , in response to changes in ambient CO2 level . CAs catalyze the synthesis of HCO3− , an essential substrate for the cell's carboxylation reaction that sustains gluconeogenesis , ureagenesis or lipogenesis [30] , [31] . We hypothesize that CA is critically involved in cellular metabolism and a feedback loop involving Rca1p could exist to regulate its expression ( Figure 8 ) . Furthermore , as CA controls the level of HCO3− , the regulation of CA expression driven by cellular metabolism could have an indirect impact on the capacity of the cells to differentiate through activation of the cAMP-PKA pathway . While HCO3− is an essential cofactor for cellular metabolism in all fungi tested , the fungal requirement for CA is conditional , depending on the environmental CO2 concentration . CA mutants will not grow in ambient air where CO2 is scarce but will thrive in niches where the atmosphere is enriched with this gas [6] , [7] , [8] , [9] , [10]; the higher concentration allows sufficient spontaneous hydration to HCO3− to serve as substrate for the above carboxylation reactions . Regulation of CA expression by CO2 has been reported for S . cerevisiae , S . macrospora , A . fumigatus and A . nidulans [8] , [9] , [17] , [18] and we extend these observations to C . albicans . Though the regulation of CA by CO2 has been observed in many fungi , the current work is the first to report the identification of a fungal-specific CO2-responsive transcription factor , Rca1p in C . albicans and of one of its orthologs in S . cerevisiae . We identified Rca1p , a previously uncharacterized bZIP-family DNA binding protein , via a functional genetic screen of a transcription factor knockout library . Rca1p functions by inducing CA protein and transcript when C . albicans faces low ambient CO2 level . Loss of CA induction in high CO2 level could result from a phosphorylation or another posttranslational modification on serine 124 which leads to the inability of Rca1p to bind NCE103 promoter , integrating Rca1p in a signal transduction pathway . The impact of Rca1p in cell growth and cell wall biogenesis , independently of CO2 concentration , points to a general involvement of Rca1p in the cellular metabolism of C . albicans . Rca1p function as CA regulator is conserved in S . cerevisiae , though they also have additional functions: Cst6p has already been shown to be involved in functions such as growth on non-optimal carbon sources [23] , and our results now highlight the importance of Cst6p in cellular metabolism via its role as an inducer of CA . However , the impact of Cst6p on cell physiology differs when compared to C . albicans since the cts6Δ mutant growth defect in air is complemented by addition of environmental CO2 for S . cerevisiae . Furthermore the observation that RCA1 expression itself is regulated by CO2 underlines the importance of this regulator outside the scope of CA expression . Notably , Rca1p orthologs can be identified in S . macrospora , A . fumigatus and A . nidulans known to posses CA's which expression is influenced by ambient CO2 level [8] , [9] . Importantly , Rca1p is distinct at both the sequence and functional level from the best characterized regulator of eukaryotic CAs , HIF-1α , which induces human CA IX expression in response to hypoxia [15] , [32] . CA IX leads to extracellular acidification of hypoxic tissue , and is as such abundant in tumors [33] . By contrast C . albicans NCE103 , is not regulated by either hypoxia or changes in pH ( [34]; Figure S7 ) . Similar to HIF-1α , which binds to the HRE motif present in CA IX promoter [32] , the regulation of NCE103 by Rca1p appears to be direct . ChIP-chip and ChIP-qPCR confirmed that Rca1p binds to the CA promoter of C . albicans , specifically under low ambient CO2 , leading to the induction of NCE103 expression . An additional 84 genes were associated with Rca1p , suggesting a much broader role of this new CO2 signaling pathway . Although the majority of these genes are currently of unknown function , in depth analysis of two ( OCH1 and CHT2 ) showed Rca1p's potential to act as both an inducer and repressor of gene expression . This dual function could also be explained by an ability of Rca1p to recruit different co-factors at the associated loci . Using SCOPE [35] and other predictive programs , we obtained a relatively low number of results with significant value regarding binding sites or processes associated to Rca1p and Cst6p . These observations could be due to a higher complexity in the binding motif of Rca1p to the DNA , as well as on our limited knowledge about genes function ( 75% of ORF in C . albicans are still considered as uncharacterized; http://www . candidagenome . org ) . Reflecting the 235 million years [36] of separation between the Candida and Saccharomyces clade , we observed profound divergence in the associated genes of Rca1p and Cst6p . However , the CA remains a conserved target for both species . Such a fundamental wiring re-arrangement between closely related transcriptional factor of C . albicans and S . cerevisiae has already been reported [37] . In S . cerevisiae , CO2 regulation of CA expression by Cst6p involves the TGACGTCA palindrome motif . Database searches with this sequence retrieved the motif in 49 promoter regions of S . cerevisiae , and 40 promoters of C . albicans . While this motif is absent in the NCE103 promoter in C . albicans , it is present in a single gene associated to Rca1p ( orf19 . 4246 ) demonstrating the lack of motifs conservation between the two species . Interestingly , the promoter of human CA IX presents a bZIP binding motif , TGAGTCA [38] , which is closely related to the one identified in S . cerevisiae . Furthermore this motif is the binding sequence of the oncogen C-jun in human , which presents some sequence similarity with Rca1p ( Score 42 . 7 , e value 0 . 003 ) , particularly around the bZIP domain . To date , the expression of CA IX in response to CO2 changes in the body has not been investigated . Our results may point to an additional level of regulation in human CA's; in fact HIF-1α sole predominance in CA IX regulation has recently been questioned [39] . The regulation of CA by CO2 is likely to be complex; however , using a new antibody that we generated , we show that CO2 affects CA proteins levels dramatically – highly induced in normal and undetectable when cells are grown in an elevated CO2 atmosphere . However , CA transcript levels were only decreased by 50% relative to ambient air . This type of regulation may suggest additional levels of post-transcriptional control on CA messenger . Maintaining CA transcripts in high CO2 level would allow a shortened response in CA enzyme synthesis when cells encounter a switch from high to low CO2 atmosphere , thus ensuring sufficient supply of the essential HCO3− ion . We have begun to uncover the complex physiology associated with variations in CO2 concentrations . C . albicans cells phagocytosed by macrophages induce CA , as seen by qRT-PCR , despite being in a high CO2 ( 5% ) environment , and this is Rca1p-dependent . This suggests that the phagocyte might restrict CO2 availability , as it does for other nutrients . Furthermore , using high resolution two-photon excitation confocal microscopy with GFP-tagged CA's we visualize for the first time the impact of CO2 build-up on gene expression in single cells but also in entire fungal populations . The data presented in this report not only confirm previous observations made in C . albicans that CO2 is compartmentalized in yeast populations , specifically inducing developmental change [13] , but importantly connects micro-environments enriched in CO2 to metabolic specialization of individual members of a fungal populations . However , it is important to also consider the possibility that an unknown , CO2-independent , pathway is involved in the regulation of Nce103p at colony level . CA regulation is exquisitely sensitive to change in CO2 availability in both S . cerevisiae and C . albicans ( Figure S8 ) and our results may be highly applicable to a range of conditions in which fungi expand and act as populations rather than individual cells including the formation of drug-resistant biofilms in pathogenic yeasts such as C . albicans . For S . cerevisiae , biofilms are naturally isolated on fruit surface ( grape ) , and have a major application in industrial fermentation [40] , [41] . The effects of CO2 on fungal physiology are integrated through more than one regulatory circuit . We previously showed that elevated CO2/HCO3− is sensed by the adenylyl cyclase via a lysine residue of the enzymes catalytic core , increasing the production of the second messenger cAMP , thus linking adenylyl cyclase and CA in fungal CO2 sensing [7] , [12] . Adenylyl cyclase/cAMP are particularly important mediators of fungal virulence determinants . Although there is cross-talk between the activities of the two enzymes , we now show that the CO2 control of CA expression acts independently from the cAMP-PKA pathway . This mechanism has to be compared to the identification of a cAMP-independent CO2-sensitive pathway involved in white opaque switching which results in Wor1p phosphorylation [42] . However , the putative overlap of these two uncharacterized pathways remains to be defined . In conclusion , carbon dioxide is sensed in yeast by two independent pathways . One , previously described by us , involves the fungal adenylyl cyclase and cAMP [12] . We have now identified a second pathway and found the transcriptional regulator , Rca1p , at the core of the new pathway in yeast . Activation or inactivation of the transcription factor may involve phosphorylation which ultimately programs cellular metabolism to allow optimal adaptation to the environment inside a macrophage in the case of a fungal pathogen or yeast populations . Investigating the function of the additional Rca1p-associated genes will bring better understanding on how organisms sense the universal gas carbon dioxide .
All strains and plasmids used or constructed in this study are reported in supporting information ( Table S3 , S4 , S5 ) , as is the composition of the respective media and additional protocols ( Protocol S1 ) . C . albicans were incubated at 37°C , or 30°C for S . cerevisiae , either in ambient air or air enriched with 5 . 5% ( vol/vol ) atmospheric CO2 ( Infors HT Minitron ) when required . Strains were inoculated in 50mL of YPD at OD600 0 . 1 and grown at the suitable temperature in air or air enriched with CO2 . After 4h , cells were collected and quickly frozen . Samples were disrupted using a Mikro-dismembrator S ( Sartorius ) and resuspended in 500 µl lysis buffer ( 50 mM HEPES , 150 mM NaCl , 5 mM EDTA , 1% Triton X-100 , protease inhibitor [Roche] ) . Protein concentrations were quantified using Bradford reagent ( Sigma ) . 30 µg of protein were loaded for each sample on a 12% SDS-acrylamide/bis-acrylamide gel , and proteins were transferred to a PVDF membrane ( Millipore ) . Membranes were incubated with the appropriate antibodies diluted as follows: anti-NCE103 at 1∶500 , anti-LacZ at 1∶1000 ( Millipore ) , anti-GFP at 1∶2500 ( Roche ) or anti-Act1p 1∶1000 ( Sigma ) . This was followed by a second incubation with a peroxidase tagged antibodies of goat anti-rabbit , diluted 1∶2000 , ( Sigma ) for the anti-NCE103 , anti-LacZ and anti-Act1 primary antibodies , while a goat anti-mouse , diluted 1∶5000 ( Sigma ) , was used for the anti-GFP primary antibody . Luminol Electrochemiluminescence was used to detect signal on the membrane . Culture and samples were prepared in an identical manner as for the protein extraction , apart from total RNA extraction which was carried out with the RNeasy Kit ( Qiagen ) according to the manufacturer's recommendations . Transcripts level were determined by semi quantitative RT-PCR using the iScript One-Step RT-PCR Kit with Syber Green ( BioRad ) . Levels were normalised to ACT1 from the respective species , and calculated using the Delta C ( t ) analysis of the Opticon Monitor 3 software ( Bio-Rad ) . Values are represented as mean +/− SD from three independent experiments . 50ml cultures in YPD of strains RCA1/rca1Δ+RCA1 ( untagged ) and RCA1/rca1Δ+RCA1−HA3 ( tagged ) were inoculated at OD600 0 . 1 by overnight culture and incubated 4h at 37°C in air or air enriched with 5 . 5% CO2 , 140 rpm . Three independents cultures were grown for each strain in both conditions . The subsequent steps of DNA cross-linking , DNA shearing , chromatin immunoprecipitation ( ChIP ) , DNA labeling with Cy3 or Cy5 dyes , hybridization to intergenic DNA microarrays , and data analysis were conducted exactly as described [42] . Cy5-labeled DNA from the tagged strain ( RCA1/rca1Δ+RCA1−HA3 ) and the corresponding Cy3-labeled DNA from the untagged control strain ( RCA1/rca1Δ+RCA1 ) were mixed and hybridized to a C . albicans whole-genome tiled oligonucleotide DNA microarray [24] . After hybridization and scanning of the slides ( n = 3 for each condition ) , results were process [43] . Quantile normalization was applied to the data [25] . The parameters used were: a window size of 400 bp , a maximum genomic distance of 60 bp , and a minimum length of 120 bp . The replicate data were combined , and peak finding ( i . e . , determining the Rca1-HA3p binding sites ) was done using a pseudomedian signal threshold of at least 1 . 5 fold and a P value cutoff of 0 . 01 [25] , [44] . Chromatin Immuno Precipitations were processed according to the above protocol with the identical strains RCA1/rca1Δ+RCA1 ( untagged ) and RCA1/rca1Δ+RCA1−HA3 ( tagged ) and growth conditions . The resulting purified DNA was used in quantitative PCR using SYBR Green Master Mix ( Applied Biosystems , Inc . ) with primers: Ca-ChIP-NCE103-F/Ca-ChIP-NCE103-R for the detection of the NCE103 promoter ( a 195bp region identified to be significantly associated with Rca1p ) and Ca-ChIP-ACT1-F/Ca-ChIP-ACT1-R for the control ACT1 promoter , a gene without known association for Rca1p . Levels of detection were normalized to ACT1 and calculated using the Delta Delta C ( t ) method . Values are represented as mean +/− SD from two independent experiments . C . albicans wild-type ( SC5314 ) and rca1Δ were grown to log-phase in YPD medium , washed in water , and counted . They were then incubated with J774A . 1 macrophages at an MOI of 2∶1 ( C . albicans: macrophages ) in RPMI+10% FBS at 37°C in 5% CO2 in 750 cm3 vented flasks . Control cells were grown in the same media without macrophages at 37°C in 5% CO2 . After incubation for one hour , the flasks were rinsed with PBS , then cells were collected by scraping into ice cold water and transferred to conical tubes . They were washed twice more with water , then pellets were frozen on dry ice . RNA was prepared using the Turbo DNA-free kit ( Ambion ) . 50 ng of total RNA were used for each qRT-PCR reaction using the Power SYBR Green reaction system ( Invitrogen ) . Actin ( ACT1 ) was used as the normalization control . Primers are listed in supplemental data . ScNCE103-GFP cells were observed with a Olympus IX-81 fluorescence microscope with a 150 W xenon-mercury lamp and an Olympus 60X Plan NeoFluor oil-immersion objective . For high resolution two-photon excitation confocal microscopy of entire yeast colonies of ScNCE103-GFP , ScNCE103−GFP+cst6Δ and BY4741+pTEF−GFP , cells were grown for 4 days on YPD at 28°C . Colonies were then embedded in low-gelling agarose ( Sigma-Aldrich ) directly on the plates [28] . After solidification , agarose-embedded colonies ( an area of approximately 10×10 mm ) were sectioned vertically down the middle and transferred to the cover glass . All samples on the cover slip were enclosed with a thick agarose layer to prevent them from drying . Image acquisition was realized following published protocol [28] , using 20 x/0 . 7 water immersion planachromat objective . Statistical analyses were performed using Student's t test . P values are indicated as detailed in the figure legends . Error bars in figures represent SD .
|
Skin infection , oral and vaginal thrush , or bloodstream candidiasis are some of the diseases caused by the human pathogen Candida albicans . The high versatility of infection niches reflects the capacity of this yeast to respond to strong variations in its environment such as CO2 concentration . This molecule initiates the regulation of an essential protein: carbonic anhydrase , not through the known adenylyl cyclase CO2 sensor but as we discovered via a novel fungal CO2 sensing pathway involving the transcriptional regulator Rca1p . This protein is additionally implicated in growth , yeast-to-hyphae morphological switch and cell wall stability of C . albicans . The ortholog of Rca1p in Saccharomyces cerevisiae demonstrated a conserved function in the induction of the carbonic anhydrase in low CO2 concentration atmospheres pointing to the broad significance of Rca1p in fungal CO2 sensing .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"biology"
] |
2012
|
The bZIP Transcription Factor Rca1p Is a Central Regulator of a Novel CO2 Sensing Pathway in Yeast
|
Schistosomiasis has decreased significantly in prevalence and intensity of infection in China , thus more accurate and sensitive methods are desperately needed for the further control of schistosomiasis . The present work aimed to assess the utility of the loop-mediated isothermal amplification ( LAMP ) for detection of light intensity infection or false-negative patients and patients post-treatment , targeting the highly repetitive retrotransposon SjR2 of Schistosoma japonicum . LAMP was first assessed in rabbits with low intensity infection ( EPG<10 ) . Then 110 patient sera from Hunan Province , China , and 47 sera after treatment by praziquantel were used to evaluate the diagnostic validity of LAMP . Meanwhile , 42 sera from healthy individuals in a non-endemic area , and 60 sera from "healthy” residents who were identified as being negative for feces examination and immuno-methods in an endemic area were also examined . The results showed that LAMP could detect S . japonicum DNA in sera from rabbits at 3rd day post-infection . Following administration of praziquantel , the S . japonicum DNA in rabbit sera became negative at 10 weeks post-treatment . Of 110 sera from patients , LAMP showed 95 . 5% sensitivity , and even for 41 patients with less than 10 EPG , the sensitivity of LAMP still reached to 95 . 1% . For 47 patients after treatment , the negative conversion rate of S . japonicum DNA in patient sera increased from 23 . 4% , 61 . 7% to 83 . 0% at 3 months , 6 months and 9 months post-treatment , respectively . No false-positive result was obtained for 42 human sera from non-endemic area , while for the 60 “healthy” individuals from endemic area , 10 ( 16 . 7% ) individuals were positive by LAMP , which suggested that these individuals might be false-negative patients . The present study demonstrated that the LAMP assay is sensitive , specific , and affordable , which would help reduce schistosomiasis transmission through targeted treatment of individuals , particularly for those with negative stool examinations who may yet remain infected . The LAMP assay may provide a potential tool to support schistosomiasis control and elimination strategies .
Schistosomiasis remains one of the most important chronic parasitic diseases in tropical regions and affects approximately 200 million people , despite the continued implementation of control measures [1] . Schistosomiasis japonica , caused by infection with Schistosoma japonicum , is mainly endemic in China . In areas where infection occurs , the prevalence and infection intensity is now low due to long-term and large-scale chemotherapy campaigns [2] . Methods that allow infections to be correctly diagnosed are a prerequisite for effective disease control . All present schistosomiasis control measures , including targeted treatment of all infected individuals , especially those with low-intensity infections , and large-scale surveillance of disease transmission are strongly dependent on sensitive and accurate diagnostic assays [3] . Indeed , the move to identify and evaluate highly sensitive diagnostics has become increasingly necessary as the prevalence and intensity of schistosome infections continues to decline worldwide [4] . The Kato-Katz fecal smear technique is the most commonly used method to diagnose schistosomiasis . However , various studies have shown that its sensitivity is less appropriate for low endemic areas , post-treatment situations , and for determination of incidence [5 , 6 , 7] . Moreover , traditional stool examinations always underestimate the prevalence of schistosomiasis due to their low diagnostic sensitivity [8] . Antibody detection methods generally have high sensitivities , but the slow reduction of specific antibody levels after treatment and the inability to discriminate between active and past S . japonicum infection constituted great disadvantages of the antibody-based assays [9 , 10] . Moreover , the level of antibodies persisted after treatment , which could not indicate whether the patients were cured or not [11 , 12] . In recent years , detection of circulating antigens , such as circulating anodic antigens ( CAA ) in serum or urine , seems a promising tool for diagnosis of Schistosoma infection [7 , 13] . A field study done by van Dam et al . [7] , using lateral-flow assay for determination of CAA in urine and serum , showed that the method was at least 6 times more sensitive than triplicate Kato-Katz thick smears . However , due to the lack of “golden” standard reference test , different diagnostic methods are not comparable , thus a more sensitive standard is still needed [14] . With the development of molecular techniques , polymerase chain reaction ( PCR ) -based methods have shown great sensitivity and specificity for detection of Schistosoma DNA in a variety of samples [9 , 10 , 15–23] . However , the dependence on expensive apparatus and on specialized training in molecular biology restricts their widespread applications for field conditions . As an alternative , the loop-mediated isothermal amplification ( LAMP ) may provide a potential tool for diagnosis of schistosomiasis . The LAMP assay , firstly reported by Notomi et al . [24] , has been rapidly accepted for detection of various pathogens , including Plasmodium falciparum , S . mansoni , S . haematobium and S . japonicum , due to its simplicity and rapidity [25–32] . This technique does not require sophisticated equipment for DNA amplification or for amplicon detection [33] , which is of great value for field use . Our previous study established a LAMP assay based on the sequence of highly repetitive retrotransposon SjR2 , which is able to detect 0 . 08 fg S . japonicum DNA , and 104 times more sensitive than conventional PCR [30] . To extend from our previous work , in this study , we assessed the utility of this LAMP assay for detection of light infection or false-negative patients of schistosomiasis and evaluation of chemotherapy efficacy , with the aim of providing a potential tool to support schistosomiasis control and elimination strategies .
This study was funded by A Project Funded by the Priority Academic Program Development of Jiangsu Higher Education Institution ( No . YX13400214 ) ; the National Basic Research Program of China ( 973 Program , Grant No . 2007CB513100 ) ; Key Laboratory of Control and Prevention of Parasitic Disease of Healthy Ministry , No . wk014-001 ) . The funders had no role in study design , data collection and analysis , decision to publish , or preparation of the manuscript . This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The protocol was approved by the Committee on the Ethics of Animal Experiments of the Soochow University ( Permit Number: 2007–13 ) . All surgery was performed under sodium pentobarbital anesthesia , and all efforts were made to minimize suffering . Schistosoma japonicum-infected snails ( Oncomelania hupensis ) were obtained from Jiangsu Institute of Parasitic Diseases , China . The living snails were putting into a breaker filled with 4/5 volume of water and then exposed to a light source to induce shedding of live S . japonicum cercariae . Eight adult female New Zealand rabbits , weighing 2 . 0–2 . 5 kg , were randomly divided into four groups of two rabbits each . Each rabbit in Group 1 was percutaneously infected with 30 mixed sexual cercariae . On the 7th week post-infection , all the rabbits were sacrificed , and adult worms ( male and female ) were collected to confirm the real intensity of infection . Rabbits of group 2 and group 3 were given the same infective intensity as group1 . Rabbits of group 4 administrated with saline were used as uninfected control . The faecal samples of rabbits for Kato-Katz egg examination were collected weekly from 1 week post-infection to 30 weeks post-infection . Rabbits in group 2 were treated with a single dose of 150 mg praziquantel per kilogram body weight on the 7th week post-infection . This dose was known to eliminate S . japonicum in our model [9 , 30] . Rabbits of group 3 remained as untreated control , and received no drug treatment . Blood samples from rabbits of group 2 to group 4 were collected on the 3rd day and then weekly until 30 weeks post-infection ( Fig 1 ) . Serum of each blood samples was separated by centrifugation ( 2 , 000 rpm for 10 min ) after storage at 37°C for 2 h . All of the serum samples were stored at -20°C until use . On the 30th week post-infection , all of the rabbits were sacrificed , and the portal system and liver of the rabbits in group 2 were examined for evaluation of therapy efficacy . One hundred and ten serum samples positive for S . japonicum infection determined by triplicate Kato-Katz stool examinations were obtained from 30 females and 80 males living in endemic areas in Hunan Province , China . Written consent letters were obtained from all participants . Ethical approval was obtained from the village ( local government ) , county ( anti-schistosomiasis office ) and provincial ( schistosomiasis headquarters ) authorities . Five mL blood sample was collected from each participant , and about 2 mL serum was obtained from each blood sample . Then the serum samples were divided into 3 groups according to the eggs per gram of feces ( EPG ) : 93 samples with <100 EPG , 13 samples with 100–400 EPG , 4 samples with >400 EPG . According to the new classification in China , which sets 100 EPG as the lower limit for a heavy infection; 40–99 EPG denotes a moderate infection , while 40 EPG is the upper limit for a light infection [34] , the 93 samples with less than 100 EPG was divided again: 10 samples with 40–99 EPG , 42 samples with 10–39 EPG , 41 samples with < 10 EPG ( Table 1 ) . All the 110 cases were treated with praziquantel ( single oral dose of 40 mg/kg ) , two stool samples were obtained from each participant at the 3rd month , the 6th month and the 9th month post-treatment , respectively , and each sample was subjected to triplicate Kato-Katz thick smears to evaluate the effectiveness of chemotherapy . Meanwhile , 2 mL serum samples were collected from the study population accordingly . However , only 47 serum samples from the participants had the complete data records . Forty-two serum samples from uninfected individuals collected from Wuxi , Jiangsu Province , China were used as negative control to evaluate the specificity of the LAMP assay . Meanwhile , 60 serum samples from individuals living in fishing villages in endemic areas , which were negative for fecal eggs by triplicate Kato-Katz thick smears , and also negative for serology diagnosis by triplicate IHA and ELISA tests , were used to evaluate the validity of the LAMP assay . DNA from all of the collected serum samples was extracted using the method described previously [9] . Briefly , 200 μl sera were dissolved in 400 μl serum lysis buffer , incubated 20 min at 55°C , and then extracted with phenol–chloroform–isoamyl alcohol ( 25:24:1 ) and precipitated with dehydrated alcohol . Serum samples were tested for anti-Schistosoma antibodies by IgG-ELISA . Serology was performed according to the protocol described by Xia et al . [9] . Sera were considered positive when the OD value exceeded the mean±3 SD absorbance of sera from non-infected samples . Indirect hemagglutination kit containing human erythrocytes coated with soluble egg antigen is commercially available from the Anhui Provincial Institute of Parasitic Diseases ( Wuhu , China ) . The test procedure followed a previous study [2] . The test result was considered positive when a positive reaction appeared at a titer 1:10 . The LAMP assay was based on a previously study targeting the highly repetitive retrotransposon SjR2 of S . japonicum ( GeneBank Accession no . AF412221 ) [9 , 30 , 35] , with slight modifications . The LAMP assay was carried out with a total of 25 μl reaction mixture containing 2 . 5 μl 10×Bst-DNA polymerase buffer , 6 mmol/L MgSO4 , 1 . 4 mmol/L dNTP , 0 . 2mmol/L F3 , B3 , 1 . 6 mmol/L FIP , BIP , 0 . 8 mol/L betain , 8 U Bst-DNA polymerase , and 5 μl template DNA . The reaction mixture was incubated at 64°C for 90 min . The LAMP amplification results were identified by adding 5 μl 1:80 diluted 10000×SYBR Green I after incubation , positive reactions were detected by an orange to green color change visible under normal light . Diagnostic sensitivity , specificity , PPV and NPV were calculated using the following formulae: Sensitivity = True positives/ ( True positives+False negatives ) Specificity = True negatives/ ( True negatives+False positives ) Positive predictive value ( PPV ) = True positives/ ( True positives+False positives ) Negative predictive value ( NPV ) = True negatives/ ( True negatives+False negatives ) The SPSS 19 . 0 software was used for data analysis .
On average , 3 female and 10 male adult worms ( 3 pairs of worms ) were collected in rabbits of Group 1 , and the mean EPG of the rabbits infected with 30 cercariae was 16 , which is very low intensity infection . In addition , for fecal examinations of the rabbits , the eggs could not be found untill 7 weeks post-infection from group 1 to group 3 . No adult worms were found in the portal system and liver of the rabbits in group 2 , indicating thoroughly treatment of the rabbits in group 2 . While for the rabbits without praziquantel treatment , 3 female and 18 male adult worms were found , and egg-granulomas were clearly observed ( Table 2 ) . ELISA and IHA examination of rabbit serum samples gave positive results at 5 weeks and 4 weeks post-infection , respectively , and the antibody sustained at high level even at 23 weeks post-treatment ( 30 weeks post-infection , Table 2 ) . Whereas the S . japonicum DNA was detectable by LAMP at the 3rd day post-infection in serum of rabbit model with very low intensity infection ( EPG = 16 , Table 2 , Fig 2 ) . Following administration of praziquantel , the detection of S . japonicum DNA in rabbit sera became negative at 10 weeks post-treatment ( 17 weeks post-infection , Table 2 , Fig 3 ) . Of the 110 patient serum samples with confirmed S . japonicum infection by stool examination , the sensitivity and specificity of LAMP assay was 95 . 5% and 100% , respectively . Whereas , the sensitivity and specificity of ELISA and IHA was 84 . 6% and 85 . 7% , and 91 . 8% and 88 . 1% , respectively . There was significant difference between the sensitivity of LAMP and ELISA ( χ2 = 7 . 273 , P = 0 . 007 ) . Table 3 shows the PPV and NPV of the three detection methods , and the LAMP assay had the highest PPV ( 100% ) and NPV ( 89 . 4% ) . Although all the three methods showed 100% sensitivity for patients with high-intensity infections ( EPG>400 ) , LAMP assay had higher sensitivity ( 95 . 7% ) than ELISA ( 83 . 9% ) and IHA ( 93 . 4% ) for patients with <100 EPG ( χ2 = 7 . 627 , P = 0 . 022 , Table 4 ) . For 42 patients with 10–39 EPG , the detection rate of LAMP achieved 97 . 6% , which was significantly higher than the two immunoassays ( χ2 = 6 . 157 , P = 0 . 046 , Table 5 ) . Even for 41 patients excreting <10 EPG , which was very low-intensity infection , the sensitivity of LAMP assay sustained at a high level ( 95 . 1% ) , and only 2 cases were missed ( Table 5 ) . After treatment with praziquantel , all the stool samples were negative for faecal eggs by triplicate Kato-Katz method . However , the negative conversion rate of S . japonicum specific SjR2 DNA detected by LAMP assay increased remarkably from 23 . 4% , 61 . 7% to 83 . 0% at 3 months , 6 months and 9 months post-treatment , while for the two immunoassays , the negative conversion rate of antibodies sustained at a low level even after 9 months post-treatment ( Table 6 ) . Statistical analysis showed that there was significant difference between LAMP and the two immunodiagnostics for the negative conversion rate of S . japonica at 6 months and 9 months post-treatment ( χ2 = 17 . 63 and 37 . 43 , both P value <0 . 0001 ) . Additionally , 42 serum samples of individuals from non-endemic areas , and 60 serum samples of residents in endemic areas who were identified with negative results for all the three commonly used diagnostic methods , including triplicate stool examinations , IHA and ELISA tests , were used to assess the validity of LAMP assay for field diagnosis of schistosomiasis . No positive result was observed for 42 uninfected human serum samples , indicating an excellent specificity of LAMP assay ( Table 7 ) . However , for the 60 serum samples from residents in endemic areas , which were negative for faecal eggs and serological methods , 10 ( 16 . 7% ) samples were positive diagnosed by LAMP assay ( Table 7 ) , which might be missed by the commonly used diagnostic methods , such as Kato-Katz method , IHA and ELISA .
Diagnosis is central to control of schistosomiasis [36] . The prevention and control of the disease need rapid and reliable diagnostic techniques to identify target population accurately for treatment [37] . However , the currently available diagnostic assays are not ideal , since the search for eggs in stools and detection of circulating antigens lack sensitivity in low prevalence and post-treatment situations , and antibody detection lacks specificity [38] , and cannot distinguish current and cured infections , which results in the difficulties in determining prevalence , identifying true infected individuals for selective chemotherapy and assessing the effectiveness of intervention including follow-up of chemotherapy [34 , 39 , 40 , ] . Our results further demonstrated the limitations of the direct parasitological technique ( Kato-Katz method ) and antibody detection assays ( ELISA and IHA ) . In rabbit models infected with S . japonicum cercariea , we did not find faecal eggs until 7 weeks post-infection , and for the two immunoassays , the earliest positive detection result was obtained at 4 weeks post-infection . Our rabbit experiment results confirmed that both the stool examination and antibody detection methods could not give early diagnosis at the beginning of S . japonicum infection . Moreover , 4 egg-positive cases were negative for antibodies by ELISA and IHA , with ages ranging from 43 to 65 , indicating that antibody detections were insufficient for diagnosis of schistosomiasis in patients of older ages , due to immune down-regulation in chronic infection stages [7 , 41] . However , DNA amplification assays , which had identical diagnostic value with that of parasitological methods , may provide alternative approaches for sensitive and specific diagnosis of schistosomiasis . Lier et al . [21] reported a real-time PCR assay for detection of low intensity S . japonicum infections in a pig model . Subsequently , this method was used in a clinical trial . This real-time PCR assay detected slightly more positive faecal samples than the microscopy method , but was consistently negative in serum and urine samples [22] . To our knowledge , most of the PCR-based methods were focused on the detection of specific Schistosoma DNA in faeces , all of these methods could be seen as improvements on the stool examination approach [7] , thus the sensitivity was influenced by the large day-to-day egg fluctuations in infected individuals [42] . In general , serum and urine samples are easier to obtain and more accepted in many populations than faecal samples , and unlike eggs in faecal samples , schistosome DNA would be equally distributed throughout the serum of the patient , resolving the restrictions of uneven distributions of eggs in stool samples [43] . Therefore , the detection of Schistosoma DNA from serum samples would be more accurate in field conditions . In our previous study , we found that the S . japonicum DNA in host serum primarily comes from the residual body of dead schistosomula in the first 4 weeks post-infection , while during the spawning stage of the female schistosome , the parasite DNA mainly comes from the disintegration of inactive eggs [44] . Furthermore , a 230-bp sequence from the highly repetitive retrotransposon SjR2 was identified in our previous study , and showed high sensitivity and specificity for detecting S . japonicum DNA in sera of rabbit model and patients [9] . Although PCR-based methods have the potential for sensitive and specific detection of schistosomiasis , the dependence on expensive apparatus restricts their wide application in the field . Unlike PCR , LAMP assay does not require amplification cycles by thermocycling or amplicon detection by electrophoresis . Given these features , LAMP is potentially useful for work in the field and has already used in rural laboratories in developing areas for the diagnosis of tuberculosis [45] . Our previous study established a LAMP assay targeting S . japonicum SjR2 , and the method was capable of detecting as little as 0 . 08 fg S . japonicum DNA , which was 104 times more sensitive than common PCR assay . In particular , the LAMP assay was able to detect S . japonicum DNA in rabbit sera at 1 week post-infection , and become negative at 12 weeks post-treatment [30] . However , the rabbit models used in our previous study were of high and moderate intensity infections , and the detection of S . japonicum DNA in low-intensity infections is more consistent with the current epidemiological situation . In this study , the utility of LAMP assay was firstly assessed in rabbit models with very low grade intensities of infection ( EPG = 16 ) . It was able to detect S . japonicum DNA in serum at 3 days post-infection , and the detection results became negative at 10 weeks post-treatment , indicating that the LAMP method was useful for diagnosis of schistosomiasis , especially with low-intensity infection , and had potential for evaluation of chemotherapy effectiveness . Then the field diagnostic value of the LAMP method , and its ability for evaluation of effectiveness of drug treatments was tested using 110 patient serum samples with confirmed S . japonicum infection by stool examination . Meanwhile , two of the most extensively used immunoassays ( ELISA and IHA ) in the field were also used to assess the validity of the LAMP assay for diagnosis of schistosomiasis . Our detection results showed that the LAMP assay performed better than the commonly used immunoassays in terms of higher sensitivity in patients with low-intensity infection ( Table 4 and Table 5 ) . After treatment with praziquantel , the negative seroconversion rate of IHA and ELISA sustained at low levels , while for the LAMP assay , the negative conversion rate of S . japonicum DNA in serum increased from 23 . 4% , 61 . 7% to 83 . 0% after 3 months , 6 months and 9months post-treatment ( Table 6 ) . All of the results confirmed that the LAMP assay was efficient for diagnosis of cases with low-intensity infections , and had potential for assessment of effectiveness of drug treatment . Finally , 60 residents living in endemic areas with negative detection results of Kato-Katz , IHA and ELISA , were recognized as “healthy” residents , and were employed to assess the ability of LAMP assay for accurate diagnosis of schistosomiasis . Of the 60 serum samples from “healthy” individuals , 10 ( 16 . 7% ) were diagnosed as positive by LAMP assay ( Table 7 ) , who might had been missed by parasitological methods , indicating that traditionally used methods lack sensitivity for diagnosis of individuals with low intensity infections . A field study done by Xu et al . [3] further confirmed our results . In this study , of 1371 enrolled residents , parasitological detection identified only 74 ( 5% ) individuals as being egg-positive by Kato-Katz thick smears , of whom all the individuals were also positively diagnosed by LAMP detection of SjR2 DNA . More importantly , additional 368 ( 27% ) individuals were positive for SjR2 DNA [3] . Professor Clive Shiff of Johns Hopkins Bloomberg School of Public Health , USA , commented on this paper as “New diagnostics reform infectious parasite epidemiology” [12] . The comment suggested that continuous surveillance to predict any resurgence of infection by accurate and sensitive measuring methods is highly recommended . It is very important to help reduce schistosomiasis transmission through targeted treatment of individuals , particularly those people who are presumed to be free of infection ( false-negative ) may actually remain infected and capable of infecting snails when their faeces get into the water . In conclusion , the LAMP assay with rapidity , simplicity , sensitivity and specificity is suitable not only for case detections , but also for disease surveillance in schistosomiasis-endemic areas . Application of this method may improve the identification of cases with low-intensity infections and targeted treatment , which is of great significance for schistosomiasis control and elimination programmes .
|
Accurate diagnostic tests play a key role in patient management and control of schistosomiasis , especially in China where the prevalence and intensity of Schistosoma japonicum infection is low in recent years . The present study aimed to assess the utility of the loop-mediated isothermal amplification ( LAMP ) assay for detection of light intensity infection or false-negative patients and for the evaluation of chemotherapy effectiveness in patients , targeting the highly repetitive retrotransposon SjR2 of S . japonicum , using 110 serum samples of schistosomiasis patients . The results showed that the LAMP assay had high sensitivity of 95 . 1% for the diagnosis of S . japonicum infection with the lowest intensity ( EPG<10 ) . For the assessment of efficacy after treatment with praziquantel , the negative conversion rate increased from 23 . 4% , 61 . 7% to 83 . 0% at 3 months , 6 months and 9 months post-treatment by LAMP , whereas for ELISA and IHA , the negative conversion rate remains at a low level ( 25 . 5% by ELISA and 31 . 9% by IHA ) even at 9 months after treatment . Our results demonstrated that the LAMP assay may provide a valuable tool for the diagnosis of Schistosoma infection , especially for cases of light infection which is coincident with the current endemic situation .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
DNA Detection of Schistosoma japonicum: Diagnostic Validity of a LAMP Assay for Low-Intensity Infection and Effects of Chemotherapy in Humans
|
Sensing extracellular changes initiates signal transduction and is the first stage of cellular decision-making . Yet relatively little is known about why one form of sensing biochemistry has been selected over another . To gain insight into this question , we studied the sensing characteristics of one of the biochemically simplest of sensors: the allosteric transcription factor . Such proteins , common in microbes , directly transduce the detection of a sensed molecule to changes in gene regulation . Using the Monod-Wyman-Changeux model , we determined six sensing characteristics – the dynamic range , the Hill number , the intrinsic noise , the information transfer capacity , the static gain , and the mean response time – as a function of the biochemical parameters of individual sensors and of the number of sensors . We found that specifying one characteristic strongly constrains others . For example , a high dynamic range implies a high Hill number and a high capacity , and vice versa . Perhaps surprisingly , these constraints are so strong that most of the space of characteristics is inaccessible given biophysically plausible ranges of parameter values . Within our approximations , we can calculate the probability distribution of the numbers of input molecules that maximizes information transfer and show that a population of one hundred allosteric transcription factors can in principle distinguish between more than four bands of input concentrations . Our results imply that allosteric sensors are unlikely to have been selected for high performance in one sensing characteristic but for a compromise in the performance of many .
Sensing is fundamental to life . All cells detect chemicals in their environment and modify their physiology in response to the chemicals detected . Yet fluctuations in the concentrations of extracellular chemicals and stochasticity in intracellular biochemistry confound the cellular “decision” of which physiological change is most appropriate [1] . Sensing extracellular changes is one of the first stages of such decision-making , but why different biochemical networks use different sensing biochemistry is largely unknown . In eukaryotes , signal transduction in some networks is initiated by , for example , G protein-coupled receptors [2] and by receptor tyrosine kinases [3] in others . Initiation can even occur directly in the case of nuclear receptors [4] . In bacteria , the sensing of extracellular changes can be relayed through two-component signalling systems [5] or directly transduced into changes in gene regulation through allosteric transcription factors [6] . Here we consider the advantages and disadvantages of one of the simplest sensors , common in microbes , the allosteric transcription factor . The activity of allosteric sensors is regulated by their interaction with the molecules they sense . Thinking of information transfer , we will refer to these molecular signals as input molecules ( to the sensing system ) . Such input molecules bind to a site that is distinct from the DNA-binding site of the sensor . Allosteric sensors are often considered to have two main conformations [5] , [7] , [8] , [9] , [10] , [11] and stabilise into one of these conformations upon binding an input molecule . The stabilized conformation may lead to new gene expression and either favours binding of the sensor to DNA if , for example , the sensor is a transcriptional activator or disfavours DNA-binding if the sensor is a transcriptional repressor . In general , sensors require several different characteristics to perform well . A sensor should generate outputs that are distinguishable through , for example , having a wide extent of possible outputs ( a high dynamic range to use terminology from engineering ) . For some systems , a sensor ought to respond only to changes in the input that are sufficiently large: the input-output response curve should be sigmoidal rather than hyperbolic . A sensor should not be too “noisy” because changes in the output should be related as best as possible to changes in the input and not be generated by intrinsic fluctuations of the sensing biochemistry if the sensor is to transfer information despite these intrinsic fluctuations . It may also be beneficial if sensors filter any fast dynamics of the input because such changes may be “input noise” and unrelated to the slower extracellular change of interest . A sensor ought to be able to detect small changes in the input by amplifying these changes to large changes in the output: it should have a high gain . Finally , the time taken to sense is important – organisms with precise , but slow sensors may be outcompeted by organisms that respond quickly if not always appropriately – as too is the metabolic cost of synthesizing and maintaining sensors and of the sensing itself . For any particular biochemical network , it is challenging to know which of these sensing characteristics is favoured . Fast sensing may be important for responding appropriately to an increase in temperature whereas slow but accurate inference of the state of nutrients in the environment may be preferred before initiating sporulation . Using an established model of allosteric transcription factors , we will determine how six sensing characteristics change as the biochemical parameters of individual sensors and the number of sensors alter . Our approach is inspired by that of Detwiler et al . who studied G-protein signalling [12] . A similar methodology has also been applied to small RNAs [13] and to aptamers [14] , but we extend the approach by using mutual information to quantify interdependences between the characteristics of a collection of sensors , each sensor having randomly chosen biochemical parameters . Our goal is to understand biological design . We wish to develop biophysically plausible hypotheses to explain why one sensing system might have , for example , hundreds of allosteric sensors that are dimers and another have tens of sensors that are tetramers . To do so , we should discover which biochemical parameters predominantly determine which sensing characteristic and how the different characteristics “play off” against each other . In engineering , for example , a compromise must be reached between the gain and the bandwidth when designing an amplifier [15] . We will investigate whether analogous “rules-of-thumb” exist for allosteric sensing . From the perspective of synthetic biology , we also wish to know which parameters to manipulate to determine particular sensing properties and whether all regions in the space of characteristics can be reached with biophysically realistic values of parameters . We begin with the Monod-Wyman-Changeux ( MWC ) model of allostery [10] and consider six characteristics of allosteric sensing: the dynamic range and the Hill number of the average response , the intrinsic noise , the capacity , the mean static gain and the average response time . We wish to uncover the trade-offs between these characteristics and understand the constraints they pose on allosteric sensing .
The classic description of an allosteric sensor is the MWC model [10] , which forms the basis of our analysis . We assume that sensors can exist in two conformational states , which differ in their quaternary structure and properties of interaction . We will call these states the T state , or the inactive state , and the R state , or the active state . In the absence of any input molecules , the sensor has an intrinsic bias towards the T state . The sensor detects a signal by binding to input molecules . Such molecules bind preferentially to the R state and so counteract the intrinsic bias towards the T state ( Fig . 1A ) . Consequently , an individual sensor will spend more time in the R state when bound by an input molecule than when unbound , and the equilibrium between R and T states of the population of sensors that are unbound by input molecules will also shift to favour more R-sensors . Both effects encourage the binding of additional input molecules . We considered sensors that consist of a number of identical subunits , each with its own allosteric binding site . Our analysis , though , applies for any kind of protein with allosteric binding sites that have identical properties when interacting with the same type of input molecule . Following Monod et al . [10] , we assumed concerted transitions: all subunits change simultaneously . The active and inactive conformations are therefore properties of the sensor as a whole and not just of the individual subunits . Transitions between the R and T states can occur regardless of how many molecules of the input are already bound to the sensor , but the behaviour of the system is largely unchanged if they are assumed to only occur when the sensor is not bound by input molecules . Transitions between the R and T states when the sensor is bound do not change the concentrations at equilibrium because the product of the kinetic rates in the cycle ( Fig . 1A ) is the same as the product of the rates in the cycle ( at equilibrium , the average time taken to go round the cycle should be the same in each direction ) . Hence , the kinetic rates fL , bL , fR , bR , fT , and bT in Fig . 1A are sufficient to completely determine the concentrations of the various states at equilibrium , even in the presence of additional transitions between the R and T states . We defined the activity of a population of sensors by the fraction of sensors in the active R state . At equilibrium , the activity , , satisfies [10] ( 1 ) where is the allosteric or equilibrium constant of the transition between the R and T states; is the ratio of the dissociation constant of the sensor and the input molecule when the sensor is in the R state to the dissociation constant when the sensor is in the T state; is the concentration of free input molecules in units of the dissociation constant of the R state; and is the number of subunits , or allosteric binding sites , on each sensor . We can think of as the bias of a sensor towards the T state in the absence of any input signal and as the counteracting bias towards the R state in the presence of the input . In our analysis , we treat the biases and as macroscopic properties of a sensor that do not depend on its number of subunits . If the number of subunits is greater than one , then , once some sensors in the population have already bound input , the increased probability of additional input molecules binding to the sensors usually generates a sigmoidal response curve , i . e . a non-linear increase in the mean activity for a linear increase in the input . This increase can be sharp and the population of sensors can switch from being mostly inactive to mostly active for a small change in the concentration of input . We used six properties to characterize the reliability and efficiency of sensing: 1 ) The dynamic range ( or amplitude ) , , is the difference between the mean basal level of activity , when no input is present , and the mean saturated level , when a high ( infinite ) concentration of input is present ( Fig . 1B ) . Depending on the strength of the biases and of the sensor , the system can have non-negligible activity in the absence of any input ( as high as 50% if ) or saturate below the 100% level of activity ( saturation will occur at the basal level if ) . The calculation of the dynamic range gives ( 2 ) where and . 2 ) The Hill number , , is a measure of the steepness of the switch as the sensors change from being mostly inactive to mostly active , or vice versa , as the concentration of input changes ( Fig . 1C ) . It quantifies the degree of cooperativity of binding of the input molecules . If , then the binding of one molecule encourages the binding of the next; when , there is no cooperativity and the activity increases hyperbolically with the concentration of input . This characteristic describes , then , the non-linearity of the response and so its ability to generate self-perpetuating dynamics [16] , such as bistabilities , through its interactions with downstream components . Mathematically , the Hill number is proportional to the derivative of the response curve at half saturation ( in log space ) . Defining as so that the normalised activity lies between zero and one , we can then write [17] , and so ( 3 ) where is the fraction of inactive sensors ( ) and is evaluated at the value of the input , , that produces an activity of : with . 3 ) The intrinsic noise , , quantifies the relative size of the fluctuations generated by the biochemistry of sensing around the mean level of activity ( Fig . 1D ) . Such fluctuations arise from the stochastic timing of individual chemical reactions . We define as [18] ( 4 ) where is the number of active sensors , , and is the total number of sensors in the system . By approximately solving the master equation that describes the MWC model , we found that ( 5 ) 4 ) The capacity , , provides an upper bound on the number of levels of input that can be sensed and distinguished given intrinsic noise ( Fig . 1E ) . The capacity is found by maximising the mutual information [19] between the input and the activity . For low levels of intrinsic noise [20] , we found that ( 6 ) 5 ) The static gain , , describes the mean change in activity in response to a small step increment in the input ( Fig . 1F ) . The frequency-dependent gain can be found by linearizing an ordinary differential equation model of the system around the equilibrium concentrations ( Methods ) and takes the general form [15] ( 7 ) where all quantities are Laplace transformed and is the angular frequency . Fluctuations in the input represent extrinsic fluctuations . The frequency-dependent gain measures the response of the system to extrinsic variation , and consequently the system’s ability to track small changes in the input . The static gain is defined as and can be obtained from Eq . ( 7 ) , but it can also be calculated by differentiating the steady-state activity with respect to the input [12]: ( 8 ) From the frequency-dependent gain , Eq . ( 7 ) , we find that the sensing system is a low pass filter ( see Methods ) . It can adapt its activity to slow fluctuations of the input , but gradually loses the ability to respond as the fluctuations become more rapid ( Fig . 1F ) . Beyond a cut-off frequency , the frequency-dependent gain declines , and we find that the response time determines this cut-off frequency , as expected . The filtering properties of the system at high frequencies are independent of the number of allosteric subunits in each sensor: the frequency-dependent gain falls as for all n ( Methods ) . 6 ) The response time , , is the time the system takes to reach the level of activity that is equidistant between the basal level and the maximum level for a particular concentration of input . We assume that initially there is a basal level of activity and that input undergoes a step increase from zero ( Fig . 1G ) . Our results imply that nature is not free to choose each sensing characteristic independently . Specifying certain characteristics restricts the values of others and some regions of characteristic space may even be inaccessible . We used numerical methods to explore generally the effects of one characteristic on another . We considered the properties of the characteristics for a randomly sampled set of parameter values . We sampled the two biases from a uniform distribution in log-space and so assumed that all orders of magnitude are equally probable . To determine the average response time of the system , we also need the kinetic rates . We sampled these rates similarly to our sampling of the biases and use the values of the biases already sampled to calculate one rate . For example , is the ratio of the forward and backward rates in the transition between R0 and T0 ( from Fig . 1A , with fL being the rate of transitioning from R0 to T0 and bL being the rate of transitioning from T0 to R0 ) . Given we can then sample freely one of the two rates , but the other is constrained . Similarly , and constrains another kinetic rate . Some allosteric sensors may undergo transitions between active and inactive states even when bound by the input ligand , unlike the model of Fig . 1A . Such effects will not change five of the characteristics , but will typically diminish the response time . We therefore compared a model in which only transitions between R0 and T0 exist to one in which transitions are possible between all states of ligand occupancy . We sample the additional kinetic rates as described above but now constrained to the condition that at each level of occupancy – Ri and Ti – the ratio of transition rates satisfies . For , we find that the mean response time is about two orders of magnitude lower when all transitions are considered . Our analysis is not concerned with the absolute values of the response time , but its relation to the other characteristics , and throughout we use the model with transitions only between R0 and T0 states because , at least for , the qualitative behaviour of these relationships is model-independent . As well as the general relationships between characteristics we have already discussed , we also observed a weak correlation between the average response time and the Hill number ( Fig . 5B ) . Generating a highly sigmoidal response is biochemically more complex and usually requires more chemical reactions . Consequently , there is a greater probability than one reaction will be slow reducing the overall response time . The correlation is maintained for the model with all active-inactive transitions but disappears when we only consider simulations where all extra kinetic rates are bound between 103 s−1 and 10−3 s−1 ( Methods ) . We found that large regions of the space of characteristics are inaccessible . Plotting dynamic range versus Hill number for our randomly sampled parameters ( Fig . 5A ) , we observe a well-defined forbidden region of characteristics space that organisms using allosteric sensing cannot access . There is a tight constraint on the dynamic range when the Hill number is greater than one ( Fig . 5A ) . We can analytically determine the boundary of this region . For example , when all data points in the characteristic space fall on a line that bounds the points in Fig . 5A from below . Similarly , a scatterplot of capacity and dynamic range ( Fig . 5C ) shows that the data fall only in a narrow area of characteristic space . Most of the space is inaccessible . Further , high Hill numbers constrain the intrinsic noise to a narrow band of possible values ( Fig . 5D ) . Attempts at reducing intrinsic noise below this constrained band require sharply reducing the Hill number ( for all sensors with more than one subunit ) . As we have seen , high Hill numbers imply high dynamic ranges ( Fig . 5A ) with basal levels close to zero and saturation levels close to one and , consequently , the threshold of the response curve will coincide with 50% of the sensors being active . From Eq . ( 5 ) , the intrinsic noise is then fixed at , where is the number of sensors , and is insensitive to the values of the two biases . Hence the narrow peak observed in Fig . 5D . When looking at the six-dimensional space of all characteristics , we found most of the space is empty and is inaccessible to allosteric sensing systems: for all numbers of subunits , we find 90% of all sampled systems are contained within less than 2% of the space of characteristics , and all systems lie within less than 6% of the space ( in contrast , 10 , 000 samples of six randomly distributed characteristics would with this measure occupy 100% of the six-dimensional space ) . The densest regions have relatively high dynamic range and capacity , low intrinsic noise , and relatively low static gain for all numbers of subunits . As the number of subunits increases , both the static gain and the Hill number are more evenly distributed across the densest regions ( Fig . 2 ) , and the response time takes average values . To quantify further the trade-offs between pairs of characteristics , we calculated the normalised mutual information [19] between all possible pairs for our randomly sampled parameters ( Methods ) . We normalised by either the entropy of the first or the second characteristic of the pair ( Fig . 6 ) . For example , if we normalise by the entropy of the first characteristic , the normalised mutual information measures the fraction of the entropy of the first characteristic that is constrained by specifying the second . When one characteristic is fixed , such as the Hill number for , then it shares no mutual information with others ( inset of Fig . 6 ) , and changing any other characteristic cannot alter the Hill number . There is a trend for the constraints between characteristics to increase slightly as number of subunits of the sensors increase ( compare the colour of the matrix for with the matrix for in Fig . 6 ) . The characteristics are therefore more independent for lower numbers of subunits . The normalised mutual information need not be symmetric: specifying one characteristic can therefore constrain another more than specifying the second characteristic constrains the first . In the matrix , we emphasize the trade-off between the Hill number and the dynamic range . The entropy of the Hill number is higher than that the entropy of the dynamic range ( inset of Fig . 6 ) . Hence specifying the Hill number constrains the dynamic range more than specifying the dynamic range constrains the Hill number . A similar phenomenon occurs for the constraints between the dynamic range and the intrinsic noise ( Fig . 6 ) . We compared the results of the normalised mutual information involving the response time for both the model of Fig . 1A and the alternative model with transitions between all active and inactive states when . For our sample of simulations the maximal deviation observed between both models was about 70% . The normalised mutual information between the response time and any other characteristic , however , remained low and qualitatively Fig . 6 is unchanged .
For allosteric transcription factors , one of the simplest biological sensing systems , we found several relationships between the system’s sensing characteristics and that specifying one characteristic strongly constrains others ( Figs . 2 , 5 and 6 ) . Using the Monod-Wyman-Changeux model , we showed that the dynamic range of a collection of sensors reaches its maximum for most values of the two biases , K and c , particularly as the number of subunits comprising each sensor increases . We found that the Hill number of the mean input-output response curve and the capacity – its ability to transmit outputs that distinguish changes in the system’s input – are both strongly correlated with the dynamic range and that the Hill number is inversely correlated with the static gain . Further , we showed that the intrinsic noise typically decreases as the number of subunits on each sensor increases . Perhaps surprisingly , we discovered that most of the space of characteristics is inaccessible for typical values of biochemical parameters ( Fig . 5 ) . For the collection of parameter values we considered ( 10 , 000 sets of parameters in all ) , constraints between characteristics caused less than 6% of the space of characteristics to be occupied . A sensing system must therefore trade high performance in one characteristic for low performance in another . For example , the intrinsic noise in the response is highest when the number of active sensors is low and therefore will be reduced by increasing the system’s basal level of activity ( Fig . 3D ) . Such an increase , however , diminishes the dynamic range and also therefore the system’s capacity ( Fig . 2 and Fig . 5C ) . The fall in intrinsic noise is not enough to overcome the decrease in capacity caused by reducing the dynamic range . Reducing the noise would also decrease the Hill number ( Fig . 5D ) . The response time , however , would be expected to become faster ( Fig . 5B ) . Such constraints tighten as the number of subunits in each sensor increases ( Fig . 6 ) . To maximise information transfer , the system should generate discriminative outputs for as many bands of input concentrations as possible . We found that the distribution of output that maximized the mutual information between the input and the output is indeed discriminatory being peaked only at low and high values . Similarly , the corresponding optimal input distribution has high probabilities for those inputs where a small change in input gives a large change in the mean output activity and has low probabilities for inputs that give little mean change in the output ( Fig . 4 ) . We find that the capacity has a maximum value of around two bits for a system with a population of 100 sensors ( Fig . 2 ) , and so allosteric transcription factors can therefore distinguish between four bands of input concentrations , at least when our assumptions of low intrinsic noise and more input molecules than sensors hold . Selection , though , need not favour the ability to distinguish multiple bands and therefore , say , multiple states of the extracellular environment , but rather the ability to quickly and reliably determine a few states , such as the presence or absence of a toxin . Further , some cellular information-processing is likely to be dynamic [24] with the system not having time to reach steady-state as we have assumed here . The number of subunits a sensor has enables the sensing system to access different regions of the space of characteristics ( Fig . 2 ) . Comparing 33 allosteric transcription factors in Escherichia coli [25] , we found that 33% are monomers and 48% are dimers with only 18% having more than two subunits ( Table 1 ) . Being a dimer helps a transcription factor recognise palindromic sequences in promoters [6] , but having two subunits also perhaps gives a profitable compromise between the fragility and robustness of the sensing characteristics [26] , [27] . Dimeric systems have both substantial regions of parameter space where some characteristics , such as the dynamic range , intrinsic noise , and capacity , vary and equally substantial regions where the dynamic range and the capacity are large and the intrinsic noise small ( Fig . 2 ) . Sensors with four or eight subunits , however , will have a near maximal capacity for most values of the two biases , at least when there are 100 sensors and given our approximations ( Fig . 2 ) , but do not appear common in Escherichia coli . To include the higher biochemical costs of synthesizing sensors with four or eight subunits , we can compare the capacity for equal numbers of subunits rather than equal numbers of sensors ( e . g . , a system with 50 dimers versus a system with 25 tetramers ) . From Eq . ( 6 ) , however , the maximal capacity declines with the total number of sensors favouring dimers over tetramers when the total number of subunits is limiting . An important caveat to our results is that natural selection presumably acts on the entire biochemical system , not only on upstream allosteric sensing but also on both the direct and indirect downstream gene expression . Our analysis , however , is perhaps best extended to specific systems because to understand trade-offs and constraints in those systems we need to know the biochemical details of their control and regulation , the probability distribution of typical inputs , the costs and benefits of potential responses , and ideally how these responses correlate with fitness [1] , [28] . Nevertheless , we can make a few general predictions about how some individual characteristics will change if the output of the system is taken to be the level of expression of a downstream gene . For example , the maximum Hill number describing the response of the last species of a biochemical cascade is given by the product of the Hill number at each stage of the cascade [29] . We can expect the capacity to at best remain unchanged with the addition of each new stage in the cascade from the data-processing inequality [19] . Intrinsic noise in the output can increase because each stage of the cascade is itself a new source of stochasticity [30] , [31] , but need not do so if the input at each stage saturates the output at that stage [32] . Further , assuming that regulation of the downstream gene can be described by a Hill function , we can show by simulation that the dynamic range of the expressed protein is more sensitive to changes in this function’s threshold rather than its Hill number . If we consider that natural selection acts to improve the performance of a sensing system , our results indicate that the performance of allosteric sensors is likely to be a function that balances the values of all the sensing characteristics . Changing one or two necessarily changes others . Indeed , there are other factors , such as structural constraints , difficulties in regulating the molecular assembly of large oligomers , fluctuations in the number of sensors , and energetic costs , that we have not considered and that will impact selection . Nevertheless , the biochemistry of allosteric sensing prevents random changes in the values of biochemical parameters generating random changes in sensing , and , as such , the constraints we have determined here may themselves have been selected to enable allosteric sensors to be evolvable and reduce catastrophic mutations [33] .
The biases and are sampled from a uniform distribution in logarithmic space across six orders of magnitude for each case , i . e . , and . We consider the number of subunits on each sensor to be and 8 . For each , we sample 10 , 000 different sets of and pairs . The kinetic rates , , and ( Fig . 1A ) are sampled a posteriori with the constraint that the previously sampled values of the biases are maintained . Taking the typical volume of an E . coli cell , 10-18 m3 [34] , and the diffusion-limited upper bound on association rates , which is in the order of 107 – 1010 M−1 s−1 [35] , we sample normalised kinetic rates between 10−3 s−1 and 103 s−1 . We sample the kinetic rates for the model that includes transitions between all active and inactive states similarly , maintaining that . We find the response time by simulations of an ordinary differential equation model using the Facile software [36] and MATLAB ( The MathWorks , Massachusetts ) . We normalise the response time by dividing by 1s , which is the timescale associated with the central value of the range we chose for the kinetic rates . We can calculate the intrinsic noise of the system analytically if we assume that the number of input molecules is much greater than the number of the sensors . We can then consider each sensor to act independently: any input molecule binding to one sensor does not affect the availability of input molecules for other sensors because the number of free input molecules is always assumed to be much greater than the number of sensors proteins , or more exactly the total number of allosteric binding sites . We proceed by considering a single allosteric molecule . In the presence of input molecules , the sensor will transition between R and T states according to the transition probabilities for each trans x transitioning from a state to any of its neighbours and the transition probabilities are independent of how the system reached that state . Denoting the time-dependent state vector of the system by and the transition rate matrix by , where is the transition probability between states i and j and , then ( 13 ) and we have and at steady-state . Eq . ( 13 ) is identical to the system of deterministic rate equations that describe the dynamics of the mean activity for a fixed amount of free input . The probability of a sensor being in the active state is given by ( 14 ) These results then show the probability of an individual sensor being active is given by from Eq . ( 1 ) – the deterministic ( mean ) activity for a population of sensors . Furthermore , in a model that includes transitions between the R and T states when they are bound by input molecules , neither the mean activity ( Eq . ( 1 ) ) nor the probability of a sensor being active ( Eq . ( 14 ) ) are altered because of the thermodynamic constraint between parameter values created by the presence of cycles of reactions ( involving , , and ) . To extend our results to a stochastic population of sensors , we need to consider all possible configurations of the individual sensors that correspond to a particular activity of the population . For example , for a system with two sensors and one subunit , , our approximation implies that both sensors are independent and the probability that both are active is then ( 15 ) Similarly , the probability that only one sensor is active is and the probability that no single sensor is active is . Extending the argument for systems with higher numbers of sensors , the activity obeys a binomial distribution [22]: the probability of having m of N sensors active is ( 16 ) which corresponds to the probability of an activity of m/N . The capacity is the upper bound of the mutual information between the input and output , i . e . , the activity , given a model for the intrinsic noise . The mutual information is defined as [19] ( 17 ) Using a small Gaussian noise approximation , Tkačik et al . [20] derived the optimal solutions of the input and output that maximise the mutual information , i . e . , the capacity . Their solution is ( 18 ) where . Using the expressions for and in Eqs . ( 1 ) and ( 9 ) , we find the capacity in the MWC system satisfies Eq . ( 6 ) . All stochastic simulations were performed using the EasyStoch software [38] , which implements the Gibson-Brück [39] version of the Gillespie algorithm [40] . Given the general form of a linear system , , , where , , … , are the state variables , , , … , is the input and , , … , is the output , it can be shown that the frequency response is given by with I being the identity matrix [15] . The frequency-dependent gain measures the relation between the input and the output of linear systems , and here we use it to quantify how the system responds to extrinsic fluctuations in the input signal . We ignore intrinsic fluctuations and consider the system to be at equilibrium , or at an ‘operating point’ , and introduce a small perturbation term to the input , L , so that it becomes . This fluctuation propagates through the system and each variable gains a small correction whose dynamics we follow by linearizing an ordinary differential equation model of the system . We can then write down the matrices and calculate the frequency-dependent gain directly . The frequency-dependent gain has the form ( 19 ) where is the complex argument of the Laplace transform . The roots of the numerator , , are the zeros of the system; those in the denominator , written as , are the poles . The exact number of zeros and poles varies with the number of subunits on the sensors , and the magnitude of the gain rises for each zero and falls for each pole [15] . We found the number of poles always to be greater than the number of zeros and therefore the system is a low-pass filter . We consider the pole with the lowest frequency to be the cut-off frequency . For all numbers of subunits the number of poles exceeds the number of zeros by two , so in the limit of large frequencies the transfer function declines with . To give an estimate of the density and occupancy of the space of characteristics , we divided the space of characteristics into equally sized hypercubes and counted the number of sampled sets that fall into each hypercube . We binned each characteristic into low , medium and high levels , thus obtaining a total of 729 hypercubes . Mutual information is a statistical measure that quantifies how much knowing one random variable informs on another [19] , [41] . It is symmetric , but we normalise it in two ways , by the entropy of each characteristic in the pair . For characteristics and , the normalised mutual information in its discretized form is ( 20 ) where is the mutual information between two characteristics and is the entropy of one characteristic . We estimate the probability distributions for the characteristics by binning values of the characteristics from our randomly sampled parameter sets into 30 bins . Although the probabilities are then dependent on the numbers of bins , we varied that number without observing a substantial qualitative change in the mutual information .
|
Sensing environmental changes is the first step in the process of cellular decision-making , but many different biochemical sensors exist and why one sensor is selected for a particular task over another is not known . Here we study the sensing properties of a simple and generic allosteric sensor to understand the effectiveness and limitations of its “design” . We begin by defining and calculating a set of six engineering-inspired characteristics of the sensor’s response and investigate how specifying a high performance in one characteristic constrains the sensor’s performance in others . We determine many such trade-offs and , perhaps surprisingly , that much of the space of characteristics is inaccessible given biophysically plausible ranges of parameters . Our results suggest that allosteric sensors are not under selection for high performance in one sensing characteristic but for a compromise in performance between many . Our approach provides both quantitative and qualitative insights about the function and robustness of allosteric sensors and as such is applicable to both the study of endogenous systems and the design of synthetic ones .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"systems",
"biology",
"biochemical",
"simulations",
"theoretical",
"biology",
"biology",
"computational",
"biology",
"signaling",
"networks"
] |
2011
|
Trade-Offs and Constraints in Allosteric Sensing
|
Wolbachia are vertically transmitted , obligatory intracellular bacteria that infect a great number of species of arthropods and nematodes . In insects , they are mainly known for disrupting the reproductive biology of their hosts in order to increase their transmission through the female germline . In Drosophila melanogaster , however , a strong and consistent effect of Wolbachia infection has not been found . Here we report that a bacterial infection renders D . melanogaster more resistant to Drosophila C virus , reducing the load of viruses in infected flies . We identify these resistance-inducing bacteria as Wolbachia . Furthermore , we show that Wolbachia also increases resistance of Drosophila to two other RNA virus infections ( Nora virus and Flock House virus ) but not to a DNA virus infection ( Insect Iridescent Virus 6 ) . These results identify a new major factor regulating D . melanogaster resistance to infection by RNA viruses and contribute to the idea that the response of a host to a particular pathogen also depends on its interactions with other microorganisms . This is also , to our knowledge , the first report of a strong beneficial effect of Wolbachia infection in D . melanogaster . The induced resistance to natural viral pathogens may explain Wolbachia prevalence in natural populations and represents a novel Wolbachia–host interaction .
Wolbachia are obligatory , intracellular α-proteobacteria that infect a wide range of arthropods and filarial nematodes . They are found in 17% to 76% of surveyed arthropods and have recently been estimated to be present in 66% of all arthropod species [1–3]; therefore Wolbachia are one of the most widespread intracellular bacteria . Although the phylogenies of Wolbachia strains and their arthropod hosts show horizontal transmission of the bacteria on an evolutionary time-scale [4] , these endosymbionts are mainly transmitted maternally . Consequently , Wolbachia strains and the species they infect form long-term associations . Wolbachia were first discovered infecting the mosquito Culex pipiens in 1924 [5] , but interest in these bacteria mainly arose when it was shown that infected mosquito males do not successfully breed with noninfected females [6] . This phenomenon is termed cytoplasmic incompatibility ( CI ) and has , since then , been found in many other insect species infected with Wolbachia [7] . In some hosts , Wolbachia can also cause feminization , male killing , or parthenogenesis [7] . All these mechanisms profoundly alter the reproductive biology of their hosts and are thought to increase the success of bacterial transmission through the female germline . In the majority of known cases , Wolbachia behave like reproductive parasites of their hosts . Interestingly , in the parasitic wasp Asobara tabida , a Wolbachia strain is required for the inhibition of apoptosis in the germline and , consequently , normal oogenesis [8 , 9] . Similarly , Wolbachia is required for normal development and fertility in many filarial nematodes [10–13] . In all these cases , the endosymbionts are obligatory mutualists—they are essential for the survival of their host species . Curiously , examples of Wolbachia infections that are facultative and provide a fitness benefit are rare ( e . g . , [14 , 15] ) . One would , however , expect them to be frequent , since these are long-term symbioses , and Wolbachia fitness ultimately depends on the host fitness . The model organism Drosophila melanogaster can also be infected with Wolbachia . In fact , they are detected in a large proportion of flies of natural populations and laboratory stocks [16–18] . Interestingly the presence of wMel , the Wolbachia strain associated with D . melanogaster , does not seem to cause a strong phenotype . wMelPop , a Wolbachia variant from a laboratory stock , does causes tissue degeneration and significantly shortens the lifespan of its carriers [19] . The appearance of this strain may be an artifact of conditions in which laboratory stocks are kept , because no wMel variant from natural populations with these characteristics has been discovered . Wolbachia also rescues the sterility of Sex-lethal hypomorphic mutants [20] . However , it is not known how this translates to the interaction of Wolbachia with wild-type flies . How natural variants of Wolbachia affect wild-type D . melanogaster has been extensively addressed . wMel only induces a weak and transient CI phenotype in D . melanogaster [21–23] , although some Wolbachia strains induce strong CI in the closely related D . simulans [24] . This low CI cannot explain how Wolbachia spreads and is maintained in wild-type populations , especially considering that , in the wild , infection is not vertically transmitted with 100% fidelity [25] . A strong hypothesis to explain Wolbachia presence in natural populations is that Wolbachia gives a fitness benefit to D . melanogaster [26 , 27] . Several studies have either been unable to find differences in fitness parameters or found only slight beneficial or detrimental effects of Wolbachia infection [25 , 27–31] . Moreover , even when effects were observed , they were dependent on the Wolbachia variant or the fly's genetic background . A clear strong beneficial effect of Wolbachia infection in D . melanogaster has still not been shown , and it remains a puzzle why these bacteria are so prevalent in natural populations . D . melanogaster is a valuable tool in the study of resistance to pathogens , with many components of innate immunity signaling pathways conserved between Drosophila and mammals [32] . The epitome of its utility was the discovery of the involvement of Toll-like receptors in innate immunity . Toll was first discovered to be important in the resistance of Drosophila to fungi [33] , later Toll-like receptors were shown to have fundamental functions in mammalian innate immunity [34] . Moreover , Toll-like receptors are important in the activation and modulation of mammalian adaptive immunity . The responses of Drosophila to systemic infection by fungi and bacteria are increasingly well known [35] . Although less extensively , Drosophila has also been used as a model system to study resistance to viruses . Recent research has shown conservation between flies and mammals in their immune response to viruses . Mutations in hopscotch , the gene that encodes the kinase of the JAK-STAT pathway , reduce resistance to Drosophila C virus ( DCV ) infection and increases viral titers [36] . Interestingly , in mammals , JAK-STAT pathways are involved in cytokine signaling , including anti-viral type I interferon [37] . Work in Drosophila has also been important in showing that RNA interference is involved in anti-viral resistance in animals . Flies mutant in genes that encode components of this pathway , Dicer-2 , Argonaute-2 , and r2d2 , are more sensitive to infection by several RNA viruses and have higher titers of viruses than the wild type [38–41] . To identify new genes involved in Drosophila resistance to viruses , we initiated a screen for DCV-sensitive flies . In doing so , we found that flies infected with intracellular bacteria were much more resistance to DCV infection than those that were uninfected . We identified these bacteria as Wolbachia and show that DCV titers are much lower in Wolbachia-infected flies . Moreover , resistance to infection extends to two other RNA viruses but not to a DNA virus . These results identify a new major factor involved in Drosophila resistance to RNA viruses , and the first strong beneficial effect associated with Wolbachia infection in D . melanogaster .
In order to identify new genes involved in D . melanogaster resistance to viruses , we are conducting a genetic screen for virus-sensitive mutants ( LT , AF , MA , unpublished data ) . We have generated a collection of mutant lines by P-element insertional mutagenesis using the set of w1118 iso isogenic lines described in Ryder et al . , 2004 [42] . We chose an isogenic background to minimize variability in the response to viral infection and , for the same reason , cleaned the initial set of lines of potential chronic viral infections using the protocol described in Brun and Plus , 1978 [43] . We test the resistance of each insertion line to DCV infection . DCV is a small , non-enveloped virus with a single-stranded , positive-sense RNA genome that belongs to the Dicistroviridae family , an insect specific family of viruses very similar to picornaviruses [44] . This virus is a natural pathogen of D . melanogaster , it is sequenced and relatively well characterized , and its infection has an easily scored lethal phenotype [43 , 44] . In the initial screen , we assayed adult survival after intra-thoracic DCV injection and realized that , unexpectedly , the control w1118 iso line was much more sensitive to DCV than most of the tested P-element insertion lines ( Figure 1 and unpublished data ) . When injected with a dose of 500 times the median tissue culture infective dose ( TCID50 ) , all the w1118 iso males died within 12 d , whereas very little death was observed in the males of the P-element insertion lines ( Figure 1A ) . Moreover , a large proportion of males of the P-element insertion line survived until 21 d after infection . Preliminary analysis showed that the P-element was not responsible for virus resistance ( unpublished data ) . We then tested the hypothesis that a previous treatment of the w1118 iso line with tetracycline could have rendered it more sensitive to DCV . We raised flies of a P-element insertion line , VF-0058–3 , on tetracycline-containing medium or control medium and compared the adults' resistance to DCV infection ( Figure 1A ) . The tetracycline treatment made the flies die much faster upon DCV infection , with a sensitivity similar to that of w1118 iso line . This result strongly suggested that a tetracycline-sensitive bacteria , associated with the resistant stocks , conferred resistance to the viral infection . We discarded the possibility that the effect was an artifact of the tetracycline per se because raising w1118 iso flies on medium with tetracycline did not make them more sensitive to DCV ( Figure 1B ) . We also treated the VF-0058–3 and VF-0097–3 P-element insertion lines with tetracycline for two generations and moved them back to normal medium for at least five generations in order to negate any side effects of tetracycline itself ( these stocks will be referred as VF-0058–3t and VF-0097–3t ) . We then repeated the assay , comparing resistance to DCV of these treated stocks to the non-treated stocks ( Figure 1C ) . The tetracycline treatment makes the P-element insertion lines stably more sensitive to DCV than non-treated lines and equally sensitive to DCV as the w1118 iso line . A similar result was obtained when females of these lines were injected with DCV , with the difference that females are less sensitive to DCV than males ( unpublished data ) . Importantly , in the timeframe of this assay , the survival of tetracycline-treated and non-treated stocks do not differ when only injected with buffer ( Figure 1D ) . In summary , these results show that tetracycline-sensitive bacteria , not easily acquired from the laboratory environment , confer on D . melanogaster resistance to DCV . The increased resistance to DCV could be due to increased resistance to the damage caused by the viral infection or decreased viral proliferation . To test this , we probed by Western blot the levels of viral proteins in VF-0058–3 and VF-0058–3t adult flies after infection with DCV ( Figure 1E ) . While viral proteins were not detectable on extracts of DCV-infected VF-0058–3 flies , they were clearly detectable on extracts of DCV-infected VF-0058–3t flies , and their levels increased from 3 to 6 d post-infection . We extended this analysis by quantifying , in cell culture , the viral titer in these flies after DCV infection ( Figure 1F ) . DCV is detected after infection in flies from both stocks and slightly increases from 3 to 6 d post-infection . However , DCV levels are considerable higher , by approximately 10 , 000 times , in VF-0058–3t flies . These experiments show that the bacteria that confer resistance to DCV infection interfere with the virus proliferation . The fact that the tetracycline treatment permanently renders the flies sensitive to DCV shows that the bacteria are not easily acquired from the laboratory environment . To test if the resistance could be horizontally acquired , we raised together the progeny of w1118 iso females ( without resistant-inducing bacteria ) with the progeny of VF-0058–3 females ( with resistant-inducing bacteria ) and then assayed the levels of viral proteins in infected flies ( Figure 2A ) . The progeny of these females can be distinguished by their eye color , due to the presence of a functional white gene , in the RS3 transposon , only in the progeny of VF-0058–3 flies . w1118 iso flies do not acquire the resistance to DCV when raised mixed with the progeny of VF-0058–3 flies . Therefore , the bacteria that confer resistance to DCV are not acquired horizontally . We then tested if the resistance was vertically transmitted by crossing males and females from VF-0058–3 and VF-0058–3t stocks in all four possible combinations and assaying the survival of the adult progeny after DCV infection ( Figure 2B ) . The results clearly show that the determinant factor of the progeny resistance is the mother's resistance; therefore , the bacteria in question are maternally transmitted , which strongly suggests they are intracellular . However , the bacteria could , in theory , be only transmitted by the mother but not be intracellular ( e . g . , they could be deposited on the egg surface ) . This did not seem to be the case , because flies that were raised from VF-0058–3 surface-sterilized embryos did not become more sensitive to DCV ( Figure 2C ) . Moreover , we could visualize the presence of intracellular bacteria by DNA staining in embryos from VF-0058–3 and VF-0097–3 stocks but not from VF-0058–3t , VF-0097–3t , or w1118 iso stocks ( Figure 2D ) . We can therefore conclude that the viral resistance is mediated through maternally transmitted intracellular bacteria . To identify the intracellular bacteria in question , we extracted DNA from surface-sterilized embryos of resistant-to-DCV flies ( VF-0058–3 ) and performed PCR amplification using prokaryotic 16S rRNA universal primers . We analysed the product of this amplification by cloning it and sequencing over 100 independent clones . All the 104 sequences of inserts in the cloning plasmid we obtained were at least 99 . 5% identical to the sequence of a fragment of the 16S rRNA gene of Wolbachia ( GenBank accession number EU096232; http://www . ncbi . nlm . nih . gov/Genbank/ ) . Therefore , these embryos , which carry the resistance to DCV inducing bacteria , are most probably only infected with Wolbachia . To verify the presence of Wolbachia , we performed PCR amplification using primers for the Wolbachia specific genes wsp and wspB [45 , 46] ( Figure 2E ) . Wolbachia is present in VF-0058–3 and VF-0097–3 and absent from the VF-0058–3t and VF-0097–3t embryos' extracts . The sequence of the wsp-specific primers' PCR amplification product from the VF-0058–3 flies is identical to the wsp sequence of wMel ( GenBank accession number DQ235407 ) , the only Wolbachia strain known to infect D . melanogaster . In a recent survey in 35 different Drosophila species , that screened over 4 , 500 individuals , only two kind of heritable endosymbiotic bacteria were found: Wolbachia and Spiroplasma [47] . We specifically tested for the presence of Spiroplasma in the VF-0058–3 line ( and the wt-1 to -6 lines used in Figure 3B ) using primers specific for the 16S rRNA gene of Spiroplasma [48] . We detect Spiroplasma in a positive control , RED-67 [48] , but not in any of the other tested lines . Therefore it is not Spiroplasma that confers resistance to viruses . This result and the sequencing results strongly suggest that the maternally inherited intracellular bacteria that confer resistance to DCV are Wolbachia . These results show that resistance to DCV is associated with the presence of Wolbachia , however it could be possible that other cryptic intracellular bacteria were responsible for the virus resistance , and Wolbachia would merely be present in these flies by chance . Wolbachia cannot be cultured and therefore we cannot infect a sensitive stock with a pure cultured isolate and verify acquired resistance to DCV . Wolbachia can be artificially transferred from an infected host to a new host . However , if we did transfer Wolbachia from infected flies to non-infected flies and show concomitant transfer of resistance to DCV , we could not discard the possibility that we were also transferring the hypothetical cryptic bacteria . We addressed this problem by treating the Wolbachia-infected stock VF-0058–3 with a suboptimal dose of tetracycline for one generation and then establishing isofemale lines from the progeny . We expected to obtain lines that kept the Wolbachia infection and other lines that lost it . The segregation of Wolbachia should be independent of the segregation of any hypothetical other bacteria . From two independent sets , one set of ten lines and another set of 23 lines , we established , in total , three lines that conserved Wolbachia infection and 30 lines that lost it . We then tested these lines for resistance to viruses ( Figure 3A ) . Wolbachia presence and viral resistance fully segregate with each other; the probability that the presence of Wolbachia and resistance to DCV are independent is very low ( Fisher's exact test , p = 0 . 0002 ) . These data strongly indicate that it is Wolbachia infection that induces DCV resistance . To corroborate that it is Wolbachia infection that protects D . melanogaster from DCV , we analyzed this interaction in other independent fly stocks . We screened , by PCR , for Wolbachia presence in a collection of wild-type stocks kept in our laboratory and we found six infected lines . After establishing tetracycline-treated stocks derived from these lines , we compared , by Western blots , their resistance to DCV with the original lines ( Figure 3B ) . In all the six cases , the loss of Wolbachia makes the flies more sensitive to DCV . The same procedure was applied to six stocks that did not carry Wolbachia initially ( Figure 3B ) . There is much heterogeneity in the levels of DCV proteins 6 d after infection in these stocks ( e . g . , line 7 is very resistant to DCV infection ) , but , importantly , there is no increase in the levels of DCV proteins in the tetracycline treated stocks . These results show that increase sensitive to DCV after tetracycline treatment is always associated with an initial Wolbachia infection . From this set of data , plus the same result of increased sensitivity with tetracycline treatment of VF-0058–3 , we can state that the probability that initial Wolbachia infection and increased sensitivity upon tetracycline treatment are independent is very low ( Fisher's exact test , p = 0 . 0006 ) . In conclusion , we can confidently state that it is Wolbachia that protects D . melanogaster from DCV infection . Wolbachia and DCV are commonly found in D . melanogaster natural populations and laboratory stocks; their interaction could be very specific . We investigated if Wolbachia protection extends to infections by two other RNA viruses and a DNA virus . Nora virus is a recently described common natural pathogen of D . melanogaster [49] . Similar to DCV , Nora virus is a small , non-enveloped virus with a single-stranded , positive-sense RNA genome . It is similar to picornaviruses and dicistroviruses , but it has a unique genome organization . Using reverse-transcription PCR ( RT-PCR ) with Nora virus–specific primers , we found that the VF-0058–3 and VF-0058–3t stocks are not infected with it while another laboratory stock , Oregon R , is ( Figure 4A ) . An extract of Oregon R adult flies was injected into VF-0058–3 and VF-0058–3t adult flies , and the levels of Nora virus replication were accessed , by semi-quantitative RT-PCR , after 3 d ( Figure 4A ) . We observed , in four independent repeats , that Nora virus levels are lower in Wolbachia-infected flies; Wolbachia infection also protects Drosophila from Nora virus infection . Flock House virus ( FHV ) belongs to the Nodaviridae family of insect viruses . These are small viruses containing two single-stranded , positive-sense genomic RNAs [50] . FHV was isolated from a coleopteran [51] and is not a natural pathogen of D . melanogaster . However , it can be cultured in D . melanogaster cells and proliferates and causes death in adult flies when injected [38 , 50] . We injected FHV in VF-0058–3 and VF-0058–3t adult flies and followed their survival ( Figure 4B ) . Wolbachia-infected flies were much more resistant to FHV infection; with an infection dose of 50 TCID50 , all VF-0058–3t flies die by day 13 , while only 40% of VF-0058–3 flies die by day 21 . We can detect , by Western blot , increase in FHV proteins with time in both infected stocks ( Figure 4C ) . Surprisingly , we only detected a slight increase in FHV proteins in the infected VF-0058–3t flies , compared with the infected VF-0058–3 flies . We confirmed this result by determining , in cell culture , the viral titer per infected fly , 6 d post-infection ( Figure 4D ) . We find , on average , only 1 . 8-fold more FHV in Wolbachia-free flies . and the difference is not statistically significant ( Mann-Whitney test , p = 0 . 05764 ) . We can conclude that Wolbachia presence also increases the resistance of D . melanogaster to FHV infection although it does not or only slightly affects FHV levels . Finally , we wanted to test the effect of Wolbachia on a DNA virus infection; however , there is no known DNA virus that is a natural pathogen of D . melanogaster . Insect Iridescent Virus 6 ( IIV-6 ) ( also named Chilo iridescent virus ( CIV ) ) is a large virus with a double-stranded DNA genome from the Iridoviridae family [52] . It was first isolated from a lepidopteran but can infect a large number of different insects and cultured insect cells [53 , 54] . IIV-6 can infect and replicate in D . melanogaster cells [54] and cause adult flies death upon injection ( Peter Christian , personal communication ) . We confirmed that IIV-6 infection causes premature death in adult flies , approximately halving their lifespan when 1 , 000 TCID50 are injected per fly ( Figure 4E ) . Importantly we can show that IIV-6 replicates in adult D . melanogaster . Infected flies become iridescent as they accumulate virions , a characteristic of iridoviruses due to their paracrystalline packing ( Figure 4F ) . Moreover , IIV-6 titer per fly after 10 d of infection , determined in cell culture , was approximately 109 TCID50 , when a dose of 103 TCID50 was injected ( Figure 4G ) . Contrary to the results obtained with DCV and FHV infection , Wolbachia-infected flies actually died faster than Wolbachia-free flies when infected with IIV-6 ( Figure 4E ) . This probably represents just a cumulative effect of the deleterious effects of Wolbachia and IIV-6 infection . In fact , Wolbachia infection has a long-term deleterious effect that results in a shorter lifespan in the absence of viral infection ( Figure 4E ) . In accordance with this interpretation , the average IIV-6 titer , 10 d after injection , is only 1 . 8-fold higher in Wolbachia-infected flies compared with Wolbachia-free flies , and not significantly different ( Mann-Whitney test , p=0 . 08118 ) ( Figure 4G ) . In conclusion , Wolbachia presence does not protect D . melanogaster from IIV-6 infection .
We have shown that Wolbachia infection in D . melanogaster induces resistance to DCV infection . Several lines of evidence lead to this conclusion . The resistance to DCV was maternally transmitted and sensitive to tetracycline , as is Wolbachia; in embryos infected with bacteria inducing resistance to DCV , we can only detect the presence of Wolbachia; all tested D . melanogaster lines that carried Wolbachia became more sensitive to DCV after tetracycline treatment; lines that did not carry Wolbachia did not become more sensitive to DCV after tetracycline treatment . Finally , when transmission to the next generation was imperfect , due to treatment of larvae with a low dose of tetracycline , Wolbachia and resistance to DCV co-segregated . Following Occam's razor principle—Pluralitas non est ponenda sine necessitate . “Plurality should not be posited without necessity . ” —the simplest and most plausible hypothesis is that Wolbachia is the causative agent of resistance to DCV . Infection by Wolbachia considerably increased the lifespan of DCV-infected flies . This is due to a strong reduction in viral titers , as observed by Western blot and titration by cell culture . At 3 d post infection , the DCV titer in Wolbachia-infected flies was 10 , 000 times less than that in Wolbachia free flies . This difference is larger than that reported between the wild type and mutants in the anti-viral resistance genes Dcr-2 , ago-2 , and hop [36 , 38 , 40] . Wolbachia is clearly a major factor affecting Drosophila resistance to DCV . Wolbachia and DCV are common symbionts of D . melanogaster . However , the interaction is not specific to DCV; we found that Wolbachia also induced resistance to two other RNA viruses . In the case of Nora virus , there was also reduction in the viral titer of infected flies . FHV infection , in terms of mortality , was also much less severe in the presence of Wolbachia , to a degree similar to that seen with DCV . But , with this virus , Wolbachia only slightly affected viral titer . The resistance to FHV is most probably an increase in resistance to the damaged caused by the viral infection rather than an ability to inhibit virus proliferation . However , we cannot exclude the possibility that there is strong inhibition of FHV proliferation in certain essential adult tissues or that a small decrease in viral titer is enough to significantly increase the lifespan of infected individuals . DCV and Nora virus differ from FHV in two ways: they are both natural pathogens of Drosophila and both are picornavirus-like . An endogenous virus and its host could be co-adapted so that a small advantage , in this case provided by the bacteria to the host , would profoundly tilt the equilibrium between virus and host , whereas an exogenous pathogen may be less sensitive to bacterial infection of its host . On the other hand , Wolbachia could interfere with the life cycle of picornavirus-like viruses but not of FHV , a nodavirus . We cannot distinguish between these possibilities with such a small sample of viruses; it would be interesting to extend the analysis to other RNA viruses that infect D . melanogaster ( e . g . , Sigma ( a rhabdovirus ) [43] , Drosophila X virus ( a birnavirus ) [55] , and Drosophila A virus ( picornavirus-like ) [43] ) . We have also tested the interaction of Wolbachia with a DNA virus , IVV-6 . Wolbachia did not protect Drosophila from this virus; it actually decreased the lifespan of infected flies . We think this is due to the cumulative effect of Wolbachia and IIV-6 infection , since , in the genetic background of the flies we were using , Wolbachia had a negative effect on long-term survival . It would be interesting to also extend the analysis to other DNA viruses , however there are no DNA viruses known to infect D . melanogaster . To our knowledge , this is the first report of a DNA virus proliferating in adults of D . melanogaster . An obvious question is how Wolbachia induces resistance to RNA viruses . The different effect on DCV/Nora virus and FHV raises the possibility that this effect is multifactorial; interfering with virus replication in some cases and increasing resistance of Drosophila to viral infection damage in others . One important question to address is whether the effect is cell-autonomous or systemic . Wolbachia is widespread throughout tissues of the infect host [18 , 56] , so both hypotheses are possible . This could be investigated in tissue culture with Wolbachia-infected cells . If the effects are cell autonomous , one explanation for increased resistance to viruses could just be competition for resources , since both microorganisms occupy the same niche , the host's cytoplasm . For example , Wolbachia is thought to acquire much of its energy from the metabolism of amino acids imported from the host cytoplasm [46] . DCV , on the other hand , is very sensitive to perturbations in host translation [57] . The presence of Wolbachia could reduce the pool of cytoplasmic amino acids to a point that interferes with translation of viral proteins . Another possibility is that Wolbachia infection could trigger cell-autonomous mechanisms of resistance to intracellular pathogens , such as a reduction in cellular metabolism . A further explanation for a cell-autonomous effect would be that Wolbachia has been selected to actively interfere with virus replication in co-infected cells . Wolbachia has a complete type IV secretion system [46] , which many bacteria use for translocation of effector molecules into host cells ( e . g . , Legionella and Agrobacterium ) . Genes encoding proteins with ankyrin repeats , involved in protein–protein interactions , are over-represented in the Wolbachia genome [46 , 58] and are good candidates for mediators of anti-viral resistance . If the effect is systemic , a strong hypothesis is that Wolbachia could alter the host–immune response , increasing resistance to viral infection . The pre-activation of the host immune system , for example , could allow for a faster response upon viral infection . This would be similar to what happens in a herpesvirus-induced resistance to Listeria in mice , due to the production of cytokines [59] . It was also reported recently that the presence of gut flora slightly increases the resistance of Aedes aegypti to Dengue virus , presumably through activation of the Toll pathway [60] . In tissue culture of D . melanogaster cells , infection with Wolbachia slightly increases the expression of innate immune genes [61] . There is also a report that Wolbachia increases resistance of D . melanogaster to the pathogenic fungus Beauveria bassiana [62] . All these reports support a model of general activation of innate immunity . However it has also been shown that in adult D . simulans and Aedes albopictus Wolbachia does not activate the expression of anti-microbial peptides [63] , in D . simulans , Wolbachia infection does not alter sensitivity to Beauveria and renders the host more sensitive to parasitoid wasps [64] , and in D . melanogaster , Wolbachia presence does not affect Spiroplasma levels [65] . In summary , it is not clear if there is a general activation of innate immunity in adult D . melanogaster infected with Wolbachia that would render them more resistant to other pathogens . It would be interesting to identify immune pathways involved in anti-viral resistance activated by Wolbachia infection . It would also be important to analyze Wolbachia-induced resistant to other microorganisms , including pathogenic bacteria . A different hypothesis would be that Wolbachia infection actually inhibits some of the immune responses against viral infection and that increases the lifespan of infected D . melanogaster . This may be true if the host response to infection damages the host itself , as in the case of septic shock in mammals . This could explain the increased resistance to FHV infection without a strong effect on viral titers . Finally , a similar hypothesis would be that Wolbachia inhibits , cell-autonomously or systemically , apoptosis induced upon viral infection . Some published data support this hypothesis; FHV induces apoptosis in tissue culture cells [66] , Wolbachia inhibits apoptosis in the germline of Asobara tabida [8 , 9] , and the Wolbachia protein Wsp inhibits apoptosis in human cells [67] . This new host-microorganism-microorganism interaction adds to the perception that the response of a host to a particular pathogen also depends on its interactions with other microorganisms . Other examples are herpesvirus latency-induced protection to Listeria in mice mentioned above [59] , the suppression of HIV-1 infection by human herpesvirus 6 in human cells [68] , symbiotic bacteria protection against fungi in a shrimp and an aphid [69 , 70] , symbiotic bacteria protection against parasitic wasps in an aphid [71] , and symbiotic bacteria protection against fungal infection in a wasp [72] . As also mentioned above , there is a recent report that gut flora has a protective role against Dengue virus in A . aegypti [60] . However , this is , to our knowledge , the first report where bacteria that confer protection against viruses have been identified . This interaction has some practical consequences . Researchers working on Drosophila immunity against viruses should take in consideration the presence of Wolbachia in the stocks they are analyzing . On the other hand , researchers working on Wolbachia should consider that any observed effects of Wolbachia could be mediated through effects on viral infections . A practical application of this discovery would be , if possible , to induce resistance to viruses , by infection with Wolbachia , in insects that are beneficial to humans ( e . g . , honeybee ) or transmit arboviruses ( e . g . , mosquitoes ) . However , introducing Wolbachia to virus-transmitting vectors could be a double-edged sword . If the interaction Wolbachia-vector-virus were similar to the one seen in this report with DCV , then it would be beneficial because it could decrease the probability of the vector being infected or transmitting the disease . If , however , it were similar to the interaction with FHV , then there would be the risk of having healthier infected vectors with high titers of viruses , therefore increasing disease transmission . This latest possibility should be taken into account in proposed strategies of introducing Wolbachia in vectors of arboviruses [73 , 74] . Finally , this is , to our knowledge , the first report of a strong beneficial effect of Wolbachia infection in D . melanogaster . The induced resistance to natural viral pathogens may explain the prevalence of Wolbachia in natural populations . It also indicates that the fitness benefit of having Wolbachia is dependent on the viral infection status of the population . This may explain differences in Wolbachia infection frequencies between populations [17 , 25] and variable fitness effects in different D . melanogaster lines [28 , 30] . It would be interesting to broaden the analysis to other Wolbachia strains and to other Wolbachia–host combinations . If Wolbachia induces resistance to viruses in other hosts , this would have major implications for our understanding of the very widespread presence of this endosymbionts in arthropods and filarial nematodes . After the submission of this manuscript , an independent report with similar findings to ours was published [75] . In agreement with our data , Hedges et al . show that the treatment of Wolbachia-infected flies with tetracycline renders them more sensitive to three RNA viruses: DCV , Cricket Paralysis virus , and FHV . Moreover , they also show that the levels of DCV increase in infected Wolbachia-free flies .
The set of w1118 iso isogenic flies were obtained from the DrosDel collection in our laboratory [42] . These lines were cleaned of viruses similarly to the protocol in Brun and Plus , 1978 [43] . Flies were aged to 30 d at 25 °C and their eggs were collected in agar plates , treated with 50% bleach for 10 min , washed with water , and transferred to fresh vials . The wild-type laboratory lines used in Figure 3B have the origins described in Table 1 . Stocks were treated with tetracycline ( cleaned of Wolbachia infection ) by raising them for two generations in ready-mix dried food ( Philip Harris ) with 0 . 05 mg/ml of tetracycline hydrochloride ( Sigma ) . Sub-optimal tetracycline treatment was done by raising flies , for one generation , in food with 0 . 00625 mg/ml of tetracycline hydrochloride . Virgin adult females , emerging from these vials , were collected and individually crossed with males to establish isofemale lines . Sensitive and resistant-to-viruses flies were raised together by placing in a vial one w1118 iso female , one VF-0058–3 female , and two w1118 iso males . The progeny of the two different females could be distinguished by the eye color , because only the progeny of VF-0058–3 females have a functional white gene . Flies were only collected from vials that had adults of both phenotypes . DCV-C [77] was kindly provided by Dr . Peter Christian and raised as in Johnson and Christian 1999 [78] . Nora virus extract was prepared from a naturally infected Oregon stock present in the laboratory . Thirty adult flies were squashed in 900 μl of 50 mM Tris-HCl , pH 7 . 5 . Extract was then frozen at −80 °C , thawed and twice centrifuged for 10 min at top speed in a tabletop centrifuge , at 4 °C . The supernatant was aliquoted and stored at −80 °C . IIV-6 [79] was kindly given by Dr . Peter Christian and raised in Schneider Drosophila line 2 ( DL2 ) cells [54] . DL2 cells were kept in Schneider's Drosophila Medium ( Invitrogen ) supplemented with 10% Fetal Bovine Serum , 2mM L-Glutamine , 100 U/ml penicillin , and 100μg/ml streptomycin ( all Invitrogen ) . Seven days after infection , the cell culture was collected and frozen at −80 °C . The culture was thawed , frozen , and thawed again to disrupt cells and centrifuged twice at 600g for 10 min to remove cell debris . Virus was pelleted by centrifugation at 10 , 000g for 10 min and re-suspended in water . Virus suspension was placed over a 30% sucrose solution and the virus was pelleted again by centrifugation at 30 , 000g for 30 min and re-suspended in water . The virus was then pellet by centrifugation at 15 , 000g for 10 min and re-suspended twice . The final re-suspension was done in 50 mM Tris-HCl , pH 7 . 5 , aliquoted , and stored at −80 °C . All centrifugations and re-suspensions were done at 4 °C . FHV was kindly provided by Dr . J . -L . Imler [38] and dilutions of this aliquot were used in this work . Three-to-six–d-old flies were injected with a Nanoject II injector ( Drummond ) . Viruses were re-suspended or diluted in 50 mM Tris-HCl , pH 7 . 5 , and 69 nl of virus solution was injected , per fly , in the thorax , between the mesopleura and the pteropleura . Flies were injected while anesthetized with CO2 . Fifty flies were injected per sample , ten flies were placed per vial , and vials were changed twice a week . DCV injected flies were kept at 18 °C . Nora virus– , FHV- and IIV-6–injected flies were kept at 25 °C . Flies were counted daily for all survival curves except for VF-0058–3 and VF-0058–3t injected with buffer , as shown in Figure 4E , which were counted at least twice a week . Five flies were pooled per sample . Flies were squashed in 50 mM Tris-HCl , pH 7 . 5 , frozen , thawed , and centrifuged for 10 min at 20 , 000g and supernatant was collected ( DCV and FHV ) . For IIV-6 , centrifugation was done twice at 600g and the supernatant passed through a 0 . 45-μm filter before the assay . Viruses titers were determined in cell culture and calculated by the Reed and Muench end-point calculation method [80] . DL2 cells in 96-well plates were infected with the serial dilutions of virus suspensions . DCV and FHV infection was scored by the presence of cell death , IIV-6 was scored by non-proliferation of cells and presence of very large cells . Extracts of non-infected VF-0058–3 or VF-0058–3t flies did not cause any cytopathic effect in tissue culture cells . Five to eight males were pooled per sample . Rabbit polyclonal antibodies raised against purified DCV was kindly given by Dr . Peter Christian . Rabbit polyclonal antibodies raised against FHV capsids was kindly given by Dr . Jean-Luc Imler [38] . Specificity of antibodies was verified by lack of signal on Western blot lanes of non-infected flies ( Figures 1E and 4C ) . E7 mouse monoclonal anti-β-tubulin was acquired from Developmental Studies Hybridoma Bank [81] . Embryos 0–2-h-old were collected , treated with 50% commercial bleach for 10 min , fixed for 30 min in 4% formaldehyde , 50% heptane , the vitelline membranes were removed by vortexing the embryos in 50% heptane , 50% methanol , then embryos were washed briefly in methanol and for 10 min in 50% methanol , 50% PBS and finally placed in PBS 0 . 1% tween-20 . Embryos were treated with RNAse H 0 . 25 μg/μl for 30 min at 37 °C , washed in PBS 0 . 1% tween-20 , stained with PBS 0 . 1% tween-20 and 1 μg/ml propidium iodide ( Sigma ) for 30 min , washed in PBS 0 . 1% tween-20 and mounted in Vectorshield . Images were taken in a confocal microscope . A fragment of bacterial 16S rRNA gene was amplified from DNA of Drosophila embryos surface sterilized by treatment with 50% commercial bleach for 10 min . DNA was extracted using Wizard Genomic DNA purification kit ( Promega ) . Primers used were 27f ( 5′-GAGAGTTTGATCCTGGCTCAG-3′ ) and 1495r ( 5′-CTACGGCTACCTTGTTACGA - 3′ ) . The PCR program was: 94 °C for 4 min; 25 cycles of 94 °C for 30 s , 58 °C for 1 min , and 72 °C for 2 min; 72 °C for 10 min . The PCR product was ligated into pCR 2 . 1 TOPO vector ( Invitrogen ) and transformed into DH5α cells . Nineteen plasmid DNA preparations and 96 bacteria cultures were sent for sequencing . Three sequencing reactions failed and 8 clones did not carry an insertion in the cloning plasmid . From all the other 104 sequences , we selected a sequence of at least 600 bp with good quality and aligned it with a fragment of the 16S rRNA gene of Wolbachia ( GenBank accession number EU096232 ) using Clustal W 2 [82] . PCR amplification of Wolbachia-specific genes was done either on the DNA extracts of embryos as described above ( Figure 2E ) or on DNA extracts of adult flies ( Figure 3 ) . Adult flies were squashed in 25 mM NaCl , 10 mM Tris-HCl pH=8 . 0 , 1 mM EDTA , 200 μg/ml proteinase K and incubated for 30 min at 37 °C . Proteinase K was inactivated at 95 °C for 5 min . The supernatant was directly used for PCR amplification . wsp primers were wsp 81F ( 5′-TGGTCCAATAAGTGATGAAGAAAC-3′ ) and wsp 691R ( 5′-AAAAATTAAACGCTACTCCA-3′ ) [45] . wspB primers were wspB-F ( 5′-TTTGCAAGTGAAACAGAAGG-3′ ) and wspB-R ( 5′-GCTTTGCTGGCAAAATGG-3′ ) [46] . As a positive control for cytoplasmic DNA extraction we used the primers for mitochondrial 12S rRNA , 12SAI ( 5′-AAACTAGGATTAGATACCCTATTAT-3′ ) and 12SBI ( 5′-AAGAGCGACGGGCGATGTGT-3′ ) [4] . The PCR program used was: 94 °C for 4 min; 30 cycles of 94 °C for 1 min , 55 °C for 1 min , and 72 °C for 1 min; 72 °C for 10 min . The wsp primers amplification product from VF-0058–3 embryos was purified , as described above , and sequenced . The sequence obtained was identical to a fragment of the wsp sequence of wMel ( GenBank accession number DQ235407 ) . PCR amplification with primers specific for Spiroplasma 16S rRNA gene was done on DNA extracts of adult flies . Primers used were SpoulF ( 5′-GCTTAACTCCAGTTCGCC-3′ ) and SpoulR ( 5′-CCTGTCTCAATGTTAACCTC-3′ ) [48] . The PCR program was: 94 °C for 4 min; 30 cycles of 94 °C for 30 s , 55 °C for 1 min , and 72 °C for 1 min; 72 °C for 10 min . As a positive control for cytoplasmic DNA extraction , we used the primers for mitochondrial 12S rRNA , as described before . Nora virus presence in VF-0058–3 , VF-0058–3t , and Oregon R stocks was analyzed by RT-PCR . RNA of 100 flies , per sample , was extracted using Trizol ( Invitrogen ) . cDNA was synthesized from 5 μg of total RNA with Superscript III ( Invitrogen ) , using random hexamers primers , at 50 °C . The Nora primers used for PCR were Nora-F ( 5′-TTAAGGTGTTAGAGAACAGC-3′ ) and Nora-R ( 5′-CGTAAACACCAACTTACTTC-3′ ) [49] . RpL32 primers , used as a positive control for RNA extraction , were RpL32-F ( 5′-TCCTAC CAGCTTCAAGATGAC-3′ ) and RpL32-R ( 5′-CACGTTGTGCACCAGGAACT-3′ ) . The PCR program used was as follow: 94 °C for 4 min; 10 cycles of 94 °C for 30 s , 60 °C minus 0 . 6 °C per cycle for 1 min and 72 °C for 1 min; 20 cycles of 94 °C for 30 s , 54 °C for 1 min , and 72 °C for 1 min; 72 °C for 10 min . The PCR amplification fragment obtained with the Nora primers was purified , as described above , and sequenced . The sequence was 98% identical to a fragment of Nora virus genome sequence ( GenBank accession number DQ321720 ) . For semi-quantitative analysis of Nora virus in infected flies , the procedure was as above except that 25 flies were used per sample and the PCR amplification for each sample was done with a total of 20 , 25 , and 30 cycles .
|
Many symbiotic bacteria confer fitness benefits to the organisms that they infect . Wolbachia are one of the most widespread intracellular bacteria , infecting a great number of species of insects . Here we show that in the fruit fly Drosophila melanogaster , infection with Wolbachia increases resistance to a natural pathogen of Drosophila , an RNA virus called Drosophila C virus . Furthermore , we show that Wolbachia also increases resistance of Drosophila to two other RNA viruses ( Nora and Flock House virus ) but not to a DNA virus ( Insect Iridescent Virus 6 ) . These results identify a significant new factor that regulates D . melanogaster resistance to infection by RNA viruses . Our results add to a growing body of literature showing that the response of an organism to a particular pathogen is modulated by prior or contemporaneous interactions with other microorganisms . That the fruit fly clearly benefits from increased resistance to viruses may provide a solution to the longstanding puzzle as to why Wolbachia is so common in natural populations of D . melanogaster .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"infectious",
"diseases",
"virology",
"immunology"
] |
2008
|
The Bacterial Symbiont Wolbachia Induces Resistance to RNA Viral Infections in Drosophila melanogaster
|
Biofilm formation is an important virulence trait of the pathogenic yeast Candida albicans . We have combined gene overexpression , strain barcoding and microarray profiling to screen a library of 531 C . albicans conditional overexpression strains ( ∼10% of the genome ) for genes affecting biofilm development in mixed-population experiments . The overexpression of 16 genes increased strain occupancy within a multi-strain biofilm , whereas overexpression of 4 genes decreased it . The set of 16 genes was significantly enriched for those encoding predicted glycosylphosphatidylinositol ( GPI ) -modified proteins , namely Ihd1/Pga36 , Phr2 , Pga15 , Pga19 , Pga22 , Pga32 , Pga37 , Pga42 and Pga59; eight of which have been classified as pathogen-specific . Validation experiments using either individually- or competitively-grown overexpression strains revealed that the contribution of these genes to biofilm formation was variable and stage-specific . Deeper functional analysis of PGA59 and PGA22 at a single-cell resolution using atomic force microscopy showed that overexpression of either gene increased C . albicans ability to adhere to an abiotic substrate . However , unlike PGA59 , PGA22 overexpression led to cell cluster formation that resulted in increased sensitivity to shear forces and decreased ability to form a single-strain biofilm . Within the multi-strain environment provided by the PGA22-non overexpressing cells , PGA22-overexpressing cells were protected from shear forces and fitter for biofilm development . Ultrastructural analysis , genome-wide transcript profiling and phenotypic analyses in a heterologous context suggested that PGA22 affects cell adherence through alteration of cell wall structure and/or function . Taken together , our findings reveal that several novel predicted GPI-modified proteins contribute to the cooperative behaviour between biofilm cells and are important participants during C . albicans biofilm formation . Moreover , they illustrate the power of using signature tagging in conjunction with gene overexpression for the identification of novel genes involved in processes pertaining to C . albicans virulence .
Candida albicans is the most predominant human fungal pathogen , causing both superficial and hematogenously disseminated infections [1] . These infections are complicated by C . albicans' ability to form biofilms , which are complex three-dimensional microbial structures attached to either biotic or abiotic surfaces and encased in an extracellular matrix [2]–[5] . Biofilms play a crucial role in C . albicans virulence as they result in decreased susceptibility to both antimicrobial agents and the host immune system [2] , [5]–[7] . C . albicans biofilms are composed of yeast and hyphal cells , and the ability to switch between these morphotypes is essential for normal biofilm formation [8]–[10] . Additional understanding of the mechanisms of biofilm formation in C . albicans has been gained over recent years with the discovery of various regulators and effectors involved in this process ( reviewed in [11] ) . In this respect , several cell wall proteins have been shown to play crucial roles during biofilm formation . For instance , the Bcr1 transcription factor , required for biofilm formation , was shown to control the expression of genes encoding cell wall proteins , among which the ALS3 , ALS1 , and HWP1 genes contribute to biofilm formation and integrity [12]–[14] . Heterotypic interactions between Als1 and Als3 , members of the Als family of glycophosphatidylinositol ( GPI ) -anchored agglutinin-like cell wall proteins , and the hyphal wall protein Hwp1 , appear crucial for cell-cell interactions within biofilms [15] . Other GPI-anchored proteins play positive or negative roles at different stages of biofilm formation , such as Ywp1 ( Pga24 ) , Eap1 ( Pga47 ) , Pga26 , Pga1 , and members of the CFEM family ( Pga10 , Rbt5 and Csa1 ) [16]–[21] . To date , the investigation of molecular determinants of biofilm formation in C . albicans has largely relied on phenotypic analyses of loss-of-function mutants for genes predicted to play a role in this process , based on their expression profile , function or cellular location [12] , [22]–[25] . Gene overexpression is an alternative strategy for studying gene function . It mimics gain-of-function mutations , provides a complement to loss-of-function phenotypes and allows the role of both essential and non-essential genes or individual genes within multi-gene families to be studied [26] . Gene overexpression has been successfully used in Saccharomyces cerevisiae to reveal new signalling pathways [27] and identify transcription factor targets [28] . More recently , overexpression approaches in C . albicans identified genes involved in fitness , adherence , morphogenesis , pheromone response and antifungal resistance [29]–[33] as well as the characterization of transcription factor targets [13] , [23] , [34] . To date , the largest collection of overexpression plasmids that exists for C . albicans genes has been developed in our laboratory [30] . This collection includes 337 uniquely barcoded plasmids allowing tetracycline-inducible overexpression of genes encoding components of signalling networks , in particular protein kinases , protein phosphatases and transcription factors [30] . Here , we have extended this plasmid collection to include genes encoding C . albicans predicted cell wall proteins and genes involved in genome dynamics . We took advantage of the molecular barcoding of the cognate overexpression strains to develop a signature-tagged overexpression ( ST-OE ) screen aimed at identifying genes whose overexpression affects fitness during planktonic cell growth and/or biofilm development in C . albicans . Our results specifically highlight the impact of overexpressing specific cell surface GPI-modified proteins on the ability of C . albicans to form single- or multi-strain biofilms and reveal their role in adhesion and/or cell-cell interactions .
We generated a collection of signature-tagged C . albicans conditional overexpression strains using partial C . albicans ORFeome libraries ( [30]; Legrand et al . , in preparation; Walker et al . , in preparation ) and a collection of barcoded derivatives of the conditional overexpression plasmid CIp10-PTET-GTW ( [30]; Fig . 1A , See Materials and Methods for details ) . The resulting collection includes 531 strains conditionally overexpressing individual genes encoding ( or predicted to encode ) transcription factors ( 180 ORFs ) , protein kinases ( 72 ORFs ) , protein phosphatases ( 34 ORFs ) , proteins related to DNA replication , recombination and repair ( 87 ORFs ) , predicted cell surface proteins ( 61 ORFs ) and others ( Fig . 1B; S1 Table for details ) . We tested the efficiency of detecting each of the 531 tagged strains from a pool with equal strain representation using an oligonucleotide microarray carrying both forward and reverse probe sequences for every barcode ( S1 Figure and S1 Table , see Materials and Methods ) . The microarrays efficiently detected nearly 90% of the pooled strains with only 11 . 6% of total strain-matching probes ( 123 out of 1062 ) having log2-transformed Cy5 and Cy3 signal intensities lower than 8 ( S1 Figure ) , suggesting that just over 10% were underrepresented or that the corresponding tags had low hybridization efficiency . Strain detection with the microarrays was highly reproducible , as shown by the high Pearson correlation coefficients between 2 independent pool replicates ( S1B Figure ) . As a proof-of-principle for the ST-OE approach , we evaluated the impact of overexpressing each of the 531 genes on planktonic strain fitness . Pooled strains were grown for 16 generations in GHAUM medium ( see Materials and Methods ) at 30°C in the presence or absence of 50 µg . mL-1 doxycycline ( Dox ) , with dilution in fresh medium after 8 generations to avoid saturation of the culture . Genomic DNA was prepared , followed by barcode amplification , labeling with Cy3 ( untreated ) or Cy5 ( Dox-treated ) dyes and hybridization to the barcode microarray . We found that overexpression of 5 genes , namely RAD53 , RAD51 , PIN4 , SFL2 and ORF19 . 2781 , decreased strain fitness using a Z-score ( i . e . number of standard deviations from the population mean ) cutoff of −2 . 0 ( Fig . 2A and S2 Table ) . Notably , we failed to detect any gene conferring increased fitness ( Fig . 2A ) . Liquid growth assays of individual strains grown independently confirmed an increased doubling time in Dox-treated cells overexpressing RAD53 , RAD51 and ORF19 . 2781 , relative to untreated cells ( Fig . 2B ) , in agreement with our microarray data . As morphology alterations affect turbidity measurements , the SFL2-overexpressing strain which shows extensive filamentation in overexpression conditions [30] , [35]–[37] was omitted from this assay . Taken together , these data indicated that our ST-OE approach could be used to identify genes whose overexpression affects C . albicans fitness . We used the ST-OE approach to identify C . albicans genes whose overexpression alters strain abundance within a multi-strain biofilm . To this end , the 531-strain pool was grown overnight in GHAUM medium at 30°C in the presence or absence of 50 µg . mL−1 doxycycline . Cells were allowed to adhere to Thermanox slides that were subsequently incubated at 37°C for 40 h in a microfermentor under a continuous flow of GHAUM medium in the presence or absence of 50 µg . mL−1 doxycycline . Genomic DNA was extracted from the resulting mature biofilms followed by barcode amplification , Cy3 ( untreated ) /Cy5 ( Dox-treated ) labeling and hybridization to barcode microarrays . The entire experiment was performed using 8 biological replicates and two independent analyses of our microarray data were performed ( See Materials and Methods for details ) . We found 20 genes whose overexpression altered strain abundance in the multi-strain biofilm . Only one gene ( ORF19 . 2781 ) was common to both planktonic and biofilm data , suggesting that the remaining 19 genes were not linked to a growth defect ( Table 1 , Fig . 3 and S3 Table ) . Overexpression of 16 out of these 20 genes resulted in increased strain abundance within the multi-strain biofilm ( Table 1 ) . A gene ontology ( GO ) term enrichment analysis was performed using the GO Term Finder at the Candida Genome Database [38]–[40] and the 531 gene set as a reference . Strikingly , among the 20 identified genes we found a significant enrichment of the GO term “cell wall” ( P = 1 . 4×10−5 ) including 10 genes encoding ( or predicted to encode ) proteins involved in cell wall biogenesis and integrity ( IHD1/PGA36 , PGA15 , PGA19 , PGA22 , PGA32 , PGA37 , PGA42 , PGA59 , PHR2 and TOS1 , Table 2 ) . Similarly , when the enrichment analysis was performed using broad functional categories , we observed a significant enrichment for cell surface protein-encoding genes ( P = 1 . 31×10−5; Table 2 ) . Overexpression of these 10 genes did not significantly promote or inhibit morphogenesis when compared to strain SC5314 ( S2A Figure ) . Moreover , their overexpression did not significantly impact growth rates as judged by colony size on solid GHAUM medium at 30°C and 37°C ( S2B Figure ) . Thus , our data suggested that the increased occupancy of the multi-strain biofilm by the selected strains was not a consequence of altering growth rate or morphogenesis . The selected overexpression strains were tested for their ability to form single strain biofilms under conditions similar to those used for our mixed-population screen . Biofilms were grown in the continuous-flow microfermentor system . The biomass obtained after 40 h of biofilm growth was quantified and compared to the biomass formed by the wild type strain SC5314 grown under the same conditions . As shown in Fig . 4A , the strain overexpressing PGA59 displayed increased biomass , consistent with our mixed-population data . Decreased biofilm biomass was observed upon PGA22 overexpression ( Fig . 4A ) , contrasting with the positive impact of overexpressing this gene on strain abundance in the multi-strain biofilm ( Fig . 3 and Table 1 ) . Finally , overexpression of the remaining genes did not significantly affect biofilm biomass ( Fig . 4A ) . We further compared biofilm formation of the wild-type and the PGA59- and PGA22-overexpression strains during a kinetic experiment in the continuous-flow fermentor system under inducing conditions . The biofilm biomass was quantified at 18 , 24 , 40 , 48 , and 65 h of biofilm growth . The overexpression strains behaved similarly until the first time-point ( 18 h; Fig . 4B ) . The PGA59-overexpression strain formed a biofilm faster than the remaining strains from 18 h onwards ( Fig . 4B , upper panel ) . In contrast , the PGA22-overexpression strain developed into biofilms more slowly than the wild-type strain ( Fig . 4B , lower panel ) . Overall , these results were consistent with those obtained in the end-point analysis ( Fig . 4A ) and suggested that the behavior of the PGA22-overexpression strain differed when grown in single- or multi-strain biofilms . Because of the discrepancy observed for the PGA22-overexpression strain in single- or multi-strain biofilms , we assessed whether the increased abundance of PGA22- or PGA59-overexpression strains in a multi-strain biofilm could be recapitulated in a biofilm formed by either of these strains together with a control strain . The PGA22- or PGA59-overexpression plasmids were introduced into a C . albicans strain that constitutively expressed GFP ( CEC3781 , S4 Table ) . A strain constitutively expressing BFP ( CEC3783 , S4 Table ) was used as a control . A 1∶1 mixture of the GFP-tagged PGA22- or PGA59-overexpressing strains and the BFP-expressing control strain was used as an inoculum , and biofilms grown for 40 h in the absence or presence of doxycycline . Genomic DNA was isolated from each biofilm , and strain abundance quantified by qPCR using GFP and BFP specific primers . In biofilms grown under non-inducing conditions , the overexpression and control strains were equally represented ( Fig . 5 ) . By contrast , both PGA22- and PGA59-overexpression strains were significantly more abundant in the respective biofilms grown under inducing conditions ( 64% vs . 50% for the PGA22-overexpression strain and 65% vs . 50% for the PGA59-overexpression strain , with or without doxycycline , respectively; Fig . 5 ) . Consistently , confocal microscopy-acquired fluorescence images showed more GFP-labeled cells relative to BFP cells in the doxycycline-treated mixed biofilms as compared to untreated controls ( S3 Figure ) . A similar increase was observed when BFP-expressing PGA22- or PGA59-overexpression strains were grown in mixed-biofilms with a GFP-expressing control strain ( S4 Figure ) . Hence , overexpression of PGA22 and PGA59 resulted in increased strain abundance in a mixed biofilm-population . Biofilm formation is a multi-stage process that is initiated upon adherence of C . albicans yeast cells to a substrate and reinforced through cell-to-cell interactions ( reviewed in [3] ) . Differences observed for the PGA22- and PGA59-overexpression strains upon single- or multi-strain biofilm formation might reflect differences in adherence to the substrate or to other biofilm cells . Therefore , strains overexpressing PGA22 or PGA59 as well as the other cell wall-related genes identified in our screen were individually tested for their ability to adhere to Thermanox following overnight growth in GHAUM medium at 30°C in the presence or absence of doxycycline . Modest , non-significant , variations in the adherence of the ten strains were observed when they were grown under non-inducing conditions ( S5 Figure ) . In contrast , overexpression of IHD1/PGA36 , PGA15 , PGA22 and PGA59 significantly increased adherence of C . albicans to Thermanox , while overexpression of PGA19 , PGA32 and PGA37 decreased it ( Fig . 6 ) . The increased adherence phenotype was substrate-independent , as shown in an adhesion assay of the PGA22- and PGA59-overexpressing strains using a polystyrene substrate ( microtiter plate , S6 Figure ) . Results presented above indicated that over-expression of PGA59 and PGA22 increased adherence to abiotic surfaces and yet , this could have different outcomes during biofilm formation . While the PGA59-overexpression strain showed an increased ability to form biofilms in single- and multi-strain biofilms consistent with increased adherence to the substrate , the PGA22-overexpression strain showed decreased or increased ability to form biofilms when grown alone or in combination , respectively . We reasoned that the differences observed for the PGA22-overexpression strain might occur at early time points during biofilm development and therefore , examined the fate of cells overexpressing PGA22 once adhered to Thermanox and exposed to a flow of medium in the microfermentor system . To this aim , the PGA22-overexpression strain was grown overnight in the presence or absence of doxycycline , allowed to adhere to Thermanox and incubated in the microfermentor system in the presence or absence of doxycycline for 2 h . Cells attached to the Thermanox at t = 0 h and t = 2 h and those that were released during the 2 h of biofilm growth in the continuous-flow fermentor system were quantified by microscopy or by flow cytometry ( see Materials and Methods ) . At t = 0 h , overexpression of PGA22 resulted in increased adherence ( Fig . 7A ) ; however , at t = 2 h , less cells remained attached to the slide ( Fig . 7B ) , correlating with an increase of released cells during this 2 h period ( Fig . 7C ) . Moreover , careful observation of Thermanox slides under the microscope revealed that upon overexpression of PGA22 , cells tended to adhere in clusters of more than 3 cells ( Fig . 7D ) . Under a more static environment ( i . e . in the absence of flow ) , overexpression of PGA22 did not decrease biofilm formation and showed a tendency to form more biofilm ( S7 Figure ) . In this environment the PGA59-overexpression strain still formed more biofilm ( S7 Figure ) . This is in line with the effect of flow in washing out cells adhered to the substrate and consequently affecting biofilm biomass . We further analyzed our confocal microscopy-acquired fluorescence data from the competitive biofilm growth assay under continuous flow of the GFP-labeled strains overexpressing PGA22 and PGA59 versus the parental control strain ( BFP-labeled , S3 Figure ) . We quantified abundance of GFP- versus BFP-labeled cells within the bottom layer of the mature biofilm , where early events such as adhesion occur . In the bottom layer of the biofilm , the PGA22-overexpressing cells were less abundant as compared to the upper layer ( S8 Figure , compare panels A and B ) . In contrast , PGA59-overexpressing cells were more abundant within both the bottom and upper layers of the mature biofilm ( S8 Figure , compare panels C and D ) . Taken together , our results suggested that the apparent increase in adherence of the PGA22-overexpression strain was a consequence of increased cell aggregation that rendered adhered cells more susceptible to shear forces occurring in the microfermentor . This may explain why PGA22 overexpression is detrimental to single-strain biofilm formation but favorable in a potentially protective multi-strain biofilm . In contrast , increased adherence of individual cells of the PGA59-overexpression strain to the surface was directly correlated to increased biofilm formation whether alone or in combination . To further investigate the effect of PGA22 and PGA59 overexpression on the adherence of single C . albicans cells , we performed adhesion force measurements using Atomic Force Microscopy ( AFM , Fig . 8 ) . The PGA22- and PGA59-overexpression strains were grown for 16 h in the presence or absence of 50 µg . mL−1 doxycycline in YPD medium and subjected to an adhesion force measurement assay between the AFM tip , composed of Si3N4 , and the cell surface ( Fig . 8 ) . We generated adhesion maps ( Fig . 8 , left panels ) where the intensity of each pixel corresponds to the force required to dissociate the AFM tip from the sample ( i . e . adhesion force ) . Upon doxycycline treatment , adhesion events were detected in 76% of the recorded force curves for PGA22 overexpression ( Fig . 8 , + Dox , upper panels ) , versus less than 3% without treatment ( Fig . 8 , -Dox , upper panels ) . We detected only 5% of adhesion events in the recorded force curves on the parental control strain ( vector only , S9 Figure ) . The mean force of PGA22 overexpression-mediated adhesion events was 1 . 21 nN±0 . 55 nN ( Fig . 8 ) . The PGA59-overexpressing cells displayed a stronger surface adhesion rate upon induction by doxycycline ( Fig . 8 , lower panels , PGA59 , + Dox ) . Both frequency of cell surface-tip adhesion events ( 82% ) and adhesion forces ( ranging from 1 up to 5 nN ) were higher than those observed for PGA22 overexpression ( Fig . 8 , lower panels ) . Taken together , our AFM analyses indicated that PGA22 and PGA59 overexpression imparted significant adhesion forces to single C . albicans cells . The C . albicans cell wall is characterized by an inner layer containing the skeletal polysaccharides chitin , β-1 , 3-glucan and β-1 , 6-glucan , and a fibrillar outer layer enriched with O-linked and N-linked mannose polymers ( mannans ) covalently associated with proteins . The major class of cell wall proteins are GPI-modified proteins attached to the β-1 , 3-glucan skeleton by β-1 , 6 linkages [41] , [42] . We reasoned that overexpressing predicted GPI-anchored proteins could result in an abnormal cell wall structure , leading to modified adherence to the substrate and/or other cells . The cell wall architecture of strains overexpressing PGA22 and PGA59 were analyzed using transmission electron microscopy ( TEM ) . Results presented in Fig . 9 showed that the PGA22-overexpressing strain displayed a thinner outer fibrillar layer as compared to the control strain and the PGA59-overexpressing strain . In contrast , the inner cell wall thickness of these strains was similar ( Fig . 9B ) . Alterations in the cell wall structure of the PGA22-overexpression mutant may contribute to its modified ability to bind Thermanox and/or aggregate . To better understand how PGA22 overexpression affects the cell wall structure , we additionally performed transcript profiling of the PGA22-overexpression strain under the same growth conditions used for the TEM analysis ( Fig . 9 ) . Three independently grown PGA22-overexpression strains were treated or not with 50 µg . mL−1 doxycycline for 16 h followed by total RNA extraction , reverse transcription , labeling and hybridization to a custom-designed C . albicans ORF microarray that was described previously [36] . Analysis of the control strain ( see below ) indicated that doxycycline treatment did not affect gene expression under these conditions ( S10 Figure ) . Using a fold-change cut-off of 1 . 5 and a p-value threshold of ≤0 . 05 , 37 genes were upregulated upon PGA22 overexpression ( Table 3; see Materials and Methods for details and S5 Table for complete transcript profiling data ) . A significant proportion of cell wall related genes were among the induced genes , including genes encoding ( or predicted to encode ) GPI-anchored proteins ( PGA29/RHD3 , CRH11 ) , a chitinase ( CHT3 ) and a chitin synthase ( CHS1 ) , cell wall proteins ( RBE1 , SCW1 ) as well as putative adhesins ( PGA35/FGR41 , PGA38 and ORF19 . 5267 ) ( Table 3 ) . Gene Ontology term enrichment analysis revealed a strong overrepresentation of the term “Cell wall” ( p = 0 . 001 ) among the transcriptionally induced genes . On the other hand , although 100 genes were significantly downregulated ( fold-change <−1 . 5; p-value <0 . 05; S5 Table ) , no significant gene ontology enrichment was found . However , we noticed the presence of genes encoding mannosyltransferases ( MNN12 , MNT1 , KTR4 , RHD1 ) and genes linked to or affecting mannosyltransferase activity ( VRG4 , SMF12 , SKN1 ) in the set of downregulated genes ( S5 Table , See discussion ) . We confirmed the expression microarray data by RT-qPCR analyses of selected targets ( S10 Figure ) , using the parental BWP17 strain carrying the empty vector ( BWP17AH-CIp10-PTET-GTW , S4 Table ) as a negative control for doxycycline-inducible expression ( Control , S10 Figure ) . Taken together , our transcript profiling data suggested that PGA22 overexpression leads to perturbation of the expression of cell wall genes , consistent with a role of PGA22 in C . albicans cell wall structure and/or function . To test whether the Candida-specific PGA22 gene may confer cell-to-cell adhesion or an aggregation phenotype in a heterologous context , we expressed PGA22 in the non-adherent yeast S . cerevisiae using a surface display system [43] . Briefly , a version of PGA22 deleted for the predicted GPI anchor signal was fused to a S . cerevisiae cell wall targeting signal , and constitutively expressed from the TEF1 promoter . S . cerevisiae cells expressing this fusion protein were allowed to adhere to a 24-well polystyrene plate for 1 h , and the biomass was measured with crystal violet staining , after thorough rinsing of the wells . The S . cerevisiae strain transformed with the empty vector was used as a negative control . There was no significant difference in the biomass between the two strains , but examination of the polystyrene surfaces after rinsing showed aggregation of the PGA22-expressing S . cerevisiae strain ( Fig . 10A ) . The cell wall structures of the control and PGA22-expressing strains were then analyzed by TEM . When expressing PGA22 , S . cerevisiae exhibited a cell wall with a less dense fibrillar outer layer , reminiscent of the structure observed when overexpressing PGA22 in C . albicans ( Fig . 9A and 10B ) . Thus , excess of PGA22 seemed to cause major structural modifications of the outer cell wall , which may modify cell-to-cell and cell-to-surface adhesion properties . In order to get further insight on the role of PGA22 , we investigated the behavior of a pga22Δ/pga22Δ strain upon adherence to and biofilm formation on Thermanox using an existing deletion mutant ( S4 Table ) . We found that the pga22Δ/pga22Δ strain displayed increased adherence to Thermanox ( Fig . 11A ) , although it was not significantly altered for biofilm formation in the microfermentor system ( Fig . 11B ) . We also tested the effect of deleting PGA22 on competitive biofilm growth ( Fig . 11C ) . We generated pga22Δ/pga22Δ mutants expressing either GFP or mCherry ( ΔΔpga22-GFP or ΔΔpga22-mCherry; S4 Table ) and mixed each mutant with the parental strain expressing either mCherry or GFP , respectively ( BWP17-mCherry or BWP17-GFP; S4 Table ) , at a 1∶1 ratio , followed by growth for 40 h in the microfermentor to form biofilms . As a control , biofilm growth of a 1∶1 mixture of strains pga22Δ/pga22Δ expressing GFP and mCherry was used . We quantified the relative strain abundance by qPCR using mCherry and GFP as strain identifiers and two independent primer sets for each gene ( Fig . 11C , see Materials and Methods for details ) . We found that the pga22Δ/pga22Δ mutant outcompeted the parental BWP17 strain ( Fig . 11C ) , correlating with its increased adherence on Thermanox ( Fig . 11A ) . Finally , TEM revealed that the cell wall of the pga22Δ/pga22Δ mutant had a thinner outer fibrillar layer as compared to the control strain , while the inner cell wall was unchanged ( Fig . 11D ) . Taken together , these results indicated that lack of PGA22 resulted in an altered cell wall structure that contributed to increased adherence and occupancy of a multi-strain biofilm and that modifying positively or negatively Pga22 levels in the C . albicans cell wall impacted the cell wall structure and function .
We designed a screen to identify C . albicans genes that when overexpressed alter planktonic growth fitness or strain abundance in a multi-strain biofilm . Surprisingly , our study showed that over-representation of a C . albicans strain in a multi-strain biofilm did not systematically correlate with an increased biomass in single-strain biofilms . This observation reinforces the notion that cell-to-cell interactions play critical roles in the formation of C . albicans biofilms . There was a noteworthy enrichment for cell surface-related genes among the overexpressed genes that conferred increased abundance in the multi-strain biofilm . Indeed , ten genes were included in this category , encoding either predicted GPI-anchored proteins ( Ihd1/Pga36 , Pga15 , Pga19 , Pga22 , Pga32 , Pga37 , Pga42 , Pga59 and Phr2 ) , or a secreted protein of unknown function ( Tos1 ) with similarity to the predicted GPI-anchored protein Pga52 [42] , [44] . Notably , the majority ( 70% ) of these proteins are specific to pathogenic Candida species and 5 have orthologs only in C . dubliniensis ( Pga15 , Pga19 , Pga32 , Pga37 , Pga42; [42] , [45] ) . Overexpression of MSB2 also resulted in increased occupancy of the multi-strain biofilm . Msb2 is a plasma membrane-bound signaling mucin with a heavily glycosylated extracellular domain [46] , [47] . Inactivation of MSB2 leads to a defect in biofilm formation [47] , consistent with our observation that its overexpression favors biofilm formation . Relatively little is known about the functions of the nine predicted GPI-anchored proteins , and their localization at the cell membrane or in the cell wall has not been fully investigated . One exception is Phr2 , a member of the beta-glucanosyltransferase family , which is covalently attached to the cell wall and has a role in cell wall biogenesis at low pH [48] and no described role in biofilm formation . The three remaining genes in the beta-glucanosyltransferase family , namely PGA4/GAS1 , PGA5/GAS2 , and PHR3 were not tested in this study . Pga59 is a small , abundant , cell wall GPI-anchored protein whose absence negatively impacts on cell wall integrity and hyphal morphogenesis [49] . While the PGA59 gene is highly expressed in biofilms [49] , [50] , it is not strictly required for biofilm formation under the conditions analyzed to date [49] . Overexpression of PGA62 , a paralog of PGA59 , also increased C . albicans occupancy of the multi-strain biofilm but to a lower extent than PGA59 overexpression ( S3 Table ) . PGA15 , PGA22 , PGA37 and PGA42 are members of the Cell Surface-Targets of Adherence Regulators ( CSTAR ) group of genes [25] . PGA15 , PGA41 and PGA42 are members of a C . albicans-specific gene family [44] . Overexpression of PGA41 also showed a tendency for increased C . albicans occupancy of the multi-strain biofilm ( S3 Table ) . Overexpression of the different members of a gene family did not always result in similar phenotypes in the ST-OE screen . For instance , PGA37 and PGA57 are paralogs but only overexpression of PGA37 resulted in increased occupancy of the multi-strain biofilm under the conditions tested ( S3 Table ) . We found that many of the genes identified in our screen were upregulated at/in different steps and/or models of biofilm development , including PHR2 [51] , TOS1 [34] , [52] , MSB2 , PGA32 , PGA22 , BEM2 , IHD1 and PGA37 [34] . For instance , PHR2 appears induced during late steps of biofilm development ( i . e . mature biofilms ) [51] , whereas TOS1 appears induced during early events [52] . The behavior of the overexpression strains for these ten cell surface-related genes in single-strain biofilms did not follow a general pattern . Overexpression of some genes ( eg PGA59 and IHD1 ) resulted in increased adherence and an increased or unchanged biofilm biomass . In contrast , overexpression of other genes had a negative impact on adherence and no impact on biofilm formation ( PGA19 , PGA32 and PGA37 ) , a positive impact on adherence and a negative impact on biofilm formation ( PGA15 , PGA22 ) or no discernible impact ( PGA42 , PHR2 and TOS1 ) . Taken together , these results suggest that the ability of C . albicans strains to adhere to the biofilm substratum is not the only defining component of biofilm development . In addition , the discrepancies observed when testing single-strain biofilm formation and strain occupancy in a multi-strain biofilm indicate that the environment provided by a multi-strain biofilm may favor biofilm occupancy by a strain otherwise defective for single-strain biofilm formation . This is illustrated by the PGA22-overexpression strain that had increased adherence , but adhered cells were more sensitive to shear forces , due to cell-to-cell clustering , thus delaying the emergence of a single-strain biofilm . Such sensitivity to shear forces was not observed when this strain formed a biofilm in combination with one or several other C . albicans strains that expressed PGA22 at normal levels . Hence the presence of wild type cells may protect the PGA22-overexpression strain from the deleterious effects of shear forces . Cooperativeness between strains that express cell surface proteins to different levels has already been observed . Indeed , Nobile et al . [15] have shown that hwp1 and als1als3 knockout mutants are individually defective in biofilm formation . Yet , the combination of these mutant strains results in the formation of an intact multi-strain biofilm . This suggested that Hwp1 interacts with Als1 and Als3 in order to ensure efficient biofilm formation . Here , we did not identify any functional link between the predicted cell surface proteins studied , but our results indicate that they may participate in cell-to-substrate and cell-to-cell interactions , with different impacts on biofilm formation . It is currently unknown how PGA22 precisely confers increased cell-to-cell and cell-to-substrate adhesion . We observed that overexpression of PGA22 in either C . albicans or S . cerevisiae had a significant impact on the cell wall structure of both species and caused cell aggregation ( Fig . 7 and 10 ) , suggesting that the Pga22 protein might have a more direct effect on the cell wall structure . Our AFM experiments indicated a clustered repartition of the adhesion events on the cell surface of the C . albicans cells overexpressing PGA22 , not seen with PGA59 . This may indicate the presence of Pga22-enriched cell surface domains and explain the alteration of cell wall structure in both C . albicans and S . cerevisiae . Our genome-wide transcript profiling data also showed that PGA22 overexpression triggered alterations in the expression of cell wall genes , including putative adhesins ( e . g . FGR41 , and PGA38 ) or chitin synthesis and remodeling genes ( CHS1 and CHT3 ) that may participate in conferring the adhesion phenotype or be a consequence of altering the cell wall structure and/or function by Pga22 overproduction ( Table 3 , S5 Table , S10 Figure ) . PGA22 overexpression also correlated with downregulation of mannosyltransferase-encoding genes ( MNN12 , MNT1 , KTR4 and RHD1 , S5 Table ) as well as genes associated with mannosyltransferase activity and wall maintenance ( VRG4 , SMF12 , SKN1 , S5 Table ) , which may impact on the production of mannoproteins at the external layer of the cell wall as well as affect the overall wall structure and composition . Consistently , TEM analyses revealed that the PGA22-overexpressing strain had an altered mannoprotein-rich outer fibrillar layer ( Fig . 9 ) . Our preliminary results also showed that overexpression of PGA22 reduced the cell wall protein content and decreased Concanavalin A staining , indicative of reduced mannoprotein abundance . On the other hand , both overexpression and deletion of PGA22 cause increased adherence and increased biofilm formation under competitive growth with a wild-type strain ( Fig . 4 and S7 Figure ) . It is possible that cells compensate for PGA22 absence by producing other adhesion proteins . Previous studies have reported cases where a gene deletion phenocopied the overexpresser . For instance both deletion and overexpression of α-1 , 2-mannosyltransferases similarly altered cell wall integrity in Mycobacterium smegmatis and Mycobacterium tuberculosis [53] . Both overexpression and deletion of SFL1 , encoding a transcription factor that controls the yeast-to-hyphae transition , attenuated virulence of C . albicans in a mouse model of systemic infection [54] . Clearly , more studies are needed for a better understanding of the complexity of PGA22 function during C . albicans biofilm development . Further experiments will also be needed in order to understand the basis for the changes in adherence and biofilm formation of overexpression strains for other cell wall genes identified in this study . These changes may reflect subtle modifications in the physical properties of their cell walls , including altered charge and hydrophobicity and compensatory changes in the cell wall proteome , as well as modification of the extracellular matrix . Our overexpression screen uses the conditional overexpression system pNIM1 [30] , [62] that has the advantage of inducing gene expression under tightly controlled conditions and bypassing the effect of any mutation ( e . g . acquired during C . albicans transformation ) that could interfere with the phenotype . In addition , our validation experiments were performed using independently-generated overexpression strains carrying the alternative doxycycline-inducible overexpression system pNIMX that allows for higher overexpression levels to be achieved [30] . Some limitations of our system include possible doxycycline-dependent effects on the observed phenotypes , even if low concentrations of doxycycline ( in the range of 40-50 µg/ml ) were used . We provide some data arguing against doxycycline interference with both gene expression ( S10 Figure; Control ) and phenotype ( S9 Figure; Vector-only control ) . Biofilms are also notorious for their ability to exclude small molecules , such as antifungal agents and antibiotics , and it is possible that doxycycline does not reach deep areas within the mature biofilm . Our biofilm development model relies on a continuous supply of the growth medium with doxycycline , suggesting that at least cells that adhere to or are located within the external surface of the biofilm are constantly exposed to doxycycline . Furthermore , strains overproducing Pga22 express high levels of PGA22 even after 16 h of exposure to doxycycline ( S10 Figure ) indicating that doxycycline promoter-driven transcripts are sustained over a long period of time . Overexpression of several C . albicans genes has previously been shown to affect biofilm formation . For instance , overexpression of ADH5 , GCA1 , and GCA2 promoted matrix production , while overexpression of CSH1 and IFD6 inhibited it [23] and PES1 overexpression increased the dispersal step [55] . These genes were not included in our overexpression collection . Other genes , with known biofilm phenotypes when overexpressed , were included in our collection e . g . NRG1 , UME6 and GAT2/BRG1 . Overexpression of NRG1 repressed morphogenesis and resulted in the formation of yeast-only biofilms [56] . Overexpression of UME6 and GAT2/BRG1 triggered hyphal formation independently of the presence of hypha-inducing cues [32] and resulted in the formation of hyperfilamentous biofilms and increased biofilm biomass , respectively [32] , [56] , [57] . Despite these previous findings the NRG1- , UME6- and GAT2/BRG1-overexpression strains did not display altered occupancy in our multi-strain biofilm model . This reinforces the notion that the behavior of individual strains in a multi-strain biofilm cannot be directly predicted from their behavior in a single-strain biofilm model . In fact , the majority of the genes identified through the ST-OE biofilm screen did not have an impact on morphogenesis when overexpressed ( S2B Figure ) , although this process is of central importance to biofilm formation [8] . Our unpublished data showed that only overexpression of ORF19 . 3459/MCK1 resulted in a filamentation phenotype at 37°C ( but not 30°C ) in media that do not normally promote hyphal growth . Thus , this strain might show normal adherence but increased hyphal growth and representation in a biofilm . In addition , overexpression of two cell polarity genes CDC24 and BEM2 [58] resulted in increased representation in the multi-strain biofilm although the underlying mechanism remains to be investigated . We found that overexpression of four genes involved in the regulation of cell-cycle progression and the DNA-damage response , namely RAD53 , RAD51 , PIN4 and ORF19 . 2781 , and one gene known for its role in the regulation of C . albicans morphogenesis ( SFL2 ) , resulted in decreased fitness upon planktonic growth with doubling times increasing by 3-27% when strains were grown individually ( Fig . 2B ) . Our results are consistent with the S . cerevisiae phenotypes associated with overexpression of the orthologs of these five genes [27] , [59] . One observation that emerges from our study is the low number of C . albicans genes that when overexpressed resulted in decreased fitness ( 5/531; 1% ) . This contrasts with the observation of Douglas et al . [59] who used a similar setting ( growth of a pool of ∼5 , 100 barcoded PGAL4-dependent S . cerevisiae overexpression strains ) and identified 361 ( 7 . 1% ) strains with decreased fitness after 20 generations under inducing conditions . This may reflect a weaker level of overexpression from the PTET promoter used in our study . Indeed , when using a modified transactivation system that induces higher expression from the PTET promoter ( pNIMX , [30] ) and a collection of 257 overexpression strains for genes largely overlapping those analyzed here , 18 strains ( 7 . 0% ) were identified with a fitness defect , including those overexpressing RAD53 and SFL2 [30] . A second interesting observation is the lack of genes whose overexpression conferred increased fitness . This was also the case when higher levels of expression were driven from the pNIMX system ( S . Znaidi and C . d'Enfert , unpublished ) and in the study of ∼5 , 100 S . cerevisiae overexpression strains [59] . It is likely that use of a relatively rich medium provided optimal growth conditions and therefore limited discriminatory capacity for increased fitness . Our approach also identified four genes , namely ASH1 , ORF19 . 2781 , PRR2 and STB5 whose overexpression led through unexplored mechanisms to under-representation of the corresponding strains in a multi-strain biofilm ( Fig . 3 , Table 1 ) . It is notable that among these genes , only ORF19 . 2781 resulted in a lower fitness when overexpressed in planktonic culture , suggesting that fitness determinants may differ depending on growth conditions . In conclusion , our results illustrate the power of using signature tagging in conjunction with gene overexpression for the identification of genes involved in biofilm formation and more general processes pertaining to C . albicans virulence . This warrants our current development of a genome-wide collection of C . albicans overexpression strains [60] . Moreover , our results reveal how targeted changes in the cell wall proteome differentially alter C . albicans ability to form single- and multi-strain biofilms , re-emphasizing the importance of the cell wall and cell-cell interactions in C . albicans pathobiology .
C . albicans overexpression strains used in this study have been derived from 294 barcoded integrative overexpression plasmids described previously [30] , as well as 237 novel barcoded plasmids for overexpression of transcription factors and signaling components ( 90 ) , cell wall-related proteins ( 61 ) and genes involved in DNA replication , recombination and repair ( 86; see S1 Table for a list of all ORFs included , primers used for their amplification and corresponding barcodes ) . Briefly , for this latter set of 237 plasmids , the respective ORFs were PCR amplified using chimeric primers followed by recombination-mediated transfer into the Gateway donor vector pDONR207 [30] , [61] . The set of pDONR207 derivatives was fully sequenced to ascertain that no unintended mutations were introduced during PCR amplification . The pDONR207-ORF plasmids were then used in a Gateway LR reaction together with barcoded derivatives of the CIp10-PTET-GTW vector [30] , carrying a TET promoter ( PTET , [62] ) . All barcoded overexpression vectors were linearized with StuI and used to transform either CEC1121 , a derivative of SN148 [63] ( S4 Table ) , or CEC1429 , a derivative of CAI4 [64] ( S4 Table ) , both strains harbor the pNIM1 plasmid [62] for doxycycline-regulated expression from the PTET promoter . Transformants were selected and checked as described yielding 531 overexpression strains [61] . Selected overexpression plasmids were also used to transform strains CEC3783 or CEC3781 carrying the pNIMX plasmid ( S4 Table; [30] ) for doxycycline-regulated expression from the PTET promoter and either a PTDH3-BFP or PTDH3-GFP gene fusion for constitutive expression of BFP or GFP , respectively . Constructs with GFP and BFP were integrated between PGA59 and PGA62 . PTET was induced with 50 µg . mL−1 of doxycycline in all experiments . The C . albicans ΔΔpga22 loss-of-function mutant was generated in BWP17 [65] by successive replacement of the complete ORF in the two alleles using PCR-generated disruption cassettes flanked by 100 bp of target homology region as previously described [66] . The disruption cassettes were amplified using oligonucleotides 3738J5DR and 3738J3DR described in S6 Table and ARG4- or HIS1-bearing plasmids . The resulting transformants were verified by PCR and one clone was selected for subsequent transformation with plasmid CIp10 [67] yielding the prototrophic ΔΔpga22 mutant ( S4 Table ) . Alternatively , the selected clone was transformed with CIp10 derivatives harboring either the GFP gene placed under the control of the C . albicans TDH3 promoter or the mCherry gene placed under the control of the C . albicans ADH1 promoter , yielding strains ΔΔpga22-GFP and ΔΔpga22-mCherry , respectively . The 531 signature-tagged overexpression strains were thawed on Nunc omnitray plates ( Thermo Scientific ) containing YPD ( 1% Yeast Extract , 2% Bacto-Peptone , 2% D-glucose ) -agar using a 96 pin replicator and allowed to grow for 6 days at 30°C . No significant colony size alterations were recorded . 10 mL of YPD were added to each plate and colonies were scraped off using a cell spreader . Strains were pooled in ∼100 mL YPD/15% glycerol at a concentration of ∼57 OD600 ( optical density at 600 nm ) units . mL−1 , aliquoted in 2-mL tubes and frozen at −80°C . The overexpression strain pool was grown at 30°C with agitation ( 200 rpm ) for 16 generations in GHAUM medium , a synthetic defined medium ( 0 . 67% Yeast Nitrogen Base , 2% D-glucose ) supplemented with histidine , arginine , uridine and methionine ( at final concentrations of 1 mg . mL−1 , 1 mg . mL−1 , 0 . 02 mg . mL−1 and 2 mg . mL−1 , respectively ) , in the absence or presence of 50 µg . mL−1 doxycycline . Genomic DNA was extracted as described for S . cerevisiae in Rose et al . [68] from strain pools , followed by PCR-amplification of the barcodes using primers CipSAC2-UP-2 and CipSAC2-DWN-2 ( 3 min at 94°C; followed by 35 cycles of 30 sec at 94°C , 30 sec at 50°C , and 30 sec at 72°C; and a final step of 7 min at 72°C ) ( see S6 Table for primers used in this study ) . The PCR products were then subjected to indirect differential fluorescent dye labeling ( Cy5 for Dox-treated , Cy3 for untreated pools ) . Labeled DNA was resuspended in 50 µL DigEasy Hyb solution ( Roche ) , incubated at 95°C for 5 min , snap-cooled on ice and directly deposited on a barcode microarray ( Agilent Technologies , GEO platform # GPL17420 ) containing: i ) ∼12 on-chip replicates of both sense and antisense DNA sequences complementary to 657 tags ( representing 531 strain tags +126 unused tags ) and ii ) different negative control spots ( Agilent reference ) . Hybridization was performed overnight at 25°C , followed by washing and scanning of the arrays using GenePix 4200 AL scanner ( Molecular Devices ) . This experiment was repeated twice independently . Microarray data were analyzed using two distinct data processing softwares: GeneSpring GX 11 ( Agilent Technologies ) and ArrayPipe v2 . 0 [69] . Z-score ( i . e . number of standard deviations from the population mean ) calculations were performed using ArrayPipe v2 . 0 . The thresholds for GeneSpring were kept at Fold Change values equal or superior to 2 and p-values equal or inferior to 0 . 05 , while thresholds for ArrayPipe v2 . 0 were absolute Z-score values equal or above 1 . 5 and p-values equal or below 0 . 05 . Only strains that met both algorithm thresholds for both sense and antisense barcode fluorescence signals were kept as altering planktonic growth . Microarray data have been deposited at GEO under accession number GSE48647 and Z-score and fold-change data are available in S2 Table . Strains were individually grown three-times independently in 96-well plates at a starting optical density ( OD600 ) of 0 . 1 in 100 µL of YPD supplemented with or without 50 µg . mL−1 doxycycline . The OD600 was measured every 5 min using a Tecan Infinite 200 reader . Tecan OD600 readings were converted into “flask OD600” reading using the following formula: ODFlask = ODTecan ×12 . 2716–1 . 0543 [70] and doubling times were calculated within the exponential growth interval as previously described [71] . The inoculum was prepared from an early-stationary-phase culture of either the pool of overexpression strains , a combination of two equally represented strains or individual strains grown in flasks at 30°C in an orbital shaker . Cells were grown in GHAUM medium with or without 50 µg . mL−1 doxycycline , each inoculum was then diluted to an OD600 of 1 in fresh GHAUM medium with or without 50 µg . mL−1 doxycycline and left at room temperature for 30 min , to allow further overexpression . Plastic slides ( Thermanox; Nunc ) were immersed in the inoculum for 30 min at room temperature to allow adherence of cells to the plastic substrate . The plastic slides were then transferred to the glass vessel of a 40-mL incubation chamber [50] . This vessel has two glass tubes inserted to drive the entry of medium and air , while used medium is evacuated through a third tube . The flow of GHAUM medium is controlled by a recirculation pump ( Ismatec ) set at 0 . 6 mL . min−1 and pushed by pressured air supplied at 105 Pa , conditions minimizing planktonic phase growth and promoting biofilm formation . The chambers with the plastic substrate were incubated at 37°C and biofilms ( 8 independent biological replicates ) were grown for 40 h followed by genomic DNA extraction , barcode amplification and differential labeling ( Dox-treated samples with Cy5 , untreated samples with Cy3 ) and hybridization to barcode microarrays as described above . We performed two independent analyses of our microarray data using the Arraypipe [69] or GeneSpring softwares . Arraypipe analyses identified 29 genes with absolute Z-score values above or equal to 1 . 5 , fold change above or equal to 2 , and p value below or equal to 0 . 05 ( Fig . 3 , Table 1 , S3 Table ) , while GeneSpring analyses identified 21 genes when the last two selection criteria described above were used ( Table 1 and S3 Table ) . Microarray data have been deposited at GEO under accession number GSE48647 . Mixed biofilms with two strains expressing either the BFP or the GFP genes under the control of the PTDH3 promoter were grown for 40 h as described above . Plastic substrates were then recovered and immersed in 25 mL of PBS . Biofilms were detached from the plastic substrates by vortexing twice for 15 sec , and collected by centrifugation . Genomic DNA was extracted using MasterPure Yeast DNA Purification Kit ( Epicentre ) , and quantified using a NanoVue Plus ( GEHealthcare Life Sciences ) ; all samples were adjusted to a DNA concentration of 100 ng . µL−1 . 10 ng . µL−1 and 1 ng . µL−1 dilutions were used as templates for quantitative PCR ( qPCR ) , with the following protocol: 0 . 2 µM of each primer was added to SYBR Green ( Invitrogen ) and 5 µL of DNA , in a total reaction volume of 25 µL; qPCR was performed as follow: 3 min at 94°C , followed by 35 cycles of 30 sec at 94°C , 30 sec at 58°C , and 30 sec at 72°C , and a final step of 7 min at 72°C . Each sample was tested for amplification within the BFP and GFP coding regions ( using the primers BFPpFwd and BFPpRev , and GFPpFwd and GFPpRev respectively; S6 Table ) . The resulting Ct value of each amplification was analyzed in order to assess the ratio between the GFP- and BFP-strains within the biofilm , in the induced ( with Dox ) and non-induced ( without Dox ) samples . Six replicates for each strain were analyzed through a Student's t-test . For competitive growth of the ΔΔpga22 mutant versus the parental BWP17 strain , two independent sets of primers for mCherry ( primers 1 , mCherry2-FWD and mCherry2-REV; primers 2 , mCherry-FWD and mCherry-REV ) and GFP ( primers 1 , GFPpFwd and GFPpRev; primers 2 , GFP . RT . fw and GFP . RT . rv ) were used ( S6 Table ) and two independent experiments were averaged . The TEF3 gene was used as a calibrator and the ACT1 gene was used as a control ( primers ACT1-FWD and ACT1-REV , S6 Table ) . Biofilms were grown and recovered from the substrate after 18 , 24 , 40 , 48 and 65 h of growth as described above . The PBS solution containing the detached biofilm was vacuum-filtered through a 1 . 2 µm filter ( Millipore ) ; the filter was dried at 60–65°C for 2–3 days and then weighed on a precision scale ( Mettler AE200; Mettler Toledo ) to obtain the dry mass of the biofilm . A minimum of 3 replicates were analyzed through a Student's t-test . After adherence as described above , the plastic substrates were washed three times in PBS to remove non-adherent cells , and mounted on a glass slide for observation with a Leica DM RXA microscope , using an objective at 10× magnification or an oil-immersed objective at 40× magnification . Pictures of the substrates were taken and 15 fields were counted for each strain per condition ( presence or absence of doxycycline ) , except for the strains overexpressing PGA42 , TOS1 and PHR2 , with 10 fields each . The replicate measurements were analyzed through a Student's t-test . The same test was performed for the ΔΔpga22 knockout strain , without doxycycline , and the results were compared to the wild-type strain SC5314 . Pictures of the substrates were taken and 20 fields were counted for each strain . 10 mL of YPD ( 1% Yeast Extract , 2% Bacto-Peptone , 2% D-glucose ) liquid medium were inoculated with PGA22-overexpression strain , PGA59-overexpression strain , and the control strain with empty plasmid ( CEC3785 ) , and incubated ON at 30°C , in an orbital shaker ( 180 rpm ) . A 10 µL aliquot of each culture was then diluted to 10 mL in fresh YPD medium with or without 50 µg . mL−1 doxycycline , and allowed to grow for 16 hours to allow overexpression of the targeted protein . 5 mL of the cell culture were then quickly centrifuged , washed with 5 mL of acetate buffer ( 18 mM CH3COONa , 1 mM CaCl2 and 1 mM MnCl2 , pH 5 . 2 ) and resuspended in 3 mL of the same buffer . 100 µL of this cell suspension were deposited on a freshly oxygen-activated microstructured PolyDiMethylSiloxane ( PDMS ) stamp . Cells were immobilized in the PDMS stamps as described elsewhere [72] , and immersed in the same acetate buffer . To get statistical significance of the AFM data , about 10–15 cells have been analyzed from three independent experiments for each strain . AFM experiments were conducted on a Nanowizard III from JPK Instruments ( Berlin , Germany ) . We used MLCT probes from Bruker probes with a spring constant of 0 . 02 N . m−1 +/−10% measured before each experiment by the thermal noise method . Adhesion force maps were recorded in force volume mode ( 32×32 or 64×64 force curves ) . The maximum applied force has been set to 2 nN , the Z displacement to 2 µm and the retract time to 50 ms ( with a loading rate of 800 , 000 pN . s−1 ) . Force curves were analyzed using JPK data processing software to extract the maximum adhesion force on each force curve . Following adherence , the plastic substrates were introduced in the continuous-flow fermentor system for 2 h , and then observed under the microscope . 30 fields for each condition ( presence or absence of doxycycline ) were photographed . The cells that were released and washed away during these 2 h under the continuous-flow conditions were also collected , pelleted and resuspended in 1 mL of PBS , 400 µL of which were counted on a MACS Quant ( Mylteni Biotec ) . Student's t-tests were performed . Mixed biofilms with two strains expressing either the BFP or the GFP genes under the control of the PTDH3 promoter were grown for 40 h as described above . Confocal microscopy was then performed on the recovered plastic substrates , using a Zeiss LSM 700 laser scanning confocal microscope on an upright Axio Imager Z2 stand , using a Zeiss W-nACHROPLAN 40X/0 . 75 working distance 2 . 1 mm objective; z-stacks of the biofilms were obtained using the blue and green lasers , for the whole biofilm thickness . Z-stacks were then analyzed using Volocity software to acquire the volume occupied by the cells in the green channel ( overexpression mutant , expressing GFP ) and by the cells in the blue channel ( control strain , expressing BFP ) . Overexpression strains were grown overnight in YPD in the presence or absence of 50 µg . mL−1 doxycycline . S . cerevisiae strains were grown overnight at 30°C in liquid YNB N5000 medium ( 0 . 17% YNB w/o AA w/o ammonium sulfate; 1% Glc; 0 . 5% Ammonium sulfate ) supplemented with leucine , histidine and lysine at a final concentration of 0 . 1 mg . mL−1 . Samples were prepared by high-pressure freezing with an EMPACT2 high-pressure freezer and rapid transport system ( Leica Microsystems Ltd . , Milton Keynes , United Kingdom ) . After freezing , cells were freeze-substituted in substitution reagent ( 1% [wt/vol] OsO4 in acetone ) with a Leica EMAFS2 . Samples were then embedded in Spurr resin and additional infiltration was provided under a vacuum at 60°C before embedding in Leica FSP specimen containers and polymerizing at 60°C for 48 h . Semithin survey sections , 0 . 5 µm thick , were stained with 1% toluidine blue to identify areas containing cells . Ultrathin sections ( 60 nm ) were prepared with a Diatome diamond knife on a Leica UC6 ultramicrotome and stained with uranyl acetate and lead citrate for examination with a Philips CM10 transmission microscope ( FEI UK Ltd . , Cambridge , United Kingdom ) and imaging with a Gatan Bioscan 792 ( Gatan United Kingdom , Abingdon , United Kingdom ) . The thicknesses of the inner and outer layers of the cell wall were measured using Image J and by averaging 30 measurements for each cell ( n = 30 cells ) . Analyses were performed using Student's t-tests . PGA22 overexpression strain was grown three times independently in YPD medium supplemented or not with 50 µg . mL−1 doxycycline during 16 h . Total RNA was extracted using the hot phenol method as described previously [36] , followed by first-strand cDNA synthesis and Cy5 ( doxycycline-treated samples ) /Cy3 ( untreated samples ) labeling from 20 µg total RNA , using the Superscript III indirect cDNA labeling system ( Invitrogen ) . Purified labeled samples were mixed and hybridized to a C . albicans expression array ( Agilent Technologies ) designed such that two nonoverlapping probe sets target each of 6 , 105 C . albicans ORFs for a total of 15 , 744 probes , thereby allowing two independent measurements of the mRNA level for a given gene [36] . Hybridization was performed as described elsewhere [36] . Images of Cy5 and Cy3 fluorescence were generated by scanning the expression arrays using an Axon Autoloader 4200AL scanner ( Molecular Devices , Downington , PA ) . Images were subsequently analyzed with the GenePix Pro 6 . 1 . 0 . 2 software ( Molecular Devices , Downington , PA ) . GenePix Results ( GPR ) files were imported into the Arraypipe 2 . 0 for spot filtering , background subtraction ( limma normexp BG correction ) and Lowess global normalization of signal intensities [73] . Replicate arrays ( n = 3 ) were combined and fold-change and P-values ( standard Student's t-test ) were calculated . For RT-qPCR analyses , the strain CEC3785 ( S4 Table ) was grown exactly as described above and used as a negative control for doxycycline treatment . Total RNA from both the PGA22 overexpression- and control strains was extracted using the hot phenol method [36] and reverse transcription ( RT ) was performed using the SuperScript III first-strand synthesis system using 5 µg of total RNA ( Invitrogen , catalog # 18080-051 ) in a total reaction volume of 20 µL . The qPCR reaction was made of 1 µL from the RT reaction mixture combined with 4 µL of primer mix at 10 pmol . µL−1 each ( forward and reverse primers of the selected genes , S6 Table ) , 10 µL of 2X Takyon Rox SYBR MasterMix dTTP Blue ( Eurogentec ) and 5 µL of H2O ( total volume = 20 µL ) . Q-PCRs were performed in a MicroAmp Optical 96-Well Reaction Plate ( Applied Biosystems ) using an Eppendorf realplex4 Mastercycler real-time PCR instrument ( Eppendorf ) with 1 cycle at 50°C for 2 min , 1 cycle at 95°C for 10 min and 50 cycles at 95°C for 15 sec and 58°C for 1 min . Data analysis was performed using the realplex software version 2 . 2 ( Eppendorf ) . For each experiment , threshold cycle ( CT ) values were determined using the realplex software . The levels of relative gene expression ( n-fold ) for the doxycycline-treated samples as compared to the untreated controls of PGA22 , ORF19 . 5267 , FGR41 , RBE1 , CHT3 , CHS1 and the ACT1 negative control gene were calculated using the 2−ΔΔCT method , as follows: ΔCT = CT ( selected gene ) − CT ( TEF3 reference gene ) and ΔΔCT = ΔCT ( doxycycline-treated sample ) − ΔCT ( untreated control ) . The ACT1 gene was used as a negative control . Three independent experiments were performed on different days using 2 biological replicates each time ( assumed as n = 6 ) . A two-tailed Student's t-test was applied by comparing the doxycycline-treated set to the untreated set . Statistical significance is set as P≤0 . 05 . PGA22 was amplified from SC5314 genomic DNA using PGA22-GTW-fwd and PGA22ΔCter-rev as primers , and recombined into the Gateway donor vector pDONR207 ( see above ) . The resulting plasmid was sequenced , prior to being transferred into S . cerevisiae Gateway destination vector pBC542 [43] . Briefly , this centromeric plasmid bears the TEF promoter , a Gateway cassette flanked by an inframe HA tag , followed by an inframe S . cerevisiae GPI anchor sequence . The pBC542/CaPGA22 plasmid was used to transform S . cerevisiae BY4742 , and the resulting strain was tested in an adherence assay as described in Monniot et al [74] . After rinsing , the cells adhered to the 24-well plates were imaged with a Leica M80 stereomicroscope and a Leica DMI6000 inverted microscope , using a HC PLAN APOx20/0 . 70 objective .
|
Candida albicans is the most prevalent human fungal pathogen . Its ability to cause disease relies , in part , on the formation of biofilms , a protective structure of highly adherent cells tolerant to antifungal agents and the host immune response . The biofilm is considered as a persistent root of infection , disseminating infectious cells to other locations . In this study , we performed large-scale phenotypic analyses aimed at identifying genes whose overexpression affects biofilm development in C . albicans . Our screen relied on a collection of 531 C . albicans strains , each conditionally overexpressing one given gene and carrying one specific molecular tag allowing the quantification of strain abundance in mixed-population experiments . Our results strikingly revealed the enrichment of strains overproducing poorly-characterized surface proteins called Pgas ( Putative GPI-Anchored proteins ) , within a 531-strain-containing biofilm model . We show that these PGA genes differentially contribute to single-strain and multi-strain biofilm formation and are involved in specific stages of the biofilm developmental process . Taken together , our results reveal the importance of C . albicans cell surface proteins during biofilm formation and reflect the powerful use of strain barcoding in combination with gene overexpression to identify genes and/or pathways involved in processes pertaining to virulence of pathogenic microbes .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biofilms",
"cell",
"walls",
"ecology",
"and",
"environmental",
"sciences",
"genomic",
"library",
"screening",
"genomic",
"library",
"construction",
"model",
"organisms",
"dna",
"construction",
"molecular",
"biology",
"techniques",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"hyperexpression",
"techniques",
"microbial",
"ecology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"gene",
"expression",
"and",
"vector",
"techniques",
"cell",
"biology",
"ecology",
"dna",
"library",
"construction",
"genetic",
"screens",
"library",
"screening",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"yeast",
"and",
"fungal",
"models",
"candida",
"albicans"
] |
2014
|
Targeted Changes of the Cell Wall Proteome Influence Candida albicans Ability to Form Single- and Multi-strain Biofilms
|
The vaccinia virus ( VACV ) A41L gene encodes a secreted 30 kDa glycoprotein that is nonessential for virus replication but affects the host response to infection . The A41 protein shares sequence similarity with another VACV protein that binds CC chemokines ( called vCKBP , or viral CC chemokine inhibitor , vCCI ) , and strains of VACV lacking the A41L gene induced stronger CD8+ T-cell responses than control viruses expressing A41 . Using surface plasmon resonance , we screened 39 human and murine chemokines and identified CCL21 , CCL25 , CCL26 and CCL28 as A41 ligands , with Kds of between 8 nM and 118 nM . Nonetheless , A41 was ineffective at inhibiting chemotaxis induced by these chemokines , indicating it did not block the interaction of these chemokines with their receptors . However the interaction of A41 and chemokines was inhibited in a dose-dependent manner by heparin , suggesting that A41 and heparin bind to overlapping sites on these chemokines . To better understand the mechanism of action of A41 its crystal structure was solved to 1 . 9 Å resolution . The protein has a globular β sandwich structure similar to that of the poxvirus vCCI family of proteins , but there are notable structural differences , particularly in surface loops and electrostatic charge distribution . Structural modelling suggests that the binding paradigm as defined for the vCCI–chemokine interaction is likely to be conserved between A41 and its chemokine partners . Additionally , sequence analysis of chemokines binding to A41 identified a signature for A41 binding . The biological and structural data suggest that A41 functions by forming moderately strong ( nM ) interactions with certain chemokines , sufficient to interfere with chemokine-glycosaminoglycan interactions at the cell surface ( μM–nM ) and thereby to destroy the chemokine concentration gradient , but not strong enough to disrupt the ( pM ) chemokine–chemokine receptor interactions .
Chemokines ( chemotactic cytokines ) comprise a large family of small ( ∼ 7–14 kDa ) , secreted proteins that direct the migration of leukocytes into areas of infection and inflammation , as part of the innate immune response [1 , 2] . They function by binding to cell surface glycosaminoglycans ( GAGs ) and establishing concentration gradients , which are detected by their cognate seven-transmembrane G-protein coupled receptors ( GPCRs ) on the surface of immune cells . This leads to activation of leukocytes and macrophages and their consequent chemotaxis to sites of inflammation and infection . Chemokines are classified into four subfamilies called C , CC , CXC and CX3C based on the arrangement of conserved disulphide forming cysteine residues at the N terminus [3] . Despite differences in primary sequence and varied functions within the superfamily , chemokines adopt very similar tertiary structures , with an extended N-terminal loop region followed by a three β stranded sheet arranged in a Greek key motif [4 , 5] , and form strong pM interactions with their cognate receptors . Poxviruses are a family of large , double-stranded DNA viruses that replicate in the cell cytoplasm [6] . Vaccinia virus ( VACV ) , the prototype member of this family , was the live vaccine used to eradicate smallpox , caused by variola virus ( VARV ) [7] . The VACV strain Copenhagen genome contains approximately 200 genes [8] , although the number of genes varies slightly between strains [9] . About half of these genes are dispensable for virus replication but affect virus virulence , host range or interactions with the immune system . One poxvirus immune evasion strategy is to target the chemokine/chemokine receptor system by encoding viral mimics of chemokines or chemokine receptors , or to express secreted chemokine binding proteins [10 , 11] . Several poxviruses encode a secreted CC chemokine inhibitor ( vCCI ) ( also previously called vCKBP/T1/35 kDa ) that is unrelated to cellular chemokine receptors and binds tightly to CC chemokines with pM affinities [12–14] . vCCI functions by competing for and preventing the interaction of chemokines with their receptors on leukocytes and so blocking their role in the inflammatory response . In contrast , it was proposed that another poxvirus chemokine binding protein , called M-T7 from myxoma virus ( MYXV ) , blocks the binding of chemokines to GAGs and thereby prevents the interaction of the chemokines with the endothelial cell surface [15] . Consequently , a chemokine concentration gradient is not established near the site of infection and leukocytes are not recruited . The vCCI proteins from leporipoxviruses and orthopoxviruses share several conserved features and therefore might share similar binding modes with chemokines . A collection of structures of vCCI proteins from cowpox virus ( CPXV ) [16] , rabbitpox virus ( RPXV ) [17] and ectromelia virus ( ECTV ) [18] reveal a distinct β sandwich topology with no obvious relationship to host chemokine receptors , which are seven-transmembrane GPCRs . One face of the vCCI β sandwich ( sheet I ) is rather flat and electrostatically bland , whereas the second face ( sheet II ) is elaborated by a conserved loop , that contributes to the highly negatively charged solvent accessible surface . The acidic surface residues of sheet II are highly conserved in vCCI proteins and were hypothesized to be the chemokine-binding surface for these proteins [16] . A subsequent structure of the complex between RPXV vCCI and human CCL4 confirmed this [17] , and structure-based mutational analysis of the ECTV vCCI protein highlighted residues involved in high-affinity interactions between vCCIs and their CC chemokine partners [18] . These recent structures have contributed to an emerging paradigm for vCCI–chemokine interactions , allowing comparisons to be made of the binding modes of these soluble viral proteins and host chemokine receptors . This paper focuses on another immunomodulatory protein that shares sequence similarity to vCCI and is encoded by VACV gene A41L . This gene has also been called 166 using a nomenclature in which genes are numbered sequentially from left to right of the VACV strain Western Reserve ( WR ) [9] . The A41 protein from VACV strain WR is a secreted 30 kDa glycoprotein and very similar proteins are expressed by camelpox virus , CPXV and all 16 strains of VACV tested [19] . In addition , genome sequencing showed that all orthopoxvirus species and strains studied ( over 70 in total ) encode a highly conserved A41L gene [9] . VACV A41 blocks the recruitment of cells to the site of VACV infection in a rabbit intradermal model , and decreases immunopathology and viral clearance in a mouse intranasal model [19] . When gene A41L was deleted from modified VACV Ankara ( MVA ) , an attenuated VACV strain , the resultant virus induced a stronger CD8+ T-cell response and conferred better protection against subsequent challenge with a pathogenic strain of VACV [20] . Although A41 has biological properties consistent with those of chemokine binding proteins and is related to a known chemokine binding protein , hitherto , the ligands for A41 were unknown . A41 has limited sequence similarity to the poxvirus vCCI family of proteins ( ∼19% sequence identity and ∼40% conservation to both CPXV and ECTV vCCIs ) and in addition shares a similar hydropathy profile and has conserved cysteine residues [19] . A41 also shares amino acid similarity with orf virus GIF protein , another member of the vCCI protein family , which binds granulocyte macrophage colony stimulating factor ( GM-CSF ) and interleukin ( IL ) -2 [21] . Although many poxviruses encode proteins related to VACV vCCI , there are no close relatives outside poxviruses and so the origin of this gene family is uncertain . VACV and orf virus genomes have quite different G+C content ( 33% for VACV and 64% for orf virus ) and the G+C contents of the VACV gene A41L and the orf virus gene encoding the GIF protein match that of the parent virus . Therefore , these genes have been present in these genomes for a long time , or , less likely , were acquired recently from an unknown source with similar G+C content . To better understand the function of VACV A41 we have identified several chemokines as binding partners and solved the A41 crystal structure and refined it to a Bragg spacing of 1 . 9 Å . To identify ligands for A41 , we expressed and purified it from a VACV-expression system [22] in mammalian cells and screened a panel of chemokines for binding by surface plasmon resonance . A41 bound to CCL25 , CCL26 , CCL28 and CCL21 with a Kd between 10−7 to 10−9 M . A41 did not inhibit chemokine-induced chemotaxis , although the interaction between A41 and chemokines was inhibited by GAGs , suggesting A41 functions by blocking the binding of chemokines to GAGs . The A41 structure reveals the β sandwich fold observed in the vCCI family of proteins . However , there are notable differences between the structure of A41 and the vCCIs of CPXV [16] , RPXV [17] and ECTV [18] , notably in the conserved surface loop and the surface charge distribution . In light of the biochemical and structural data we propose a model for the biological function of A41 and the structural interactions that underpin it .
A41 was coupled to the surface of a CM5 sensor chip and 39 murine and human chemokines ( murine chemokines CCL1 , CCL2 , CCL4 , CCL6 , CCL7 , CCL8 , CCL9/10 , CCL11 , CCL12 , CCL19 , CCL20 , CCL21 , CCL22 , CCL24 , CCL27 , CCL28 , CXCL2 , CXCL5 , CXCL9 , CXCL10 , CXCL11 , CXCL12a , CXCL12b , CXCL13 , XCL1 , CX3CL1 , and human chemokines CCL14 , CCL15 , CCL16 , CCL17 , CCL18 , CCL23 , CCL25 , CCL26 , CXCL3 , CXCL4 , CXCL6 and CXCL7 ) were passed sequentially across the chip surface and binding was analysed by surface plasmon resonance ( BIACORE ) . Where possible , mouse chemokines were used initially because VACV strains WR and MVA lacking the A41L gene showed a distinct biological phenotype , compared to control viruses , in a mouse model of infection . A control surface coated with ovalbumin ( 45 kDa , pI 4 . 5 ) , vCCI ( as a positive control for CC chemokines ) and a blank surface were analysed in parallel channels on the same chip . The sensorgram for the blank surface was subtracted from that for each protein . Both the chemokines and the immobilised proteins differed in mass and so the binding responses recorded at the end of the injection ( RUeq ) were converted into %Rmax ( where Rmax is the calculated binding capacity ) . Table 1 shows the binding activity of immobilised A41 and ovalbumin for the different chemokines expressed as %Rmax . Binding of all chemokines to ovalbumin and the majority of chemokines to A41 were very low ( <10% Rmax ) , but binding to A41 of > 20% Rmax was observed for hCCL26 , hCCL25 , mCCL28 , hCXCL4 , mCCL21 and mCXCL13 ( Table 1 ) , suggesting that these might be ligands for A41 . vCCI also bound all the CC chemokines bound by A41 . To investigate these interactions further , kinetic analyses were performed for mouse and human versions of these chemokines using A41 produced from mammalian cells and from E . coli . Binding affinities ( Kd ) to A41VOTE and A41E . coli ( mean values from at least two experiments performed on two different chips ) are listed in Table 2 . Kd values of mouse and human CCL25 , CCL28 , CCL26 and CCL21 are in the range of 8–118 nM , whereas Kd values for hCXCL4 , hCXCL13 and mXCL1 were higher than 1 μM . Although there were some differences , the binding affinities for each chemokine were generally similar for A41VOTE and A41E . coli , indicating that A41 can bind to these ligands whether or not it is glycosylated . Overall the Kd values of A41 for these human and mouse chemokines are one to three orders of magnitude higher ( lower affinity ) than the Kd values of vCCI for many CC chemokines [13] . Poxvirus chemokine binding proteins have been reported to interact with chemokines in two distinct ways . The vCCI protein encoded by VACV and CPXV binds to chemokines such that receptor binding is blocked [13 , 14 , 17 , 23 , 24] and , the vCCI-chemokine interaction was not inhibited by high concentrations of GAGs , such as heparin [13] . In contrast , the binding of MYXV M-T7 protein to chemokines was blocked by endothelial cell associated GAGs [15] . On this basis a model of the chemokine binding was proposed in which the chemokine receptor binding site and the GAG binding site were considered distinct . The binding of vCCI to CC chemokines would block receptor binding , whereas the inhibition of GAG-chemokine interaction by M-T7 would prevent the establishment of a chemokine concentration gradient around sites of inflammation and the consequent , recruitment of leukocytes . Blocking the binding of a chemokine to its receptor will inhibit leukocyte chemotaxis in vitro , whereas blocking binding of a chemokine to GAGs will not . This distinction was exploited to investigate the mechanism of interaction of A41 with chemokines . The ability of A41 to inhibit chemotaxis induced by CCL21 , CCL25 , CCL26 and CCL28 was examined . Murine L1 . 2 cells expressing endogenous CCR7 were used to examine human and mouse CCL21-induced chemotaxis; 4DE4 cells expressing CCR3 stably were used to examine human CCL26-induced chemotaxis; and murine L1 . 2 cells were transfected with plasmids encoding wild type or HA-tagged human CCR9 and CCR10 to examine human and mouse CCL25- and CCL28-induced chemotaxis [25] . In pilot experiments , cells transfected with the HA-tagged CCR9 and CCR10 migrated in response to CCL25 and CCL28 at similar levels to cells transfected with the wild type receptors ( data not shown ) . Thereafter , the HA-tagged versions were used so that cell-surface receptor expression could be verified by FACS using an anti-HA antibody before chemotaxis experiments ( data not shown ) . The concentration of each chemokine that induced optimal chemotaxis of the appropriate cell type was determined and this concentration was used to determine whether A41 can block chemotaxis . A41VOTE , A41E . coli or ovalbumin were incubated at various molar ratios ( up to 200:1 ) with each chemokine and the number of cells that migrated through a membrane in response to this mixture was determined ( Figure 1 ) . No inhibition of chemotaxis was observed with murine and human CCL21 ( Figure 1 ) . Similarly , with murine and human CCL25 and hCCL26 , even a 200-fold excess of A41 was unable to inhibit chemotaxis by 50% compared to chemokine alone ( Figure 1 ) . Only with human and murine CCL28 was A41 able to reduce chemotaxis to ∼10% of levels induced by chemokine alone , but in the latter case chemotaxis was also inhibited ∼65% by a similar molar excess of ovalbumin . Moreover , a 50-fold excess of mCCL28 inhibited chemotaxis by only ∼35% ( Figure 1 ) and this is inconsistent with efficient blockade of chemokine–chemokine receptor interaction . Collectively , these data indicate that A41 is unable to block chemokine-induced leukocyte chemotaxis effectively . These results are consistent with the weaker binding of A41 to chemokines , compared to the binding of those chemokines to their cellular receptors , and contrast with results for VACV vCCI that inhibited chemotaxis efficiently in a dose-dependent manner and at much lower molar concentrations [13] . The failure of A41 to inhibit chemokine-induced chemotaxis and the Kd values for its interaction with chemokines ( Table 2 ) , suggest that A41 might function by blocking the interaction of chemokines with GAGs . This was investigated by immobilising A41 on a sensor chip , passing hCCL28 across the chip alone or in the presence of increasing concentrations of heparin ( sodium salt , MW 4 , 000–6 , 000 ) and measuring hCCL28 binding by BIACORE . The interaction of A41 and hCCL28 was inhibited in a dose-dependent manner and complete inhibition was achieved with 500 ng/ml heparin ( Figure 2A ) . Similarly , heparin inhibited the binding of hCCL21 , hCCL25 , hCCL26 and hCCL28 to immobilised A41 in a dose-dependent manner when A41 was produced in either mammalian cells ( Figure 2B ) or E . coli ( Figure 2C ) . The concentration of heparin used to achieve 50% inhibition was ∼100 ng/ml ( ∼20 nM ) and this was lower than used by other investigators to achieve disruption of the M-T7 chemokine binding protein from MYXV with RANTES [15] or the interaction of chemokines with the endothelial cell surface [26] . Apart from heparin , other sulphated GAGs were also tested for their ability to inhibit binding of chemokines to A41 . Heparin and dextran sulphate inhibited the hCCL25-A41 interaction , but heparan sulphate , chondroitin sulphate B , chondroitin sulphate C and hyaluronic acid did not ( Figure 2D ) . Similar results were obtained with A41 produced from E . coli or mammalian cells and with hCCL21 , hCCL26 and hCCL28 ( data not shown ) . Notably each GAG able to inhibit the hCCL25-A41 interaction was more highly charged than those that did not inhibit . Collectively these data show that A41 binds a subset of CC chemokines ( CCL21 , CCL25 , CCL26 and CCL28 ) via a site that overlaps their GAG-binding site , but A41 does not inhibit leukocyte chemotaxis . To understand the structural basis of the action of A41 , its crystal structure was determined using protein expressed in E . coli and refolded from inclusion bodies ( Materials and Methods ) . Phase determination was accomplished by MAD analysis of Seleno-methione ( SeMet ) -labelled crystals . Electron density was observed for residues 26–219 ( numbering for mature protein begins at one ) and the structure was refined to 1 . 9 Å ( final R = 20 . 4 and Rfree = 25 . 2 , Table 3 ) . A41 is a single domain protein with the distinctive β sandwich fold seen in the vCCI class of poxvirus chemokine binding proteins . The two β sheets that form the β sandwich , lie parallel to each other ( Figure 3A ) , and are linked by an array of large loops . Five anti-parallel β strands ( 6 , 7 , 1 , 12 and 13 ) form β sheet I and define the core of the structure ( the naming of secondary structure is as defined in Carfi et al , 1999 ) . The second β sheet ( sheet II ) is also composed of 5 β strands; 2 , 4 , 5 and 9 are anti-parallel whilst 11 is adjacent to and parallel with 9 . Sheet I is largely buried from solvent by two long enveloping loops on one side and sheet II on the other ( the other face of sheet II is exposed to solvent ) . One of these loops ( the 9–11 loop , residues 113–144 ) wraps around the molecule and connects β strands 9 and 11 ( Figure 3B ) , whilst the second comprises the C terminus and packs tightly against the face of sheet I , and bears strands 14 and 15 . Two short β strands ( 10 and 14 ) from these loops clip together in front of sheet I . There are also two short helical segments , the first of which ( α1 ) packs against the back of sheet II and is preceded by strand 7 . The second , α2 , helix comprises a single turn prior to strand 10 . Eight cysteine residues in A41 form four disulphide bridges ( C6-C166 , C33-C199 , C58-C104 , and C112-C152 ) . Despite sharing only 19% sequence identity with CPXV and ECTV vCCI , the structure of A41 is strikingly similar to the poxvirus vCCI family of secreted chemokine binding proteins . Superposition of A41 with the CPXV vCCI [16] ( Figure 3C ) aligns 159 Cα atoms ( out of 199 ) with an rmsd of 2 . 4 Å . The level of similarity is comparable for the recent structures of RPXV [17] and ECTV [18] vCCI proteins ( 2 . 4 Å and 2 . 5 Å rmsds for 159 and 160 residues respectively ) . In addition , there is notable structural similarity with the M3 protein of murine γherpesvirus68 ( γHV68 ) [27] , which binds all four classes of chemokines . However M3 is larger , consisting of N- and C-terminal domains , each possessing β sandwich cores similar to those of the vCCIs . The N- and C-terminal domains can be superposed on A41 with an rmsd of 2 . 5 Å and 2 . 8 Å respectively ( 86 and 47 residues equivalenced ) . The most significant structural deviations between A41 and the vCCIs occur in certain surface loops . The first of these , the 9–11 loop ( A41 residues 113–144 ) wraps around the bottom of the β sandwich before arching up the face of sheet I ( Figure 3B ) . The equivalent loop in the CPXV vCCI ( residues 133–161 , as defined by the structural alignment ) has a markedly different orientation and wraps over the top half of sheet I ( Figure 3D ) . The second major deviation between A41 and the vCCI proteins lies in the orientation of loops projecting from the face of sheet II ( Figure 3D ) . The CPXV vCCI protein [16] contains an extended and highly acidic loop ( the 2–4 loop ) between strands 2 and 4 that protrudes from the plane of sheet II and makes contact with a symmetry related molecule , giving rise to a crystallographic dimer . This feature is conserved in the RPXV [17] and ECTV [18] vCCI proteins , but is absent in A41 ( Figure 3B , 3E ) and makes a significant contribution to the charge characteristics of sheet II ( see below ) . The surface charge properties of A41 are broadly similar to those seen in other poxvirus vCCI proteins , although sheet I exhibits a large patch of positive charge ( Figure 4A ) whereas in CPXV vCCI sheet I is comparatively uncharged ( Figure 4B ) . On the opposite face of A41 ( sheet II ) the dominant electrostatic feature is a negatively charged patch , and although this is not conserved in sequence between A41 and the vCCIs this region is negatively charged in the vCCIs and is conserved within that family ( E46 , D49 , E125 and Y62 in particular , CPXV numbering ) . The complex of RPXV vCCI with chemokine CCL4 demonstrated that this negatively charged surface forms crucial electrostatic interactions with the positively charged 20s and 40s loop of the chemokine ( Figure 5A , 5B ) . This charged surface includes the acidic 2–4 loop that harbours residues E46 and D49 and protrudes from sheet II to lock the chemokine in place . This loop differs in length in the vCCIs ( it is 16 and 27 aa long in CPXV and RPXV respectively ) and is absent in A41 . Mapping structure based sequence alignment between A41 and the vCCIs onto the surface of A41 reveals that below this unconserved charged patch is a region of conservation ( Figure 6A , 6B ) , central to which is a strictly conserved phenylalanine ( F181 in A41 ) which forms a hydrophobic depression on the edge of the β sandwich in A41 and the vCCIs ( Figure 5D ) . To try and understand this pattern of sequence conservation , we modelled the binding of A41 to its chemokine binding partners using the structure of RPXV vCCI bound to CCL4 ( pdb code 2ff3 ) . Although only simple rigid body superpositions were performed to separately position A41 and chemokine hCCL26 ( pdb code 1g2s ) onto the vCCI-CCL4 complex ( maximizing the structural overlap with the appropriate component , program SHP [28] ) , the quality of the fit of docking is surprisingly good , with few serious steric clashes ( 11 residues of CCL26 are within 4 Å of 14 residues of A41 , Figure 5C , 5D ) . This A41-CCL26 model reveals that for A41 both the region of sequence conservation , in particular the hydrophobic patch around F181 , and the negatively charged patch make close contacts to the chemokine . Analysis of the RPXV vCCI-CCL4 complex shows that two regions of the chemokine are important for binding , both of which show amino acid sequence conservation . The first region is the N-terminal loop of CCL4 , in particular a highly conserved phenylalanine at position 13 in CCL4 , which makes contact with the conserved vCCI residues ( homologous to F181 in A41 ) at the edge of the β sandwich that forms the shallow hydrophobic depression . Secondly , positively charged residues in the 20s and 40s loops of CCL4 ( particularly positions corresponding to 17 , 23 , 45 , 47 and 48 CCL26 , Figure 5E ) interact with the negatively charged surface formed primarily by the 2–4 loop of the vCCI . The molecular determinants shaping this binding mechanism are highly conserved in all CC chemokines that bind vCCIs with high affinity [29] . Although the structure of A41 is largely similar to the vCCIs , only a subset of CC chemokines bind to A41 with measurable affinity , and even the tightest interaction is nearly two orders of magnitude weaker than achieved for by the VACV vCCI and CCL3 . Nevertheless , the critical residues that define the binding mode of chemokines to vCCIs are conserved in the subset of chemokines ( CCL21 , 25 , 26 and 28 ) that bind to A41 ( Figure 5E ) . The absence of the 2–4 loop in A41 may explain the specificity of A41 for certain CC chemokines and underlies the weakness of the interaction . This loop is likely to be important for the broad specificity of the vCCI proteins , providing a somewhat flexible electrostatic platform with which to sequester chemokines via their positively charged GAG-binding sites . High affinity might then be conferred by hydrophobic interactions of the N-terminal loop . The lack of the 2–4 loop may therefore restrict the selectivity of A41 to only a few chemokines , although the conservation of a negative surface patch on sheet II is sufficient to form significant electrostatic interactions with certain chemokines . Support for this model comes from an analysis of the amino acid sequence of those chemokines that bind A41 most tightly , which reveals that they possess insertions around the 20s and 40s loops ( Figure 5E ) , with longer loops conferring greater affinity for A41 ( Table 2 ) .
This paper provides a structural and functional characterisation of the VACV A41 protein . This protein was reported previously to be secreted from VACV-infected cells and to affect virus virulence [19] and the immune response to infection [20] . Although the A41 protein shares amino acid similarity with the family of poxvirus chemokine binding proteins ( vCCI ) , hitherto its ligand ( s ) and mechanism of action were unknown . Here we demonstrate that A41 binds a subset of CC chemokines ( CCL21 , 25 , 26 and 28 ) but the affinity of A41 for these chemokines is 1–3 orders of magnitude lower than the affinity of VACV protein vCCI for a wide range of CC chemokines . Consequently , A41 cannot disrupt the high affinity interactions of chemokines with their cellular receptors and so is unable to inhibit leukocyte chemotaxis in response to these chemokines . However , the affinity of A41 for these CC chemokines is still high enough to predict that A41 will block the interaction of chemokines with GAGs , and consistent with this , high concentrations of GAGs such as heparin and dextran sulphate disrupted the A41-chemokine interaction . Direct attempts to block the interaction of hCCL28 and heparin , immobilised on the surface of BIACORE chips , with excess A41 achieved a 40% inhibition at 250 nM ( data not shown ) . These experiments are complicated by the large number of binding sites present on heparin ( mol mass 15 , 000 Daltons ) . These observations suggest that A41 functions by blocking the interaction of CC chemokines with GAGs on the endothelial cell surface and thereby disrupting the establishment of a chemokine concentration gradient around the site of infection . So , in the presence of A41 , CC chemokines can still bind to their receptors on leukocytes and activate these cells , but the leukocytes would not home to the site of infection . This model ( Figure 6C ) fits well with the observed increased infiltration of leukocytes into dermal tissue infected with a VACV strain engineered to lack the A41L gene [19] and is similar to the proposed mode of action of the MYXV M-T7 protein [15] . It also fits well with observations made with mutant chemokines , such as CCL5 , that are deficient in GAG binding . Despite binding and activating chemokine receptors in vitro , these chemokines were unable to recruit cells in vivo , highlighting the importance of GAG binding to form haptotactic gradients for cells to follow [30] . The structure of the A41 protein was solved to 1 . 9 Å resolution and shows considerable similarity to the family of chemokine binding proteins ( vCCI ) from poxviruses , even though these proteins share only ∼19% sequence identity with A41 . Like the CPXV , ECTV and RPXV vCCI proteins [16–18] , A41 is composed of a β sandwich comprising two parallel β sheets connected and partially covered by extended loops . Four disulphide bonds contribute to the protein stability . The extended loops of A41 differ significantly from other vCCI family proteins and these changes are likely to affect the affinity and specificity of the chemokines bound . A notable difference between A41 and other vCCI proteins is the absence of the 2–4 loop ( acidic β strand 3 , numbered according to [16] ) , which in the RPXV vCCI-CCL4 complex helps to lock the chemokine in place via electrostatic interactions [17] . The absence of this loop in A41 may explain its reduced affinity for chemokines . However , this apart , the basic interaction of vCCI-CCL4 and the modelled A41-CCL26 interaction is remarkably similar . As for vCCI-CCL4 , the predicted binding of A41 with CCL26 is aided by the packing of conserved hydrophobic residues in the N-terminal loop of the chemokine in a hydrophobic depression on the A41 surface . This depression is formed primarily by F181 , which is strictly conserved in A41 and in vCCI proteins from all sequenced orthopoxviruses . The selectivity of A41 for CCL21 , 25 , 26 and 28 is likely to be influenced by the lack of the 2–4 loop in A41 and by the insertion of residues in the 20s and 40s loops of these chemokines ( Figure 5E ) . The nature of the chemokines bound by A41 also provides an explanation for the increased immunogenicity of VACV strains engineered to lack the A41L gene , and in particular the induction of enhanced levels of antigen specific CD8+ T cells in the secondary lymphoid organs [20] . CCL21 is a pivotal molecule for priming T-cell responses , co-stimulating the expansion of naïve CD4+ and CD8+ T cells and inducing Th1 polarization [31] . It recruits CCR7+ T cells and DCs into the lymph nodes [32] and is responsible for the movement of CD4+ T cells within the lymph node [33 , 34] . CCL21 is up-regulated during a febrile response , promoting uptake of lymphocytes into the lymph nodes across the high endothelial venules [35] . This involvement in the formation and maintenance of a specific anti-viral immune response means that CCL21 is a logical target of many viral proteins . Notably , simian and human immunodeficiency viruses [36] , hepatitis C virus [37] and murine γ-herpesvirus 68 [38] have evolved different ways to block its activity . CCL25 is also involved in the formation of a T-cell response , although its expression is mainly localised in the thymus and small intestine . It is responsible for the homing of CCR9-expressing T-cell progenitors to , and their migration through , the thymus [39 , 40] . It is also important for the development of immune responses in the gut mucosa . CCL28 is also expressed by the mucosal epithelia of the gut , where it attracts CD4+ and CD8+ resting T cells [41] and has broad antimicrobial properties [42] . An impairment of its effector functions by A41 would be expected to affect the T-cell response , but probably in the mucosa rather than the lymphoid organs . It is interesting that VACV encodes two soluble proteins that bind CC chemokines ( vCCI and A41 ) . Of these , A41 is the more conserved and is expressed by every VACV strain tested [19] and indeed every sequenced orthopoxvirus species and strain [9] , now more than 75 viruses . In contrast , vCCI is expressed by only nine out of 15 VACV strains tested [13] . However , each protein can influence virus virulence: deletion of gene A41L from VACV strain WR , which does not express vCCI [13] , caused enhanced cellular infiltration and reduced virulence in rabbit skin [19]; conversely , insertion of the vCCI gene from VACV strain Lister into VACV strain WR , which expresses A41 , attenuated the virus in a murine intranasal model , characterized by reduced mortality and weight loss , decreased virus replication and spread , and a reduced recruitment of inflammatory cells into the lungs [43] . These proteins therefore have distinct roles , consistent with their different binding specificities , affinities and modes of action . In the light of the results for A41 we can reformulate the model whereby vCCI-like molecules block the chemokine receptor epitope , whilst A41 or MT-7-like molecules block the GAG binding epitope of the chemokines . In our revised model ( Figure 6C ) the binding sites on chemokines for GAGs and chemokine receptors largely overlap and the functional distinction between the two classes of molecules arises simply from the differences in binding affinity . vCCI binds sufficiently tightly to block the interaction of chemokines with their receptors and GAG binding is also blocked . In contrast , A41 acts by blocking the establishment of a chemokine concentration gradient by competing with GAG binding only . Having both strategies may be advantageous for a virus because some chemokines , such as CCL21 , can exert some of their functions without forming concentration gradients ( a process known as chemokinesis ) [44] . This ability , coupled with the anti-viral potency of CCL21 , may mean that some viruses have evolved more than one way to target this chemokine and inhibit its various functions .
The VACV strain Western Reserve ( WR ) A41L gene was amplified from plasmid pA41 [19] by PCR using primers GGGGATTAATATGTACTCGTTAGTATTTG and GGGGGAATTCTTAACAATTATCAAATTTTTTC . The DNA was digested with AseI and EcoRI and ligated into plasmid pVOTE2 [22] that had been digested with NdeI and EcoRI . The sequence of the A41L open reading frame within the resultant plasmid pVOTE2-A41 was verified by DNA sequencing . Virus VOTE-A41 was constructed by transfection of pVOTE2-A41 into RK13 cells infected with virus vT7lacOI [22] , and selection of recombinant virus using methodology described previously [22] . To express A41 in mammalian cells , monolayers of RK13 cells were infected with VOTE-A41 at 5 p . f . u . / cell for 90 min , washed and then incubated in culture medium containing 3 mM IPTG and 190 mM NaCl . At 24 h p . i . A41 was purified from the culture medium . The medium was concentrated ∼50-fold using centrifugal concentrators with a 10 kDa cut-off ( Amicon ) . Virus particles were inactivated by treatment with 20 μg/mL psoralen and long-wave UV for 20 min on ice [45] and removed by ultracentrifugation ( 80 min , 35 , 000 x g ) . The supernatant was then applied to a 20 ml Resource Q column ( GE Healthcare ) pre-equilibrated with 2 column volumes of TE ( 50 mM Tris-HCl pH 7 . 0 , 1 mM EDTA ) containing 50 mM NaCl . Unbound proteins were washed out with two column volumes of low salt ( 50 mM NaCl ) TE buffer , and bound proteins were then eluted with a salt gradient ( 50 – 400 mM NaCl ) . The protein content of each fraction was analysed by SDS-PAGE , Coomassie blue staining and immunoblotting with anti-A41 antibody [19] . Fractions containing the A41 protein peak were pooled and concentrated to 200 μl using a 2 ml centricon ( Amicon ) . The concentrated samples were then purified further by SEC on a Superdex S75 HR 10/30 column in PBS . Fractions containing A41 were pooled and concentrated to 200 μl using a 2 ml centricon ( Amicon ) . SDS-PAGE showed a single 30 kDa protein that reacted with the anti-A41 antibody by immunoblotting . The carboxymethylated dextran surface of a CM5 sensor chip ( GE Healthcare ) was activated with 50 mM N-hydroxysuccinimide ( NHS ) and 200 mM N-ethyl-N′- ( 3-dimethylaminopropyl ) -carbodiimide hydrochloride ( EDC ) . Proteins A41 and ovalbumin in 10 mM sodium acetate pH 4 were then coupled to different flowcells on the chip at a rate of 5 μL/min using the BIACORE Application Wizard to a target of 1000 response units ( RU ) for the binding experiments . The chip surface was then deactivated and unbound protein was removed by injecting a pulse of 1 M ethanolamine hydrochloride pH 8 . 5 across all flowcells . To perform screening for analyte binding , the murine chemokines CCL1 , CCL2 , CCL4 , CCL6 , CCL7 , CCL8 , CCL9/10 , CCL11 , CCL12 , CCL19 , CCL20 , CCL21 , CCL22 , CCL24 , CCL27 , CCL28 , CXCL2 , CXCL5 , CXCL9 , CXCL10 , CXCL11 , CXCL12a , CXCL12b , CXCL13 , XCL1 , CX3CL1 , and the human chemokines CCL14 , CCL15 , CCL16 , CCL17 , CCL18 , CCL23 , CCL25 , CCL26 , CXCL3 , CXCL4 , CXCL6 and CXCL7 ( Peprotech Ec ) were passed over the chip surface sequentially , at 50 nM ( 25 μL/min , 2 min injection ) in HBS-EP buffer ( 10 mM Hepes pH 7 . 4 , 150 mM NaCl , 3 mM EDTA , 0 . 005% surfactant P20 ) . The surface of the chip was regenerated in between each analyte with 10 mM glycine pH 3 . The value observed from the empty flowcell ( Fc1 ) was deducted electronically from the other flowcells . For kinetic analyses of A41 binding , a target of 250 RU of purified A41VOTE and A41E . coli was coupled to the surface of a new CM5 sensor chip . Various concentrations ( 3 . 125 nM – 50 nM ) of each chemokine in HBS-EP were passed over the chip ( 50 μl/min , 2 min injection ) . To calculate the dissociation constant ( Kd ) for each analyte , data were analysed using a simultaneous ka/kd fitting procedure and a Langmuir ( 1:1 ) model . The quality of the fit was assessed using the residuals plot and the χ2 value . To investigate if GAGs inhibited the binding of A41 and chemokines , a new chip was prepared by immobilising 1000 RU of purified A41VOTE and A41E . coli as described above for the binding assays . Chemokines were incubated at room temperature for 1 h in the presence of increasing concentrations ( 0 , 31 . 25 , 62 . 5 , 125 , 250 , 500 ng/ml ) of heparin sodium salt ( low molecular weight , MW 4 , 000–6 , 000 , Sigma-Aldrich ) or various other sulphated GAGs ( 125 or 250 ng/ml heparan sulphate sodium salt , chondroitin sulphate B sodium salt ( =dermatan ) , chondroitin 6-sulphate sodium salt ( =chondroitin sulphate C ) , hyaluronic acid and dextran sulphate , all Sigma-Aldrich ) and BIACORE analysis performed as above . Assays were performed as described previously [46] . Briefly , different concentrations of A41 or Ova were incubated with mouse ( m ) or human ( h ) CCL21 , CCL25 or CCL28 or human CCL26 , in 31 μl in the bottom well of 96-well ChemoTxTM plates ( Receptor Technologies , Oxford , UK ) . Five to 8 samples were analysed for each concentration . A 5 μm filter was placed onto the wells and 2 × 105 target cells in 20 μl were placed onto the filter above each well . Plates were incubated for 5 h at 37 °C and cells that had migrated through the membrane were counted with a haemocytometer . Target cells were mouse L1 . 2 cells that express CCR7 naturally ( used for mCCL21 and hCCL21 ) [47] , 4DE4 cells expressing CCR3 ( used for CCL26 ) [25] and L1 . 2 cells transfected transiently with plasmids expressing wild type or N-terminally HA-tagged human CCR9 and CCR10 ( http://www . cdna . org/ ) as described [48] . The surface expression of HA was verified before each experiment by FACS analysis , as described [49] , using a monoclonal anti-HA antibody ( Covance ) at 1/100 dilution , and FITC-conjugated goat anti-mouse at 1/20 dilution ( Dako ) and the appropriate isotype control ( Sigma-Aldrich ) . Initial experiments were performed using wild type and HA tagged alleles to ensure that the HA tag did not interfere with receptor function . To determine the concentration of each chemokine that induced the optimal chemotaxis , chemokine concentrations between 5 and 150 nM were tested . Concentrations used were 10 nM for hCCL21 and mCCL21 , 20 nM for mCCL25 , 50 nM for hCCL25 , 40 nM for CCL26 , 100 nM for mCCL28 and 60 nM for hCCL28 . The A41L gene ( encoding residues 21–219 , excluding the secretion signal ) was amplified by PCR from VACV strain WR DNA with KOD Hot Start DNA polymerase ( Novagen ) using forward primer 5′-ggggacaagtttgtacaaaaaagcaggcttcgaaggagatagaaccatggcacatcaccaccaccatcacGATGATAAATCGGTATGCGATTC-3′ and reverse primer 5′-ggggaccactttgtacaagaaagctgggtctcaTTAACAATTATCAAATTTTTTCTTTAATATTTTACG-3′ . The forward primer encoded a start codon and an N-terminal His6 tag . Both primers featured the attB site of the gateway cloning system ( Invitrogen ) that was used to subclone the purified PCR product into the pDEST14 expression vector . The expression plasmid was shown to be mutation free by sequencing . Native A41 was expressed in the E . coli strain Rosetta ( DE3 ) pLysS . Cultures were grown at 37 °C for 4 h in GS96 medium ( Invitrogen ) , supplemented with Overnight Express Autoinduction System 1 ( Novagen ) , before being incubated at 25 °C overnight . Selenomethionine ( SeMet ) labelled protein was expressed in the E . coli methionine auxotrophic strain B834 ( DE3 ) ( Novagen ) . Cells were cultured at 37 °C in SeMet medium ( Molecular Dimensions Ltd . ) supplemented with 40 mg/l SeMet , until the OD595 reached 0 . 6 . Cultures were then cooled to 20 °C , A41 expression was induced by addition of 0 . 5 mM Isopropyl-βD-thiogalactopyranoside and the culture was left to incubate overnight . All cultures were harvested by centrifugation ( 6000 × g , 4 °C , 20 min ) and stored at −80 °C until use . Cell pellets were resuspended in phosphate-buffered saline ( PBS ) with 0 . 5 % ( v/v ) Tween-20 , and lysis was performed by sonication ( Sonics Vibracell ) on ice . Inclusion bodies , comprising mainly A41 , were isolated by centrifugation ( 30 , 000 x g , 8 °C , 10 min ) and subjected to several washes in triton wash buffer ( 50 mM Tris-HCl pH 8 . 0 , 100 mM NaCl , and 0 . 5 % ( v/v ) triton X-100 ) with centrifugation ( as above ) between washes to re-collect inclusion bodies . A final wash was performed in detergent free buffer ( 50 mM Tris-HCl pH 8 . 0 and 100 mM NaCl ) . The inclusion bodies were then dissolved overnight at 4 °C in a denaturing buffer containing 50 mM Tris-HCl pH 8 . 0 , 100 mM NaCl , 6 M guanidine hydrochloride and 10 mM DTT , followed by centrifugation ( 30 , 000 × g , 8 °C , 30 min ) to remove undissolved waste . The supernatants containing denatured A41 were stored at −20 °C before refolding by rapid dilution with stirring of 20 mg of denatured inclusion bodies into 200 ml of refolding buffer containing 200 mM Tris-HCl pH 8 . 0 , 1 M L-Arginine , 6 . 5 mM cysteamine , 3 . 7 mM cystamine , plus 1 EDTA-free protease inhibitor cocktail tablet ( Roche ) . Refolding reactions were incubated at 4 °C for between 24–48 h , followed by concentration to a volume of 5 ml in a vivacell 250 concentrator ( Vivascience ) with a 10 kDa cut-off . Concentrated protein was purified by SEC on a HiLoad 16/60 Superdex 200 column ( GE Healthcare ) , into a final buffer of 20 mM Tris-HCl pH 7 . 5 and 200 mM NaCl . One hundred percent incorporation of selenomethionine was confirmed by mass spectrometry . Prior to crystallization , purified A41 was concentrated by ultrafiltration to 3 mg/ml in a 10 kDa cut-off concentrator ( Vivascience ) . For both native and SeMet-substituted A41 , initial vapour diffusion crystallization experiments were performed at 21 °C in 300 nl drops ( protein/precipitant ratio of 2:1 ) using a Cartesian robot [50 , 51] . Crystals were grown from a mother liquor of 0 . 2 M potassium fluoride , 20 % polyethylene glycol 3350 . Based on this result , further optimisations were performed [52] . Crystals of A41 belong to space group P21 , ( unit cell dimensions a = 36 . 6Å , b = 60 . 8Å , c = 50 . 4Å , β = 91 . 0° ) , and contain one molecule per asymmetric unit with an estimated solvent content of 50 % . Both native and SeMet crystals were flash frozen at 100K in mother liquor containing 20 % glycerol . Diffraction data to 1 . 9 Å resolution were collected for the native crystals at the European Synchrotron Radiation Facility ( ESRF ) , beamline ID14 EH1 . A multiple wavelength anomalous diffraction ( MAD ) analysis was performed at 2 . 3 Å resolution on a SeMet crystal at beamline BM14 at the ESRF . All oscillation images were processed and reduced using the HKL software suite [53] ( Table 1 ) . The structure of A41 was solved using the MAD method . Each monomer of A41 contains four SeMet residues and the positions of these were determined using SHELXD [54] and initial phases computed with SHELXE [55] as part of the HKL2MAP package [56] . Density modification used RESOLVE [57] , and the resulting electron density map was of excellent quality , allowing an automatic chain trace to be performed with Arp/wARP [58] , which built 160 out of 199 residues . The remainder of the structure was modelled in the program O [59] . Initial refinement was performed using program CNS [60] . The 1 . 9 Å native data were isomorphous to the SeMet data , permitting refinement to be directly extended to this higher resolution in the program REFMAC [61] with iterative rebuilding in COOT [62] . For the final model the Rwork is 19 . 4 % and the Rfree is 24 . 9 % ( Table 1 ) . The stereochemical quality of the structure was assessed using the MOLPROBITY program [63] . The structure has good stereochemistry with 94 . 3 % of residues lying in the most favoured regions of the Ramachandran plot . The CC26-A41 complex was modelled by superimposing A41 onto the vCCI component of the RPXV vCCI-CCL4 complex ( pdb code 2FF3 ) ( 159 equivalences with an rmsd of 2 . 4Å ) , and then superposing CCL26 ( pdb code 1G2S ) onto the CCL4 component ( 59 equivalences with an rmsd of 1 . 8 Å ) . All superpositions were done automatically using SHP [28] .
Coordinates and structure factors have been deposited in the Protein Data Bank ( PDB; http://www . rcsb . org/pdb/ ) with accession numbers 2vga and 2vgasf .
|
As part of the innate immune response ( for example to virus infection ) , the body produces proteins called chemokines , which act by directing white blood cells ( leukocytes ) to the areas of infection and inflammation . Viruses have evolved mechanisms to fight this immune response . Indeed , so important is this need to protect themselves from the immune system that some viruses , such as poxviruses , devote up to half their genetic information to this battle . We have studied a protein called A41 , one component of the response of vaccinia virus ( the vaccine used to eradicate smallpox ) to the immune system and shown that it interferes with the function of a group of chemokines . These chemokines function by forming concentration gradients along which the white blood cells migrate , and A41 sequesters the chemokines , thereby preventing formation of the gradient . Interestingly , we show also that A41 is very similar in structure to another group of proteins , called vCCIs , that bind chemokines more tightly , blocking their attachment to white blood cells , suggesting that both mechanisms are important for virus virulence .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"viruses",
"biophysics",
"virology"
] |
2008
|
Structure and Function of A41, a Vaccinia Virus Chemokine Binding Protein
|
Most organisms are simply diamagnetic , while magnetotactic bacteria and migratory animals are among organisms that exploit magnetism . Biogenic magnetization not only is of fundamental interest , but also has industrial potential . However , the key factor ( s ) that enable biogenic magnetization in coordination with other cellular functions and metabolism remain unknown . To address the requirements for induction and the application of synthetic bio-magnetism , we explored the creation of magnetism in a simple model organism . Cell magnetization was first observed by attraction towards a magnet when normally diamagnetic yeast Saccharomyces cerevisiae were grown with ferric citrate . The magnetization was further enhanced by genetic modification of iron homeostasis and introduction of ferritin . The acquired magnetizable properties enabled the cells to be attracted to a magnet , and be trapped by a magnetic column . Superconducting quantum interference device ( SQUID ) magnetometry confirmed and quantitatively characterized the acquired paramagnetism . Electron microscopy and energy-dispersive X-ray spectroscopy showed electron-dense iron-containing aggregates within the magnetized cells . Magnetization-based screening of gene knockouts identified Tco89p , a component of TORC1 ( Target of rapamycin complex 1 ) , as important for magnetization; loss of TCO89 and treatment with rapamycin reduced magnetization in a TCO89-dependent manner . The TCO89 expression level positively correlated with magnetization , enabling inducible magnetization . Several carbon metabolism genes were also shown to affect magnetization . Redox mediators indicated that TCO89 alters the intracellular redox to an oxidized state in a dose-dependent manner . Taken together , we demonstrated that synthetic induction of magnetization is possible and that the key factors are local redox control through carbon metabolism and iron supply .
In biology , magnetism is a unique and virtually orthogonal physical property . As magnetic interactions can be contactless , remote , and permeable , integration of magnetic properties into biological systems provides another dimension for bioengineering and therapy . Magnetic functions may provide a unique interface between cells; for example , magnetic sensing as an input and induced magnetization as an output would allow not only magnetic manipulation but also magnetometric readout such as magnetic resonance imaging ( MRI ) . Only a few natural systems are known to exploit magnetic function . Magnetotactic bacteria produce a chain of organelles called magnetosomes [1] , in which ferromagnetic magnetite ( Fe3O4 ) or greigite ( Fe3S4 ) particles are formed ( reviewed in [2] ) . The cells orient and swim along geomagnetic field lines locating better growth conditions more efficiently than random swimming ( reviewed in [3] ) . In the genomes of these bacterial species , specific clusters of genes called magnetosome gene islands are conserved [2] . The recent comprehensive study has revealed several specific genes participating in various steps of formation of the magnetosome [4] , showing the complexity of biogenesis of the organelle . A putative iron transporter gene MagA from magnetotactic bacteria [5] has been shown to be sufficient for producing MRI-detectable iron-containing particles in mammalian cells [6] , [7] . However , MagA gene does not belong to the magnetosome gene island and the MagA protein localizes to the plasma membrane in Magnetospirillum magneticum strain AMB-1 [8] , [9] . To date , no other successful transgenic study for magnetosomal function has been reported . Members of magnetotactic bacteria identified so far belong to α-proteobacteria , δ-proteobacteria , Nitrospira ( reviewed in [2] ) , and γ-proteobacteria [10] , while intracellular magnetic inclusions were also found in Shewanella putrefaciens [11] and a photosynthetic purple bacteria [12] and a geo-biological study has presented magneto-fossils that are too large for bacteria—a possible remnant of eukaryotic biogenic magnetic particles [13] . Migratory animals sense geomagnetic fields , an ability called magneto-reception ( reviewed in [14] ) . The radical pair and the magnetite hypothesis are the two proposed modes for the mechanism of magneto-reception . The former may exploit a photochemical reaction affected by the magnetic field , and the latter involves small magnetic particles in the nervous system . To date , neither physiological nor molecular mechanisms for the formation of such magnetic particles in nerve tissue are understood . However , anomalous deposition of iron is found to be associated with many neurodegenerative disorders in humans , such as Alzheimer's , Parkinson's , and Huntington's disease ( reviewed in [15] , [16] ) . Being an essential element for life , iron is reactive and prone to precipitate under physiological conditions . Some modification of iron homeostasis is assumed to induce iron mineralization in nerve tissue . Recently , an iron-export ferroxidase activity of β-amyloid precursor protein [17] was identified , indicating the importance of iron homeostasis in nerve tissue integrity . Owing to its ability to catalyze the formation of reactive oxygen radical species , high concentration of free iron ions could thus be toxic . Cells may deal with this by producing ferritin , a ubiquitous iron sequestrating protein present from bacteria to human [18] . Ferritin oligomerizes to form a shell of 12–15 nm in diameter in which iron is sequestered and mineralized . Iron mishandling by ferritin causes neurodegeneration with iron deposition called neuroferritinopathy [19] . Naturally isolated ferritin from horse spleen contains paramagnetic ferrihydrite ( 5Fe2O3•9H2O ) , and every single core ( 8–10 nm at most ) is in principle too small to be manipulated magnetically . However , the composition may vary depending on the chemical environment as magnetite-formation of ferritin was demonstrated in vitro [20] exhibiting superparamagnetic behavior at room temperature . Here , we address these issues by exploring the synthetic induction of bio-magnetization using a model organism budding yeast Saccharomyces cerevisiae .
For biogenic magnetization , significant amounts of magnetic compounds need to be formed inside the organism . This may be achieved by altering iron homeostasis either physiologically or genetically . Since ferrous iron is prone to oxidation to insoluble ferric iron , citric acid , a chelator of the ferric ion , can be included to prevent precipitation with no impact on biological availability . Wild-type yeast cells can grow at as high as 5 mM ferrous ( Fe2+ ) ascorbate or 20 mM ferric ( Fe3+ ) citrate ( Figure 1A , wild type ) . Ferric citrate was less toxic and can thus be used to deliver iron to yeast without damaging the cells or forming precipitates in the media . Yeast cells lack ferritin and sequester iron in their vacuoles . The vacuolar iron transporter Ccc1p plays a major role in iron sequestration , loss of which abolishes iron tolerance [21] . Human ferritin genes consist of ferritin heavy chain FTH , ferritin light chain FTL , and the iron chaperone PCBP1 [22] . As described previously , the ccc1 knockout strain ( ccc1Δ ) showed intolerance at 5 mM ferrous ion while as high as 20 mM ferric citrate is required to see intolerance of ccc1Δ ( Figure 1A ) , suggesting mitigated iron toxicity of ferric citrate . We found that single copy expression in yeast of the human ferritin gene set conferred iron tolerance to ccc1Δ both in ferrous and ferric supplements ( Figure 1A ) , indicating that ferritin efficiently sequesters iron in these conditions . The four strains ( wild type containing empty plasmid , ferritin-expressor , ccc1Δ , and ccc1Δ ferritin-expressor ) were cultured in 20 mM ferric citrate liquid medium and tested for magnetization . The cell cultures were exposed to magnets and attraction was observed . Attraction of ccc1Δ ferritin-expressor was detectable as early as 2 min after exposure . After 10 min , attraction of all strains became observable ( Figure 1B and Video S1 ) . For quantitative characterization of the magnetic properties of the yeast cells , a superconducting quantum interference device ( SQUID ) was used . The cells were subjected to a measurement of their magnetic moment at 300 K at various magnetic fields to analyze field-dependent magnetization . Without ferric citrate supplementation all the four strains similarly exhibited negative values proportional to the applied field ( Figure 1C , no iron supplemented ) , indicating that they are diamagnetic . As is the case for most biological materials , their mass magnetic susceptibility ( m3•kg−1 ) was comparable to that of water ( −9 . 051×10−9 ) ( Table 1 ) . When supplemented with ferric citrate , all the strains exhibited positive values . At high fields ( 2 , 500 to 10 , 000 Oe ) , magnetization is proportional to field and not saturating , indicating a dominant contribution of paramagnetism . At low fields ( 0 to 2 , 000 Oe ) , an upward concave curve of magnetization was observed , indicating additional ferro/ferri-magnetic contribution , which typically saturates within this region . This suggests that the cells contain mostly paramagnetic ( or superparamagnetic ) material with a slight amount of ferro/ferri-magnetic material . Mass magnetic susceptibility of the paramagnetic constituent was given based on values at high fields ( Table 1 ) . Those of ferritin-expressor , ccc1Δ , and ccc1Δ ferritin-expressor were approximately 1 . 3 , 1 . 8 , and 2 . 8 times larger than that of wild type , respectively . Previous studies on magnetic susceptibility of isolated ferritin ranged from 3 . 7×10−8 to 9 . 4×10−8 m3/kg ( originally 2 . 95×10−6 to 7 . 5×10−6 em in cgs unit ) at room temperature , depending on the sample and measuring method [23] ( reviewed in [24] ) . Thus , we observed a gain of magnetic susceptibility due to ferritin expression , while a non-ferritin contribution was also present , indicating that ferric citrate supplementation induces basal magnetization in yeast . ccc1Δ showed increased magnetization compared to wild type , suggesting that non-vacuolar iron may have more magnetic contribution than previously thought . The synergistic effect of ferritin and ccc1Δ can be explained by higher availability of iron to ferritin in the cytosol . Ultrathin section transmission electron microscopy showed accumulation of electron-dense deposits ( Figure 2 ) . Although these varied in shape , size , and amount among cells , wild type cells typically contained round particles associated with membranous structures that are most likely the vacuoles ( Figure 2 wild-type ) , while ccc1Δ cells tended to contain aggregates within mitochondria ( Figure 2 ccc1Δ ) . As the mitochondria are where cells convert inorganic iron into heme and iron-sulfur clusters , the observed deposits could be caused by iron overload due to the defect in vacuolar iron sequestration , and may contribute to the higher magnetic susceptibility in ccc1Δ . Ferritin expression had little observable effect on the electron micrographs ( Figure 2 ferritin ) perhaps due to the small size of the iron binding center . To reveal the elemental composition of the electron-dense deposits , magnetized cells were analyzed by energy-dispersive X-ray spectroscopy ( EDS ) . Elemental maps were obtained for detectable elements , and iron , phosphorous , oxygen , and nitrogen showed characteristic distributions associated with cellular structures ( Figure 3A and 3B ) . Nitrogen was distributed throughout the cell consistent with association with biogenic molecules such as proteins . In wild type cells grown in ferric citrate , phosphorous , iron , and oxygen were slightly concentrated within membranous structures presumably vacuoles and in electron-dense round particles ( Figure 3A , magnified images ) . In ccc1Δ cell , iron showed increased localization to the clusters of electron-dense crystals ( Figure 3B ) . Phosphorous also accumulated in the clusters . Oxygen showed a similar pattern to phosphorous with less contrast . These two types of electron-dense deposits ( small round particles in wild type and clustered crystals in ccc1Δ ) thus contained iron , oxygen , and phosphorous with different composition stoichiometries ( Figure 3E and 3F ) . The elemental maps were further analyzed to estimate relative amounts of iron and phosphorous ( Figure 3G and 3H ) ; iron was higher in ccc1Δ than in wild type cells , whereas phosphrous showed only a small increase in ccc1Δ . Magnetic columns have been used for separation of biomaterials labeled with magnetic particles . To test if our yeast cells behave similarly , the cells were applied to a magnetic column . Normally grown yeast cells were not retained on the column under the conditions tested ( Figure 4A , normal ) . The cells supplemented with ferric citrate were retained by the magnetized column ( Figure 4A and 4B ) . Among the four strains , the order of rate of trapped cells is in agreement with their magnetic susceptibility measured by SQUID , indicating that this system can be used for comparison of cell magnetization , as well as to separate magnetic cells . Genetic control of magnetization would greatly expand the engineering potential of magnetic cells . To explore this possibility as well as to gain further insight into the nature of the biogenic magnetization , we sought yeast gene knockout strains that show altered magnetization . Candidate genes to be tested were selected based on their functional or phenotypic description associated with iron homeostasis or oxidative stress from Saccharomyces Genome Database ( http://www . yeastgenome . org ) . Mutant strains were grown in 20 mM ferric citrate medium and their attraction towards a magnet was observed ( Figure 5A ) . Strains showing reproducible altered attraction were selected and subjected to magnetic column separation to confirm and quantify their magnetization . From the initial screen of 60 strains ( Table S1 ) , tco89Δ was found to show consistent reduction of magnetization compared to wild type ( Figure 5A and 5B ) . Tco89p is known to be a nonessential component of TORC1 [25] . TORC1 globally regulates cell growth in response to nutrient , stress , and redox states ( reviewed in [26] , [27] ) . To ask if and how TORC1 is involved in the magnetization , Tor1p , the other nonessential component of TORC1 and Ssd1p , which coordinates with TORC1 to maintain cell integrity [25] , was tested . Both tor1Δ and ssd1Δ showed little change in the magnetization ( Figure 5B ) , indicating that challenged cell integrity is not associated with the magnetization . Magnetization of the cells positively correlated with copy number of TCO89 ( Figure 5C ) , showing a dose-dependent effect of TCO89 on the magnetization . Expressing TCO89 under a galactose inducible promoter pGal1 showed induction of magnetization ( Figure 5D ) . Loss of TCO89 in ccc1Δ decreased magnetization ( Figure 5E ) , suggesting that iron sequestration into the vacuole by CCC1 does not have a predominant effect on induction of magnetization by TCO89 . In contrast , TCO89 affects magnetism through TORC1 activity . We used rapamycin , an inhibitor for TORC1 at sub-lethal doses . Rapamycin treatment reduced magnetism in wild type and more prominently in multi-copy TCO89 , while no reduction is observed in tco89Δ ( Figure 6A ) , indicating that induction of magnetism by TCO89 is through TORC1 activity . As TORC1 processes nutritional signals , we tested if the nutritional environment affects magnetism . Compared to synthetic-defined medium , we observed a reduction of magnetism when cells were grown in rich medium with the same amount of iron ( Figure 6B ) . The magnetism then increased as extra glucose was added to rich medium . In contrast , addition of extra nitrogen ( i . e . , amino acids and nucleotides ) to synthetic defined medium decreased magnetism , suggesting that the relative availability of carbon and nitrogen has impact on the formation of magnetism . Effect of TCO89 became less prominent in rich medium or when extra nitrogen was added . These results indicate that higher carbon availability has a positive effect on magnetization , which is enhanced by TOC89 . As iron homeostasis has a close relationship with redox state , we asked if TCO89 has any function associated with redox control . Cellular redox activity can be monitored by a biocompatible redox indicator methylene blue , which loses its color when reduced . Equal numbers of cells were spotted and grown on plates containing methylene blue to observe colony staining . Compared to wild-type or plasmid-complemented cells , tco89Δ exhibited little color while multi-copy TCO89 cells were blue ( Figure 7A ) , indicating that TCO89 leads cellular redox to an oxidized state in a dose-dependent manner . tco89Δ also showed compromised cell growth in the presence of methylene blue presumably because a higher rate of methylene blue reduction interferes with cellular metabolism . Nicotinamide adenine dinucleotide phosphate ( NADP ) is a coenzyme that serves as a redox mediator in protection against oxidative stress . Cells harboring multi-copy TCO89 showed higher levels of both NADP+ ( oxidized ) and NADPH ( reduced ) , while tco89Δ had slightly lower NADP+ and higher NADPH ( Figure 7B ) . We expanded genetic screening for candidates related to carbon metabolism and redox using the magnetic column entrapment procedure ( Figure S1 ) . Gene knockouts affecting oxidative damage , such as GRX2 , GRX3 , and SOD2 , did not show significant changes , while POS5 , a gene for mitochondrial NADH kinase , showed reduction in magnetism . In contrast , UTR1 did not affect magnetism , which is a cytoplasmic ATP-NADH kinase . Gain of magnetism was seen with loss of YFH1 , which has been reported to accumulate iron in mitochondria [28] , [29] , and whose human homolog FXN is responsible for the neurodegenerative disease Friedreich's ataxia [30] . Regarding carbon metabolism , gene knockouts for SNF1 and ZWF1 showed reduction in magnetism . SNF1 is required for processing carbon stress signals ( reviewed in [31] ) and ZWF1 codes for an enzyme at the branch point of the pentose phosphate pathway [32] .
We demonstrated that generation of bio-magnetization in yeast is possible by three ways: modulating iron homeostasis , introducing iron crystallizing proteins , and controlling redox state . Intracellular redox state is normally sufficiently reductive to allow iron to exist as soluble Fe2+ . Oxidation of Fe2+ to Fe3+ facilitated by TCO89 led to an oxidative state that induces iron precipitation and yeast magnetization . The importance of redox state in magnetization offers insight into magnetotactic bacteria . Most of these bacteria thrive exclusively in microaerobic environments , showing their strong preference to certain redox conditions . Evolutionally , formation of bio-magnetic particles may have originated as a consequence of redox mediation and/or iron sequestration , although today's magnetosomes seem so specified to magnetic function that once formed they cannot be utilized as resources by the cell [3] , [33] . In the gradient of oxygen/iron distribution in deep aqueous environments , there may exist cells adapted to certain redox and chemical conditions that were optimal for the formation of magnetite or greigite [34] . In such cells magnetization and redox metabolism could have been linked . Once magneto-aerotaxis was established , the role of iron as a redox mediator became less important as seen in today's magnetotactic bacteria . Although TORC1 activity has been suggested to be redox-sensitive [35] , [36] , its redox control has not been demonstrated . Considering TORC1 function in the regulation of carbon metabolism and energy production , which are major sources of redox flux , TCO89 may control redox state through these functions . Indeed , carbon availability showed a significant impact on magnetism . In glucose-rich conditions , glycolysis and fermentation is preferred over mitochondrial TCA cycle and oxidative phosphorylation . As the TCA cycle generates reducing equivalents that then reduce oxygen , down-regulation of mitochondrial TCA cycle may result in a shift to a more oxidized state inside mitochondria . In agreement , recent studies have detected ferric phosphate nano-particles in mitochondria from fermenting yeast [37] , [38] . Combined with the results that iron deposits were found in mitochondria of ccc1Δ , oxidizing conditions in mitochondria could facilitate iron deposition that has greater magnetism . Iron deposition in neurodegeneration may also be attributed to a failure in mitochondrial redox control in concert with energy metabolism because of the brain's high demand for energy and oxygen . Redox control may also aid in applications that involve biogenic metal precipitation such as bioremediation and nano-particle production . Because our choice of organism was not based on any potential of magnetization of yeast , these ideas could be easily applied to a number of other organisms to confer them with similar or possibly greater properties , or to find key components for redox control .
Constructs were made via a BioBrick assembly method [39] , [40] . Ferritin genes , which consist of FTL , FTH1 , and Pcbp1 , were obtained from ATCC mammalian gene collection . The open reading frames of the genes were PCR-amplified and assembled under control of the CCC1 promoter from yeast . The genes were then cloned into yeast single-copy plasmid pRS316 or multi-copy plasmid pRS416 [41] . Plasmids were transformed into BY4741 ( MATa his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 ) or ccc1Δ ( his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 ccc1::kanMX4 ) . All knockout strains were obtained from Saccharomyces Genome Deletion Project [42] that have BY4741 background with gene deletion by kanMX4 . TCO89 gene including promoter and terminator region was PCR amplified and cloned into pRS316 or pRS426 . For galactose induction , TCO89 without promoter region was cloned under pGAL1 in pRS316 . Cells were grown in synthetic medium ( 0 . 67% yeast nitrogen base ) or synthetic dropout medium ( 0 . 67% yeast nitrogen base without amino acid , 0 . 2% dropout supplements ) with appropriate carbon source ( 2% glucose , 2% raffinose , or 2% galactose ) at 30°C . Ferric citrate and ferrous ascorbate were freshly added from 1 M stock solutions . Methylene blue was added from 10 mM stock solution . Each strain was pre-cultured in synthetic medium and then diluted into the medium supplemented with 20 mM iron citrate and grown overnight . The cells were collected by centrifugation and re-suspended to give 0 . 5 OD600 . Five ml of the suspensions were layered onto 1 ml Optiprep density gradient medium ( Axis-Shield PoC AS , Norway ) in four-compartmented Petri dishes . Each dish was placed over a black paper sheet and axial pole magnets R848 ( K&J Magnetics , Inc . PA ) aligned 4×4 reciprocally . For magnetic screening , 0 . 9 ml cell suspension was layered on 0 . 1 ml Optiprep density gradient medium in 24-well flat bottom plate . Block magnet BZ084 was attached from the bottom of the plate . Cells were chemically fixed and embedded as described previously [43] , [44] . Ultrathin sections ( 60–80 nm ) were cut on a Reichert Ultracut-S microtome , placed onto copper grids , and stained with 0 . 2% lead citrate . Non-stained sections were also prepared to avoid staining artifacts . Specimens were examined on a JEOL 1200EX-80 kV transmission electron microscope and images were acquired with DITABIS digital imaging plates . Non-stained sections same as for transmission electron microscopy were analyzed on JEOL JEM 2010F at 200 kV with JEOL Dark field STEM detector ( probe size 1 . 0 nm , camera length 15 cm ) . EDS analysis was performed by INCA system ( Oxford Instruments , UK ) . Cells were collected by centrifugation , washed twice with 0 . 6 M sorbitol , dehydrated in −20°C acetone , and freeze-dried . Dried samples were encapsulated and weighted . Direct-current field-dependent SQUID magnetometry was performed at 300 K from 0 to 10 , 000 Oe ( 103 A/4 πm ) using Quantum Design AC and DC Magnetic Property Measurement System . Specific magnetic moment was given by Am2/kg . Cells were collected by centrifugation , suspended in ST ( 0 . 6 M sorbitol containing 0 . 01% Triton-X 100 ) , incubated for 30 min at room temperature , and adjusted to 0 . 5 OD600 . MACS cell separation MS columns ( Miltenyi Biotec Inc . , CA ) were placed in a ring magnet R848 , equilibrated with ST , loaded with 1 ml cell suspension , washed by 1 m ST , then displaced from the magnet , and bound cells were released by 1 ml ST . The unbound fraction ( flow-through and wash-out ) and trapped fraction were measured by OD600 and the percentage of trapped cells was calculated . Cells were grown in synthetic medium and harvested at OD600 0 . 6–0 . 7 . Each culture containing about 108 cells was sampled , left to stand for 10 min , spun down at 2 , 400 rpm for 5 min , cooled on ice , washed with ice-cold PBS plus 0 . 01% Triton-X100 , and subjected to extraction and detection using Fluoro NADP/NADPH ( Cell Technology , CA ) following the manufacturer's instruction . Fluorescence was measured at 540 nm excitation and 590 nm emission in Wallac 1420 Multilabel counter ( PerkinElmer , Finland ) .
|
Most organisms do not respond to magnetic fields . However , “magnetotactic” bacteria and migratory animals can sense geomagnetic fields and alter their behavior accordingly . These organisms often contain small magnetic particles that may be responsible for sensing magnetic fields . In magnetotactic bacteria , specific genes are crucial for the formation of these magnetic particles , but no such genes have yet been characterized in migratory animals . In humans , formation of magnetic particles can be observed in the neuronal tissue in neurodegenerative diseases . One explanation for the appearance of these magnetic particles is that they are the result of alterations in metabolism , which occur in neurodegenerative diseases . Here , we explore this hypothesis by inducing magnetism in yeast cells , which are not naturally magnetic and examine how changes in metabolism contribute to particle formation and magnetism . We find that yeast cells expressing a set of human proteins that sequester iron contain iron particles and become attracted by a magnet when grown with ferric citrate . Through physiological and genetic studies we show that target of rapamycin complex 1 ( TORC1 ) signaling , which responds to nutritional signals , is important for the magnetization of these cells by altering the intracellular oxidation ( or redox ) state . We also show that genes involved in carbon metabolism affect magnetization . We propose that local redox control mediated by carbon metabolism and iron homeostasis , processes that exist in normal unmagnetized cells , are key for iron particle formation and magnetization . We conclude that magnetization of normal cells will be possible with these existing gene sets .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biotechnology",
"organismal",
"evolution",
"evolutionary",
"biology",
"synthetic",
"biology",
"microbiology",
"neuroscience",
"model",
"organisms",
"molecular",
"genetics",
"environmental",
"biotechnology",
"forms",
"of",
"evolution",
"signaling",
"in",
"cellular",
"processes",
"biology",
"molecular",
"biology",
"agriculture",
"biochemistry",
"signal",
"transduction",
"genetics",
"yeast",
"and",
"fungal",
"models",
"molecular",
"cell",
"biology",
"agricultural",
"biotechnology",
"genetics",
"and",
"genomics"
] |
2012
|
Induction of Biogenic Magnetization and Redox Control by a Component of the Target of Rapamycin Complex 1 Signaling Pathway
|
Targeting of permissive entry sites is crucial for bacterial infection . The targeting mechanisms are incompletely understood . We have analyzed target-site selection by S . Typhimurium . This enteropathogenic bacterium employs adhesins ( e . g . fim ) and the type III secretion system 1 ( TTSS-1 ) for host cell binding , the triggering of ruffles and invasion . Typically , S . Typhimurium invasion is focused on a subset of cells and multiple bacteria invade via the same ruffle . It has remained unclear how this is achieved . We have studied target-site selection in tissue culture by time lapse microscopy , movement pattern analysis and modeling . Flagellar motility ( but not chemotaxis ) was required for reaching the host cell surface in vitro . Subsequently , physical forces trapped the pathogen for ∼1 . 5–3 s in “near surface swimming” . This increased the local pathogen density and facilitated “scanning” of the host surface topology . We observed transient TTSS-1 and fim-independent “stopping” and irreversible TTSS-1-mediated docking , in particular at sites of prominent topology , i . e . the base of rounded-up cells and membrane ruffles . Our data indicate that target site selection and the cooperative infection of membrane ruffles are attributable to near surface swimming . This mechanism might be of general importance for understanding infection by flagellated bacteria .
Salmonella enterica subspecies 1 serovar Typhimurium ( referred to as S . Typhimurium in this study ) is a common food-borne pathogen . Central to the pathogenesis of S . Typhimurium is its ability to invade intestinal cells , namely M-cells , epithelial cells and possibly dendritic cells [1] , [2] , [3] . Normally , only a small fraction of the mucosal cells are being invaded [4] , [5] , [6] , [7] . The mechanisms focusing S . Typhimurium invasion to particular sites are not completely understood . Host-cell invasion by S . Typhimurium is the result of a multistep process . This includes: i ) 3-dimensional movement in the gut lumen ( motility , chemotaxis , diffusion ) ; ii ) transient interactions with the mucosal surface and particulate matter within the gut lumen; iii ) reversible binding via adhesins like type 1 fimbriae ( fimH , [8] , [9] ) ; iv ) irreversible “docking” mediated via type III secretion system 1 ( TTSS-1; [8] , [9] ) . This step commits wild type S . Typhimurium to invasion . v ) secretion of bacterial virulence factors , so called effectors , via TTSS-1 into the host cytosol; key effectors include SopE , SopE2 , SopB and SipA . ; vi ) manipulation of the host cell by S . Typhimurium effectors leading to the emergence of prominent membrane ruffles in epithelial- and M-cells [10]; vii ) host cell invasion , which often features the simultaneous entry of several bacteria through the same ruffle . While many steps have been studied in detail before , those steps targeting the pathogen to a particular site ( i . e . steps ii and iii ) have remained enigmatic . Thus , it is still unclear whether S . Typhimurium actively “selects” target sites and which mechanism would enable such a preference . We speculated that motility might affect target site selection . Like many other pathogens , S . Typhimurium employs flagella to orient and move in 3D space [11] , [12] . This has multiple well documented effects on the pathogen-host interaction . The coupling to chemo-sensing systems allows swimming towards nutrient sources ( “chemotaxis”; [13] , [14] ) . Motility is important for the invasion of tissue-culture cells and in the induction of gut inflammation by Salmonella spp . ( in vitro: [15] , [16] , [17] , [18] , [19]; animal model: [20] , [21] , [22] , [23] , [24] ) . Furthermore , the flagella might mediate adhesion [25] or elicit host-cellular signaling responses . It remained unclear whether flagella may also serve additional tasks , i . e . in target site selection . Flagellar rotation propels the bacterium with a velocity of at least 25–55 µm/s ( “motility” [12] , [26] ) . If encountering a host cell , S . Typhimurium is generally assumed to either be “deflected” back into the medium or to initiate a productive infection . However , so far , this step of the infection process has not been studied in detail . In contrast , the interaction of motile E . coli strains and solid surfaces has been extensively studied . On solid surfaces , E . coli slides in large circles , remaining in contact with the surface for extended time periods , a phenomenon called near surface swimming ( NSS ) . Two mechanisms explaining this ability of E . coli to swim along solid surfaces have been proposed ( Fig . 1A , inserts I and II ) . According to the hydrodynamic entrapment theory [27] , [28] , the bacteria experience extensive drag stress at the part of their body close to the surface , causing a “forward” rotation . This rotates the rod-shaped bacterium into an “upright” position . The upright position in turn increases the drag resistance against the fluid , resulting in an opposing rotational force . Ultimately , these two forces are in equilibrium , keeping the bacterial rod at a constant angle towards the surface , thus entrapping the organism in a tilted swimming position . The alternative DLVO model ( Derjaguin , Landau , Verwey , and Overbeek; for a review , see [29] ) explains NSS via electrostatic and van der Waals forces . Nevertheless , both models predict that motile bacteria encountering a solid surface would be “trapped” at the surface and perform a NSS motion . It remains unclear whether NSS may also occur on cellular surfaces and whether this might affect target site selection . Here , we studied target site selection by S . Typhimurium . The initial stages of S . Typhimurium-interaction with cellular or artificial surfaces were analyzed in real time . In this “pre-docking” phase of the infection , bacterial motility was of key importance . It led to characteristic near surface swimming patterns on host-cell surfaces and targeting to sites with a prominent surface topology . Our data suggest a model , in which physical forces emanating from the flagella-driven motility facilitate near-surface swimming and explain the pathogen's target preference during infection . We are discussing possible implications for the disease and for infections by other flagellated pathogens .
To study the initial interactions of S . Typhimurium with cellular surfaces , we employed time-lapse microscopy ( supplementary Videos S1 and S2 ) . HeLa cells , a commonly used tissue-culture model for studying S . Typhimurium invasion , were infected with S . TmΔ4 ( SL1344 sopEE2B sipA; Table 1 ) . S . TmΔ4behaves like S . Tmwt in all aspects of the early host-cell interaction ( steps i to v ) , but cannot trigger membrane ruffling or invasion as it lacks the key effector proteins SopE , SopE2 , SopB and SipA [8] , [9] , [30] . Therefore , S . TmΔ4 allowed us to focus on the initial surface interactions and docking ( Fig . 1A ) . First , the initial surface interactions of S . TmΔ4were analyzed using time-lapse differential-interference contrast ( DIC ) microscopy ( Fig . 1B ) . Strikingly , S . TmΔ4was swimming along the cellular surface for extended time periods ( supplementary Videos S1 and S2 ) . Our subsequent frame-by-frame analysis of the time-lapse videos identified several “stages” of this interaction: In most cases , bacteria went through all stages of interaction before taking off again . Please note that some bacteria stopped until the end of the time-lapse movie . In this experiment , we could not distinguish whether these bacteria were “docking” ( i . e . bacteria that bound irreversibly , e . g . via the TTSS-1 apparatus [8] , [9] , [30] ) , or “stopping” ( bacteria transiently stopping but continuing NSS or taking off after the end of the movie ) . Stopping did not happen randomly but occurred frequently at “obstacles” encountered during NSS . In particular , S . Typhimurium stopped and docked at cells with a round morphology ( i . e . a mitotic cell; see supplementary Video S1 ) . This provided a first indication that a transient stop may “preselect” specific sites for subsequent docking of S . Typhimurium . This would be in line with the preferential docking of S . Typhimurium onto mitotic cells observed in earlier studies ( [31] , [32] , see below ) . For a quantification of these initial bacteria surface interactions we used S . TmΔ4harboring a plasmid conferring constitutive gfp expression ( pGFP;Table 1 ) and time lapse fluorescence microscopy . This allowed precise quantification of all stages of the bacteria surface interaction including landing and take-off , since the fluorescent bacteria moving out of the focus layer appear as “rings” in the movie ( see T0s in Fig . 1C ) . Hence , the “landing” stage was defined as the time between the first detection of a “ring” ( followed by a continuous downward movement ) and the change of the direction and speed typically observed when NSS started ( T0–0 . 4 s in Fig . 1C ) . In analogy , the “take off”-stage describes the time between the end of stopping or NSS and the disappearance of the “ring” ( T41 . 9–42 . 3 s in Fig . 1C ) . Additionally , we tracked the time spent stopped or engaged in NSS . Take-off and landing occurred within <3 . 1 and <4 . 2 s , respectively ( median 0 . 4 s for both; Fig . 1D ) , while the time engaged in NSS was significantly longer ( median 1 . 5 s; range 0 . 3–41 . 5 s; Fig . 1D ) . Overall , the bacteria covered significant distances swimming along the host cellular surface ( 6 . 7–325 µm; see also below ) . Furthermore , we observed a variety of “behaviors” with respect to stopping . 33% of all imaged S . TmΔ4 ( pGFP ) bacteria did never stop , while 67% made one or more stops ( “NSS and stop”; Fig . 1E ) . Some bacteria stopped up to 5 times on the cell surface and some S . TmΔ4remained stopped ( or docked ) until the end of the imaging experiment , with the longest observed stop lasting 280 s ( Fig . 1F ) . Overall , the bacterial density at the cellular surface was significantly increased compared to the overlaying media ( Fig . 1G ) . Accumulation of bacteria at the cellular surfaces was not attributable to gravity , as indicated by a comparison between motile ( affected by gravity and NSS ) and non-motile bacteria ( affected only by gravity; suppl . Fig . S1 ) . Much rather , flagella-driven NSS and ensuing stops accounted for the increased local S . Typhimurium density at the host cell surface . This suggested that the prolonged contact time might contribute to the target site selection by S . Typhimurium . In order to determine the role of bacterial adhesins in the initial phases of the surface interaction , we analyzed two S . Typhimurium mutants , S . Tm-T1 ( SL1344 invG ) and S . Tm-T1-Fi ( SL1344 invG fimD; Tab . 1 ) . These mutants lack one or two surface structures , namely type 1 fimbriae and the TTSS-1 apparatus , which are known to mediate reversible binding and irreversible docking of S . Typhimurium to HeLa and other host cells [8] , [9] , [30] . However , a role of these adhesins for near surface swimming or stopping had not been addressed . HeLa cells were infected with S . Tm-T1 ( pGFP ) and S . Tm-T1-Fi ( pGFP ) and transient interactions were monitored by time lapse fluorescence microscopy as in Fig . 1 . All analyzed parameters , including the number of stops , were indistinguishable from those of S . TmΔ4 ( Fig . 2A–D; compare to Fig . 1 ) . These results indicated that none of the initial surface interactions were affected by TTSS-1or by fimD ( Fig . 2A , C ) . Strikingly , this also pertained to the transient stops and clearly distinguishes the initial surface interactions from later stages of the infection , i . e . reversible binding and docking . Stopping thus seems to be attributable to a different mechanism . Nevertheless , as indicated by earlier data [8] , [9] , [30] , some of the stops must result in reversible binding and docking . To estimate the relative frequency of stopping and of the docking events , the total number of stops observed on the cell layer ( = the number of potential docking events ) , during a 5 minute period was calculated: the total number of S . TmΔ4bacteria landing on the cell surface during a 5 min period was multiplied by the number of stops using the values determined in as Fig . 1E . After the end of the5-minute real time imaging experiment , the cells were washed , fixed and stained for DNA ( DAPI ) , actin ( TRITC-phalloidin ) and S . Typhimurium ( anti-LPS antibody ) . This protocol removed all “stopped” bacteria , while “reversibly bound” and “docked” bacteria remained on the cells and were enumerated by fluorescence microscopy [8] , [9] , [30] . Comparing the results from both types of analysis revealed that no more than 1–2% of the total stops ( as detected by time lapse microscopy ) resulted in a docking event . Hence , transient stops are approx . 50- to 100-fold more frequent than docking events , at least in the 5-minuteinfection experiments that we have performed , here . In conclusion , these data established that NSS requires neither TTSS-1nor type I fimbriae . The initial pathogen host cell interaction is thus clearly distinct from subsequent stages , i . e . reversible binding and docking . Bacterial flagellar movement is guided by chemotaxis . It had remained unclear , if NSS was dependent on chemotaxis or whether non-directed motility would suffice . To address this issue , HeLa cells were infected with S . TmΔ4 cheY ( pGFP ) ( SL1344 ΔsipAsopBEE2 cheY; Table 1 ) . This isogenic mutant is a “straight-swimmer” , expresses wild type numbers of functional flagella , but cannot swim along chemical gradients . The initial surface interactions were monitored by time lapse microscopy as described in Figs . 1 and 2A–D . All analyzed parameters were indistinguishable from those ofS . TmΔ4 , S . Tm-T1 or S . Tm-T1-Fi ( compare Fig . 2E , F to Fig . 1D , E and Fig . 2A–D ) . Therefore , in our tissue culture assays non-directed motility was sufficient for facilitating the initial surface interactions . Our data suggested that basic physical principles may explain the initial interactions of S . Typhimurium with host cellular surfaces . In this case , the movement patters on cellular and acellular surfaces should be quite similar . Therefore , we inoculated glass-bottom tissue culture dishes seeded with HeLa cells ( or not ) with S . TmΔ4 ( pGFP ) . The bacterial surface interactions were analyzed by time-lapse fluorescence microscopy as described in Figs . 1 and 2 . S . TmΔ4 ( pGFP ) performed NSS with equivalent speed on glass and on cellular surfaces ( approx . 30 µm/s; Fig . 3A ) . However , the median duration of an episode of NSS ( 6 . 3 s on glass vs . 2 . 95 s on cells; Fig . 3B ) and the median distance travelled during this time ( 221 µm on glass vs . 95 µm on cells; Fig . 3C ) were slightly larger on glass than on the cellular surface . These observations were in line with our hypothesis that the initial surface interactions may be governed by equivalent physical principles . Strikingly , the bacteria followed “right-handed” curved tracks on glass and , to a lesser extent , on cellular surfaces ( Fig . 3D , E ) . It is thought that this curvature of the NSS tracks is attributable to the shear force between the flagella-mediated rotation of the bacterial body and the respective surface [28] , [33] . So far , our data were in line with the hypothesis that general physical principles are responsible for NSS on cellular and on glass surfaces . In this case , other types of host cells or motile bacteria should yield similar results . Therefore , we have extended our analysis to MDCK cells , a commonly used polarized epithelial cell line , and E . coliNissle which was transformed with the GFP expression plasmid ( E . coliNissle ( pGFP ) ; Suppl . Fig . S2 ) . Both , E . coliNissle ( pGFP ) and S . TmΔ4 ( pGFP ) engaged in NSS and displayed similar movement patterns on MDCK , on HeLa cells and on glass surfaces ( Fig . 3D , E and suppl . Fig . S2A–D ) . Nevertheless , the shape of the NSS-tracks may differ slightly . This might be attributable to different morphological features displayed by the different types of surface and represents an interesting topic for future research . In conclusion , these data lend further support to the notion that initial surface interactions are governed by general physical principles and suggest that near surface swimming might be a general strategy for target selection employed by different flagellated bacteria . Analysis of our movies so far suggested that during NSS , rounded cells ( e . g . dividing cells ) represent preferential sites for stopping . We hypothesized that bacterial stopping can be explained by the prominent topological features of these rounded cells . If so , bacteria should also stop at artificial topological obstacles and bacteria should accumulate at such sites . This was tested in two different ways , i . e . in a simplified experimental setup and by a computer simulation ( see below ) . In order to experimentally test our hypothesis , we have analyzed the initial surface interactions of bacteria with small glass beads which were placed as artificial obstacles onto a glass surface . Glass-bottom tissue-culture dishes harboring glass beads ( Ø = 500 µm ) were therefore inoculated with S . Tmwt ( mCherry ) and bacterial movement patterns were analyzed by time-lapse fluorescence microscopy , as described in Figs . 1 , 2 and 3 . Again , the bacteria were moving for long distances along the glass surface , but were stopping in the immediate vicinity of the glass beads ( Fig . 4A , B; suppl . Video S3 ) . 45–70% of these bacteria continued to swim or took off again before the end of the movie , indicating that these were truly stopping ( not binding/docking ) . Thus , topological obstacles can facilitate site-specific stopping of bacteria engaged in NSS . In addition , these data allowed a rough estimation of the “altitude” at which S . Typhimurium swims above the glass surface . Based on the geometry of the glass bead , the center of the bacterial cell was 0 . 43+/−0 . 07 µm away from the surface . Similar results were obtained for beads with smaller diameters ( 150 and 30 µm , respectively ) . The rod-shaped S . Typhimurium cell has a radius of approx . 0 . 5 µm . Therefore , most of the “distance” is attributable to the bacterial cell and the NSS “altitude” is most likely <150 nm above the glass surface . These observations were in line with earlier work [27] , [28] , [29] , [34] describing bacterial NSS . The data presented above indicated that NSS affects target cell selection in two ways: by increasing the local pathogen density on surfaces and by enhancing the probability of surface contacts ( stopping ) at topological obstacles projecting from this surface . A computer simulation was used to verify whether these two phenomena are sufficient for explaining target site selection . We modeled the interaction of S . Typhimurium with a three dimensional landscape consisting of a flat surface and one spherical obstacle , partially submerged into the surface ( Materials and Methods ) . The particles ( motile , but non-chemotactic “bacteria” ) were introduced and moved linearly within the 3D virtual space above the surface . Upon contact with the sphere , the particles were either “stopping” ( 10% likelihood ) or reflected ( 90% likelihood ) . Three different scenarios were analyzed with respect to the particles encountering the flat surface . ( 1 ) the “random” scenario: particles were reflected back into 3D space and randomly assigned a new direction of movement . ( 2 ) the “billiard” scenario: particles were reflected with an angle of reflection identical to the angle of impact . ( 3 ) the “NSS” scenario: particles encountering the surface did not leave but followed the surface via NSS . If the sphere was encountered during NSS , the particles stopped at this site with a likelihood of 10% . In the NSS scenario ( scenario 3 ) but not the random or the billiard scenario , particle accumulation occurred right at the topological obstacle ( Fig . 5 ) . Therefore , NSS and stopping at physical obstacles are sufficient for explaining the target site selection observed in simplified model systems ( Fig . 4A , B ) and tissue culture infection experiments ( Fig . 1 , [31] , [32] ) . So far , our analyses of the target site preference of S . Typhimurium had focused on the initial surface interactions , i . e . landing , take-off , NSS and stopping . Next , it was important to establish how these initial interactions may affect the subsequent steps of the infection process , i . e . docking . If stopping increases the probability of binding and irreversible docking at the respective site , stopping ( as observed by time lapse microscopy; Fig . 1 , 2 ) and docking should display an equivalent target site preference . To assess the target site preference of binding/docking on host cell layers quantitatively , we have employed a well-established “standard” infection protocol [9] , [32] . HeLa cells were infected for 6 min with S . TmΔ4 at the indicated m . o . i . , washed gently , fixed and stained ( see Materials and Methods ) . As binding/docking occurred approximately 50-fold less frequently than stopping ( see above ) , we employed higher multiplicities of infection than in the time lapse microscopy experiments . Visual inspection indicated that S . Typhimurium has a pronounced targeting preference for rounded cells ( Fig . 6A , top panel; “mitotic” nuclei with condensed DNA highlighted in yellow ) . In particular , the bacteria were found to dock to the base of rounded cells ( suppl . Fig . S3A ) . In order to quantify this phenotype , we employed automated fluorescence microscopy , and an automated image-analysis routine ( Materials and Methods ) . Rounded cells were targeted ( docked ) with significantly higher efficiency than non-dividing cells ( Fig . 6C , left panel ) . This targeting preference for rounded cells was also observed by manual quantification ( S . TmΔ4; suppl . Fig . S3B ) and in infection experiments with S . Tmwt ( data not shown ) . These findings were quite similar to our observations during the initial surface interactions ( Figs . 1 and 2 ) . Rounded mitotic cells seem to represent topological obstacles within the cellular landscape . In fact , the rounded mitotic cells were significantly “higher” than the interphase cells ( 12±2 µm vs . 5 . 4±1 . 1 µm; n = 20 each; p<0 . 0001 , Mann-Whitney-U test ) . Taken together , NSS-driven stopping and binding/docking displayed equivalent preferences for topological obstacles like rounded cells . Next , we have addressed the role of flagellar-driven motility . It is well established that flagella are required for cellular invasion of S . Typhimurium [15] , [16] , [17] , [18] , [19] . The data presented above suggested that flagella-driven NSS might determine the target site preference . In order to study the role of bacterial motility in the targeting preference of binding/docking , we analyzed the host-cell interaction patterns of the non-motile mutant S . TmΔ4 fliGHI ( SL1344 fliGHI ) which does not express flagella . This mutant and other non-motile S . Typhimurium mutants did not dock efficiently at all ( Fig . S4 ) . Importantly , the few HeLa cells that were infected harbored just one bacterium , even at high m . o . i . ( Fig . 6A , bottom panel; Fig . 6C , right panel ) . In line with our hypothesis , S . TmΔ4 fliGHI did not display a targeting preference for rounded cells ( Fig . 6A , C ) . In order to increase the chances of a host cell encounter by S . TmΔ4 fliGHI , equivalent infection experiments were performed applying mild centrifugal force ( 500 g for 10 min ) to increase the collision rate between non-motile S . TmΔ4 fliGHI and the host cells ( Materials and Methods ) . This strategy is commonly used for “rescuing” invasion defects of non-motile S . Typhimurium mutants , ( e . g . [16] , [35] ) . As expected , centrifugation increased the number of binding/dockingS . TmΔ4 fliGHI in our assay ( Fig . 6B , data not shown ) . Nevertheless , S . TmΔ4 fliGHI did not display any target preference for the rounded ( mitotic ) cells , even at the highest m . o . i . , tested ( Fig . 6D; right panel ) . This was in line with our hypothesis that flagella-driven NSS determines not only the targeting preference of “stopping” , but also that of binding/docking . Finally , we analyzed the role of chemotaxis , i . e . directed motility along chemotactic gradients . S . TmΔ4 cheY is an isogenic and motile mutant incapable of chemotaxis ( compare Fig . 2E , F ) . This mutant yielded equivalent results as S . TmΔ4including a highbinding/docking efficiency and a pronounced targeting preference for rounded cells ( data not shown ) . Taken together , these results suggested that preferential infection of rounded cells is attributable to simple non-directed motility and that chemotaxis was not required , at least in this simple tissue culture model . During NSS , stopping and docking occurred preferentially at rounded cells . If targeting was indeed dictated by topological obstacles , manipulation of the host cell morphology should affect both , the initial surface interactions and docking by S . Typhimurium . To test this hypothesis , we manipulated the host cellular morphology with cytochalasin D ( Cyt . D ) . This drug depolymerizes actin filaments causing the cells to round up . First we analyzed the effect of cytochalasin D treatment on binding/docking . HeLa cells , pretreated with the indicated concentration of cytochalasin D for one hour , were infected for 12 minutes with S . TmΔ4 ( pGFP ) , S . Tm-T1 ( pGFP ) or S . Tm-T1-Fi ( pGFP ) at the indicated m . o . i . . Afterwards , cells were fixed and stained and we analyzed the actin-based cytoskeleton and bacterial docking ( Materials and Methods ) . Since the automated evaluation now focused on the percentage of infected cells ( not the number of individual bacteria per infected cell ) , overlapping but somewhat higher m . o . i . s were used than in Fig . 6 . As shown in Fig . 7A , HeLa cells were partially rounded at 2 µM and fully rounded at 10 µM cytochalasin D . All three S . Tm strains bound/docked to the rounded cells with an increased efficiency ( Fig . 7B ) . Equivalent results were obtained with another actin-disrupting drug ( latrunculin B; suppl . Fig . S5 ) . This verified that the host cell topology has a profound effect on the target site preference of binding/docking by S . Typhimurium . Next , we analyzed the effects of the cytochalasin D treatment on the initial surface interactions . HeLa cells pretreated with 10 µm cytochalasin D were infected with S . TmΔ4 ( pGFP ) , S . Tm-T1 ( pGFP ) or S . Tm-T1-Fi ( pGFP ) and landing , NSS , stopping and take-off were analyzed by time-lapse microscopy as described in Fig . 1 and Fig . 2 . Strikingly , on the cytochalasin D-treated cells , nearly all bacteria engaging in NSS made at least one stop . Only very few bacteria displayed uninterrupted NSS ( “NSS only”; Fig . 8A , D , G ) . This was quite different from our data on untreated HeLa cells where 33% of all bacteria displayed uninterrupted NSS ( see Fig . 1E ) . Furthermore , on cytochalasin D-treated cells , the time spent at each stop was significantly longer and some of these bacteria “stopped” for the entire course of the 5-min experiment ( Fig . 8B , E , H ) . In a similar analysis , the fraction of time thatS . TmΔ4 ( pGFP ) , S . Tm-T1 ( pGFP ) or S . Tm-T1-Fi ( pGFP ) spent stopping at the host-cell surface was significantly higher than in untreated HeLa cells . In fact , on the cytochalasin D-treated cells , a majority of the bacteria spent most of their time “stopping” ( “only stoppers”; Fig . 8C , F , I ) . Again , no differences were observed between S . TmΔ4 ( pGFP ) , S . Tm-T1 ( pGFP ) and S . Tm-T1-Fi ( pGFP ) , confirming that “classical” adhesins do not significantly affect these transient initial bacteria-host cell interactions . In conclusion , the shape of host cells has a profound effect on initial surface interactions and binding/docking . This provides further evidence that NSS leads to a target site preference for physical obstacles . Membrane ruffles triggered by the TTSS-1 virulence system represent a well-known topological obstacle encountered on infected cell layers [36] , [37] , [38] , [39] . We hypothesized that membrane ruffles might enhance local stopping , binding and docking at pre-existing ruffles , thus leading to cooperative invasion . To study the targeting of ruffles , we have focused on binding and docking . These two steps of the infection process can be analyzed using “standard” fluorescence microscopy assays [9] , [32] . First , we employed a co-infection strategy using a “helper strain” and a “reporter strain” . S . TmSopE ( without plasmid ) was chosen as the helper strain , since its effector SopE is able to trigger pronounced membrane ruffling [5] , [32] . S . TmSopE carries deletions of the effectors sipA , sopB and sopE2 , thus eliminating confounding pleiotropic actions of these effectors on the host cell . SopE-induced ruffles appeared within 5 minutes , are known to have a prominent shape , and represent a large physical obstacle on an otherwise much flatter cellular surface ( Fig . 9A; [40] ) . S . TmΔ4 ( pGFP ) , which does not trigger ruffles itself , was used as a “reporter” to assess docking to the ruffles . In a time lapse microscopy experiment employing DIC and fluorescence imaging both strains engaged in NSS and stopped frequently at ruffles ( Fig . 9B; supplementary VideoS4; data not shown ) . To analyze a preference of S . Typhimurium for cellular ruffles in a quantitative manner , HeLa cells were infected with a 1∶1 mixture of S . TmSopE andS . TmΔ4 ( pGFP ) at a high or a lower m . o . i for 6 min . In control experiments ( no ruffles ) , HeLa cells were infected with a mixture of S . TmΔ4 and S . TmΔ4 ( pGFP ) . Subsequently , cells were fixed followed by staining of extracellular bacteria ( using an anti-Salmonella antibody; Materials and Methods ) . After permeabilization of the cell membrane , the actin cytoskeleton was stained . The data evaluation strategy is depicted in Fig . 9C . In the experiments using S . TmSopE as the helper strain ( ruffling occurs ) , we determined the number ofS . TmΔ4 ( pGFP ) ( Fig . 9D , red bars ) and S . TmSopE ( Fig . 9E , red bars ) residing on the respective ruffle . In the negative controls ( no ruffling; co-infection with S . TmΔ4 and S . TmΔ4 ( pGFP ) ) we quantified all bacteria located on the respective cell ( Fig . 9D , D grey bars ) . Comparing targeting to a ruffle as opposed to targeting to a whole cell ( control w/o ruffle ) is a conservative strategy for detecting ruffle-specific target site preferences , since the whole cell has a much larger area than an individual ruffle . Furthermore , it should be noted that the anti-Salmonella antibody was applied before permeabilization ( Materials and Methods ) . This allowed us to discern internalized and external reporter bacteria . In all experiments , S . TmSopE and S . TmΔ4 ( pGFP ) docked more efficiently to ruffles than to non-ruffling cells ( Fig . 9D , E compare red and grey bars ) . This was true for the reporter strain ( Fig . 9D; S . TmΔ4 ( pGFP ) ) as well as for the helper strain ( Fig . 9E; i . e . S . TmSopE ) . A similar effect was also observed in 12-min infection experiments ( data not shown ) . Therefore , ruffling stimulated docking of bacteria to cellular ruffles . Moreover , ruffling facilitated internalization of S . TmΔ4 ( pGFP ) . Internalized S . TmΔ4 ( pGFP ) was detected in all co-infections with S . TmSopE ( 1 . 4 and 1 . 6 bacteria per ruffle on average for the experiment in Fig . 9D for m . o . i . s of 62 and 250 , respectively ) . No internalizedS . TmΔ4 ( pGFP ) was detected in the negative controls ( no ruffles; S . TmΔ4 as helper strain ) . Equivalent data were obtained in an analogous experiment using automated image acquisition and analysis methods ( suppl . Fig . S6 ) . In another control we tested immotile bacteria ( S . TmΔ4 fliGHI ( pGFP ) ) . As expected , these bacteria did not attach to ruffles or normal cells ( Fig . S7 ) . Again , cellular attachment was rescued upon centrifugation; however , this did not lead to any targeting preference for ruffles vs . non-ruffling cells ( Fig . S7 ) . Finally , we explored the targeting to ruffles by E . coli Nissle , a non-invasive motile bacterium . In co-infection experiments with S . Tmwt , E . coli Nissle displayed a targeting preference for host cellular ruffles ( Fig . S8 ) . Taken together , the effector protein SopE is required for triggering ruffles . Once the ruffles are formed , they seem to represent prominent physical obstacles which facilitate stopping , binding , docking and internalization of motile pathogens . However , please note that we cannot exclude that other factors , besides ruffle topology ( e . g . altered membrane structure or composition in the ruffle ) might also contribute to S . Typhimurium targeting to these sites . Finally , we reasoned that S . Typhimurium recruitment onto ruffles might lead to cooperative invasion . If multiple bacteria dock to the same ruffle , more effector proteins are delivered , thus increasing the ruffle size and enhancing the chance for stopping , binding and docking of additional bacteria . To test this hypothesis , HeLa cells were infected for 9 min with S . TmSopE ( pGFP ) at increasing multiplicities of infection , fixed and stained . We first quantified the fraction of ruffling cells ( Fig . 10A , B , left panel ) . Next we determined ( at each m . o . i . ) the number of intracellular and extracellular bacteria residing in an individual ruffle ( Fig . 10A , B , middle panel ) . Finally , we determined the number of “invaded S . TmSopE” per cell ( Fig . 10 A , B , right panel ) . This was achieved by multiplying the percentage of ruffling cells with the number of intracellular bacteria per ruffle . All these parameters increased as a function of the m . o . i . Next we wanted to determine the invasion efficiency mediated by a single bacterium without “support” from other bacteria . We therefore focused at low m . o . is . , where typically one bacterium was observed per ruffle . The few ruffles with two associated bacteria were excluded from the subsequent analysis . The single bacterium was either located outside or within the host cell ( Fig . 10B , m . o . i . 4 and 8 ) . To estimate invasion efficiency by this single bacterium alone without possible support by other bacteria , this invasion efficiency was extrapolated to higher m . o . i . , assuming a linear increase of invasion efficiency with increasing m . o . i . ( Fig . 10B , right panel , black line ) . Strikingly , this extrapolated invasion efficiency was much lower than the observed invasion efficiency at higher m . o . i . ( Fig . 10B , right panel , red line ) . This indicated that S . Typhimurium invasion occurred in a cooperative fashion , most likely by provoking stopping , binding and docking of additional bacteria engaged in NSS at sites of membrane ruffling .
The mechanism of target site selection by S . Typhimurium had remained enigmatic . We have analyzed the initial surface interactions between the pathogen and the host by time-lapse microscopy , by comparative analysis of pathogen movements on cells and glass surfaces as well as by standard infection experiments . Upon encounter with a host-cell layer , we identified a distinct early phase of the infection characterized by landing , near surface swimming and stopping , which preceded later events such as fim- or TTSS-1-mediated binding and docking . In this initial phase , flagellar motility has at least two functions in establishing the contact with the host cell , i . e . propelling the pathogen towards the host cell layer and facilitating NSS . NSS is likely attributable to physical forces emanating from the pathogen movement along the surface . Here we found that NSS increased the local pathogen concentration on the host-cell surface , and lead to stopping at topological obstacles . NSS thereby mediates preferential docking at the base of rounded cells and on pre-existing membrane ruffles . Thus , the target-cell selection for dividing cells and cooperative infection at membrane ruffles can be explained simply by physical forces between the host cellular surface and the motile pathogen , which govern the initial phase of the pathogen-host cell interaction . Different models have been proposed to explain NSS [27] , [28] , [41] , [42] , [43] , [44] . Recent approaches have combined hydrodynamic entrapment and DLVO interactions with Brownian motion , which leads to random changes of the bacterial NSS-altitude [37] . Random variations in surface distance also lead to predictable changes of the angle of bacteria towards the surface and the radius of the curved track . Therefore , Brownian motion can enable probing of the surface following different curvatures at different heights when the bacteria are swimming along . This variability may help to further expand the surface probing capacity of a motile pathogen [37] . However , it should be noted that the Brownian motion-forces are much weaker than the thrust provided by the flagella . This may explain why stopping and docking on open glass surface areas is much less frequent than at the base of elevated obstacles located on the surface . Only at these obstacles , the force provided by the flagella is fully employed to counteract repulsive forces thus driving the bacterium as close to the surface as possible . This may increase the chances for the formation of stable contacts as required for binding and docking . Are “stopping” and “docking” related ? In both cases , S . Typhimurium stays at one particular spot on the host cell surface for at least some time . However , the retention mechanisms seem to differ . Docking and binding , i . e . long-term association with host cells are mediated by adhesins [45] , [46] . In the case of HeLa cells , docking is mediated mainly viaTTSS-1 and type-1 fimbriae which is why the mutant strainS . Tm-T1-Fi has a reduced docking efficiency [9] . In contrast , stopping was not significantly affected in the case ofS . Tm-T1-Fi , implying that neither TTSS-1 nor type-1 fimbriae mediate stopping ( Fig . 2 ) . Furthermore , washing ( as performed in docking experiments; see Materials and Methods ) , removed ≥98% of all “stopped” bacteria from the host cell surface . Presumably , the remaining ∼2% were docked , suggesting that stopping is mediated by a weaker force and that it is approximately 50-fold more frequent than docking , at least during the first minutes of infection . In spite of these differences , both stopping and docking required motility ( but not chemotaxis ) and occurred with high probability at rounded cells and membrane ruffles . Based on these considerations , we propose that stopping at topological obstacles may simply extend the residence time at a given location ( close to the surface ) , thereby increasing the probability of adhesin-mediated binding and docking . It is tempting to speculate that Brownian motion might randomly drive stopping bacteria into even closer proximity of the host cell surface , thus increasing the chances for a successful engagement of TTSS-1 or type-1 fimbriae [34] . In this way , prolonged stopping and Brownian motion would foster the preferential infection of rounded cells and cell ruffles . Thus , landing , NSS and stopping may allow prolonged probing ata very limited area of the 2D surface . If docking is unsuccessful , S . Typhimurium continues NSS or takes off into and may engage in initial surface interactions at another site . NSS recruits S . Typhimurium into membrane ruffles , thus promoting cooperative invasion . Invasion of more than one bacterium at a single ruffle has been described before [47] , [48] , [49] but no mechanistic explanation and no evidence of cooperativity has been provided . Our data show that ruffles represent topological obstacles favoring stopping and productive invasion . Therefore , ruffles might be regarded as a site of “communication” between individual bacteria . Firstly , this implicates the exploitation of invasive strains by non-invasive strains , which can subsequently invade . This type of “rescue” has been described previously [37] , [50] . Secondly , it implicates cooperation between invasive strains , in two respects . When S . Typhimurium makes a favorable docking interaction and induces membrane ruffling in the host cell , this increases the chance that additional bacteria can “find” this invasion-permissive site . Furthermore , in some hosts or cell types , higher dosages of effector proteins might be required for triggering successful invasion . In this case , recruitment into ruffles might allow larger amplitudes of stimulation . Certainly , this would be of importance for the mechanistic interpretation of results from tissue-culture cell-infection experiments . In animal experiments , flagella and motility are also required for efficient gut infection [21] , [22] , [24] , [51] . Here , the flagella serve additional functions not observed in tissue culture . In the inflamed gut , flagella facilitate chemotactic movement thus mediating access to the nutrient-rich molecules secreted by the gut wall [24] ) . Thereby , chemotactic motility propels the pathogen towards the gut surface . For this reason , both , chemotaxis and motility are required in the gut [24] . This additional function of flagella , which does not affect the tissue-culture infection , has precluded straightforward animal-infection experiments addressing the “within host” importance of the NSS-based , chemotaxis-independent targeting mechanism described here . Does NSSor stopping also occur at the gut surface ? In this study we demonstrated NSS in vitro at the surface of HeLa cells and polarized epithelial cells . However , in previous studies , S . Typhimurium was found to accumulate at the surface of the gut epithelium during the first phase of the infection [24] . Furthermore , S . Typhimurium swimming along the epithelial surface has also been visualized by in vivo live microscopy in the cecum of infected mice [6] . At 4 h post infection , the pathogen was found to swim along the surface of infected crypts at a speed of 5–50 µm/sec . Subsequently bacteria stopped at the epithelium and entered into enterocytes . While these experiments did not elaborate on the individual phases of the initial surface interactions , they clearly demonstrate that NSS occurs within the host's intestine . Later on , upon triggering mucosal inflammation , S . Typhimurium seems to invade into neutrophils which are transmigrating in large numbers into the gut lumen [52] . Intriguingly , neutrophil infection was found to require motility . It is tempting to speculate that neutrophils are infected by S . Typhimurium during transmigration . While crossing the epithelial barrier , the luminal part of the neutrophil may form a physical obstacle stopping S . Typhimurium swimming along the epithelial surface . These observations suggest that NSS-mediated targeting of host cells may occur in vivo . A detailed analysis of these processes during the course of a real infection will be an interesting topic for future research . The NSS-mediated targeting mechanism identified in our work is based on general physical forces which act on any particle moving on a 2D surface . Therefore , it should pertain to many other motile bacteria , including pathogens ( e . g . enteropathogenic E . coli , Yersinia spp . ) , commensals , as well as environmental bacteria . Accordingly , we found that the motile strain E . coli Nissle can engage in NSS and target Salmonella-induced membrane ruffles . Our results imply that the NSS-mediated targeting should result in preferential binding to physical obstacles present on the respective surface , including abiotic obstacles as well as prominent topological features of surface-exposed cells of a given host . Thus , the initiation of biofilm formation and the infection of animals at particular sites might be governed by the same basic principles . Deciphering these principles will be of great interest for basic microbiology and might allow the development of countermeasures prohibiting the initial steps of bacterial colonization .
Published bacterial strains are listed in Table 1 . The construction of additional strains is described , below . Cells were seeded into glass-bottomed culture dishes ( Mat Tek ) in DMEM ( Invitrogen ) , 10% FCS ( Omnilab ) containing streptomycin ( 50 µg/ml ) , 24 hours prior to the experiment at 300 000 cells ( HeLa Kyoto , if not mentioned otherwise ) per 35 mm well; experiments were performed in HBSS ( Invitrogen ) , 10% FCS , 20 mM Hepes ( Invitrogen ) , pH 7 . 2–7 . 5 ( Invitrogen ) . After exchange of media to HBSS , cells were incubated at 37°C , 5% CO2 and infected with the indicated S . Typhimurium strain ( derivative of SL1344 , [53] ) carrying plasmid pCJLA-GFP at an estimated m . o . i . of 1 . 5 . If not stated otherwise , movies were acquired on a Leica DMI-6000B microscope , either using the differential interference contrast mode or the fluorescence mode . DIC-movies were acquired using a 63× oil objective ( HCX PLAN Apochromat from Zeiss , NA 1 . 4 ) ; for fluorescence imaging , a 20×-objective ( HC PLAN Apochromat from Zeiss , NA 0 . 75 ) with a 2-fold optovar was used . If not indicated otherwise , movies were acquired at 10 frames per second for 5 minutes . Movies were analyzed using the program Volocity ( Improvision , UK ) and the manual tracking mode . For Fig . 4 , we used a Zeiss Axio Observer equipped with a spinning disc confocal head and a 100× objective ( oil NA1 . 4 , ) . For suppl . Fig . S1 , we used a Zeiss Axiovert 200 m inverted microscope equipped with an Ultraview confocal head ( Perkin Elmer ) and a krypton-argon laser ( 643-RYP-A01 Melles Griot , The Netherlands ) and a 20× objective , 0 . 75NA ( Optovar 1 . 6 ) . In these experiments estimating the influence of gravity , glass-bottom culture dishes containing HBSS but no cells were inoculated with bacteria as described for the other time lapse experiments; For testing of bacterial near surface swimming on MDCK cells , cells were seeded in glass-bottom culture dishes in DMEM , supplemented with 10% FCS , grown to confluence and let polarize for five days . Infection with S . TmΔ4 was done exactly as described for HeLa cells . Movies were acquired on a Leica DMI-6000B microscope using a 20×-objective with a 1 . 6 optovar . The movie was acquired for 5 min at 10 frames per second and an overlay of the movie is shown . S . Typhimurium strains were grown before infection as described [32] . E . coli Nissle was grown identically except that normal Luria-Bertani media was used . For quantification of the contact of S . Typhimurium with the cells , the interaction was divided into 4 phases: 1 ) landing: the time of appearance of a S . Typhimurium until interruption of the continuous downward movement , indicated by changes in focus , direction and speed . 2 ) stopping: episodes without changes in any direction for at least 3 frames . 3 ) take off: the continuous upward movement until disappearance . 4 ) NSS: a continuous movement in the xy-direction other than 1 ) or 3 ) . S . Typhimurium that left the field of view during quantification , as well as a few cases with ambiguities were excluded . For quantification of the number of S . TmΔ4 ( pGFP ) within the focus depth , either at the cell surface or >100 um above the surface ( “in solution” ) , 40 time points from 2 independent experiments were analyzed . The same time points were used for each movie analyzed and they were spaced throughout the duration of the movie . For each bacterium within focus depth at the specific time point , the entire contact time spent on the surface was analyzed . Each bacterium was thus either counted as engaged in “NSS” or “NSS and stopping” . For the quantification of bacteria swimming in solution , any non-motile bacteria were excluded . For quantification of the distance , time and speed of S . TmΔ4 ( pGFP ) on the glass and cell surface , 20 bacteria from 2 independent experiments were manually tracked from their first contact with the cell to their last contact with the cell ( i . e . NSS quantification excluding the landing and take-off stages ) or until they left the field of view . S . TmΔ4 were selected at random from those arriving in the center of the field of view . The time indicated is NSS-time only and does not include any intermittent time spent stopping at the surface . For high-resolution images , HeLa cells were seeded on glass cover slips for 24 hours and infected with S . Typhimurium carrying plasmid pM965 for constitutive gfp expression at the indicated m . o . i . . After fixation , S . Typhimurium were stained by indirect immunofluorescence using an anti-LPS antibody ( Difco ) and goat anti-rabbit-Cy5 ( Jackson ) as a secondary antibody; the actin cytoskeleton was stained by tetra methyl rhodamine isothiocyanate ( TRITC ) -phalloidin after 5′ permeabilization with 0 . 1% Triton Tx100 . Images were taken on a Zeiss Axiovert 200 m inverted microscope equipped with an Ultraview confocal head ( PerkinElmer ) and a krypton argon laser ( 643-RYP-A01 , Melles Griot ) using a 100× oil immersion objective ( PLAN-Apochromat Zeiss with an NA of 1 . 4 ) . Stacks of 0 . 2 µm were acquired; deconvolution was performed in the actin channel with the program Volocity and a calculated point spread function . 3D-reconstruction and reconstruction of zx-layers was done using Volocity . Docking experiments were done as described [9] . In brief: cells were seeded in 96-well micro-clear plates , ( half size , Greiner ) , at 6000 cells per well 24 hours prior to the experiment and infected with the indicated S . Typhimurium strain followed by fixation , staining of nuclei and bacteria using DAPI and an anti-Salmonella antibody ( Difco ) , respectively . Images were acquired on a MD-Image Xpress microscope ( Molecular Devices ) using a 4×-objective in the DAPI- and Salmonella-channel . Images were analyzed using the open source program CellProfiler [58] and customized Matlab-scripts , available upon request . For automated analysis of individual docked bacteria ( Fig . 1 ) the analysis consisted of three major steps . First , cells and nuclei were detected using the program CellProfiler [58] and cellular properties , such as morphology , texture and intensity were extracted . Second , single bacteria were identified using the “a trous” wavelet transform [59] and every bacterium was assigned to the corresponding host cell . In a third step , mitotic cells were distinguished from interphase cells using supervised machine learning technique . For this purpose , the “Advanced Cell Classifier program” was applied and the classification was performed using the extracted cellular features with an artificial neural network classifier [60] . Finally the average amount of bacteria on mitotic and interphase cells was calculated . Docking onto mitotic cells: HeLa cells were seeded in 96-well Microclear plates ( full size , Greiner ) at 6000 cells per well 24 hours prior to infection . After infection and fixation , extracellular S . Typhimurium was stained using an anti-Salmonella LPS antibody and a Cy5-labelledsecondary antibody ( Fig . 9 ) . Due to the bright staining of extracellular bacteria , the unique bacterial shape and the high resolution used for quantification bacterial staining was clearly distinguishable from actin staining . For illustrative purposes ( Fig . 9C ) , in one experiment bacteria were stained using a Cy3-labelled secondary antibody . Subsequently , nuclei and actin were stained using DAPI and TRITC-phalloidin , respectively . Finally , bound S . Typhimurium bacteria were manually quantified using a 40×-objective ( EC Plan-Neofluar objective with a NA of 1 . 3 ) . Docking onto ruffles: The gfp-labeled S . Typhimurium strain and the helper strain were mixed at equal dilutions prior to infection . Staining and fixation was done as described for mitotic cells . Testing of cooperativity: S . TmSopE-carrying plasmid pM965 was incubated at the indicated concentration with HeLa cells; fixation and staining was done as described . To calculate invasion efficiency of S . Typhimurium without cooperativity , invasion was calculated at the lowest m . o . i . , excluding ruffles with more than one associated S . Typhimurium . At higher m . o . i . , this number was assumed to increase proportionally with the number of added S . Typhimurium . In our calculations we modeled a three-dimensional environment within a cubic space . A round sphere was placed into this space; the radius of the sphere was chosen to be 1/10 of the length of the surrounding cube . The sphere was partially submerged into the bottom surface of the cube and 75% of the height of the sphere remained within the cube . In addition , 100 single particles were modeled into the cubic space . For simplicity the size of the particles was assumed to be zero in all dimensions . The whole volume of the cube was assumed to be accessible to the particles except the interior of the sphere . All particles were assigned the same constant speed but a randomly chosen vector of movement . The positions of the particles where integrated after travelling a fraction of 1 . 2*10−5 of the length of the square . Upon hitting the bottom surface , particles followed the rules according to the chosen scenario:1 ) Reflection at a randomly chosen angle in the “random” scenario , where the new vector of movement was randomly chosen ( “impossible” movements , for instance into the surface were excluded ) . 2 ) Reflection with an angle identical to the angle of infliction in the “billiard” scenario , the vector of “take off” thus mirroring the vector of “landing” or 3 ) particles started swimming along the surface in the “NSS” scenario . In this scenario the z-vector of movement was set to zero ( particles thus following the surface ) ; the x and y-direction of movement remaining unchanged . When encountering the remaining 5 limiting surfaces of the cube , particles were simply reflected . Upon hitting the limits of the sphere , particles had a 10% chance of being attached ( simulating stopping/docking ) ; otherwise particles were simply reflected . Particles hitting the sphere during NSS also stopped/docked to the sphere with a likelihood of 10% . Upon each change of direction the movement vectors were adjusted to achieve a constant overall velocity . After an identical number of calculated increments of particle movements the simulation was interrupted and screenshots were acquired . Scatter3D is implemented in C++ and has previously been used to simulate light scattering [61] . Calculations were performed on a desktop computer . Further details are available upon request . Strains S . TmΔ4 fliGHI , S . Tm-T1 fliGHI , S . Tm-T1 -Fi fliGHI were constructed by P22 transduction [62] of the tetracycline allele of SB245 ( sipABCDsptP::aphT , fliG/H::Tn10 , K . Kaniga and J . E . Galan , unpublished data ) into strains S . TmΔ4 , S . Tm-T1 and S . Tm-T1 -Fi , respectively . flgK was deleted in SL1344 ( SB300 ) as described in [35] . Strains S . TmΔ4 flgK , S . Tm-T1 flgK , S . Tm-T1 -Fi flgK were constructed by P22 transduction of the chloramphenicol resistance containing the flgK deletion into the respective host strain . MotA/B were deleted in SL1344 ( SB300 ) using the method of Datsenko and Wanner [63] by insertion of a chloramphenicol resistance cassette that was amplified using the forward primer: ATGCTTATCTTATTAGGTTACCTGGTGGTTATCGGTACAGTGTAGGCTGGAGCTGCTTC and the reverse primer: TCACCTCGGTTCCGCTTTTGGCGATGTGGGTACGCTTGCATGGGAATTAGCCATGGTCC . S . TmΔ4 motAB , S . Tm-T1 -motAB and S . Tm-T1 -Fi motAB were obtained by P22 transduction of the chloramphenicol resistance into the respective host strain . Lack of motility of the respective strains was tested on motility agar . pM2120 ( expressing mCherry constitutively ) was constructed by PCR amplification of the mCherry gene using forward primer: CGCGGATCCCCCGGGCTGCAGGAATTCAGGAAACAGTATTCATGGTGAGCAAGGGCGAGGAG ( BamHI ) and reverse primer: GGGAAGCTTGATATATCGGAATTCTTACTTGTACAGCTCGTCCATG ( HindIII ) . Subsequently the PCR product as well as plasmid pM975 [22] was digested using BamHI/HindIII and ligated . fliG ( 11765205 ) , fliH ( 11765206 ) , fliI ( 11767468 ) , flgK ( 11767368 ) , motA ( 11765169 ) , motB ( 11765168 ) , cheY ( 11765165 ) , sopE ( 11765807 ) , sipA ( 11765948 ) , sopE2 ( 11768039 ) , sopB ( 1252609 ) , invG ( 11765959 ) , fimD ( 11764167 ) .
|
The animal body is protected by physical , chemical and immunological barriers . Identification of “promising” target sites is therefore of importance for any pathogen . This crucial step of the infection is still poorly understood . Here , we have studied target site selection by the flagellated Gram-negative gut pathogen Salmonella Typhimurium . Using a well-established tissue culture model system , we found that flagella-driven motility forces the bacterium into a “near surface swimming” mode which facilitates “scanning” of the host cell surface . The near surface swimming was found to target the pathogen towards sites with particular topological features , i . e . , rounded cells and membrane ruffles . This explains how S . Typhimurium “identifies” particular target sites and infects membrane ruffles in a cooperative manner . Interestingly , the near surface swimming is attributable to generic physical principles acting on moving particles . Therefore , our findings might be of general importance for the infection by motile pathogens .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"gram",
"negative",
"biology",
"microbiology",
"host-pathogen",
"interaction",
"bacterial",
"pathogens"
] |
2012
|
Near Surface Swimming of Salmonella Typhimurium Explains Target-Site Selection and Cooperative Invasion
|
Power laws , that is , power spectral densities ( PSDs ) exhibiting behavior for large frequencies f , have been observed both in microscopic ( neural membrane potentials and currents ) and macroscopic ( electroencephalography; EEG ) recordings . While complex network behavior has been suggested to be at the root of this phenomenon , we here demonstrate a possible origin of such power laws in the biophysical properties of single neurons described by the standard cable equation . Taking advantage of the analytical tractability of the so called ball and stick neuron model , we derive general expressions for the PSD transfer functions for a set of measures of neuronal activity: the soma membrane current , the current-dipole moment ( corresponding to the single-neuron EEG contribution ) , and the soma membrane potential . These PSD transfer functions relate the PSDs of the respective measurements to the PSDs of the noisy input currents . With homogeneously distributed input currents across the neuronal membrane we find that all PSD transfer functions express asymptotic high-frequency power laws with power-law exponents analytically identified as for the soma membrane current , for the current-dipole moment , and for the soma membrane potential . Comparison with available data suggests that the apparent power laws observed in the high-frequency end of the PSD spectra may stem from uncorrelated current sources which are homogeneously distributed across the neural membranes and themselves exhibit pink ( ) noise distributions . While the PSD noise spectra at low frequencies may be dominated by synaptic noise , our findings suggest that the high-frequency power laws may originate in noise from intrinsic ion channels . The significance of this finding goes beyond neuroscience as it demonstrates how power laws with a wide range of values for the power-law exponent α may arise from a simple , linear partial differential equation .
The apparent ubiquity of power laws in nature and society , i . e . , that quantities or probability distributions satisfy the relationship ( 1 ) where α is the power-law exponent , has for a long time intrigued scientists [1] . Power laws in the tails of distributions have been reported in a wide range of situations including such different phenomena as frequency of differently sized earth quakes , distribution of links on the World Wide Web , paper publication rates in physics , and allometric scaling in animals ( see [1] and references therein ) . A key feature of power laws is that they are scale invariant over several orders of magnitude , i . e . , that they do not give preference to a particular scale in space or time . There are several theories with such scale invariance as its fingerprint , among the most popular are fractal geometry [2] and the theory of self-organized critical states [3] . Conspicuous power laws have been seen also in the field of neuroscience [4] , among the most prominent the observed power laws in the size distribution of neuronal ‘avalanches’ [5] , [6] and in the high-frequency tails of power spectral densitites ( PSDs ) of electrical recordings of brain activity such as electroencephalography ( EEG ) [7] , [8] , electrocorticography ( ECoG ) [9]–[12] , the local field potential ( LFP ) [13]–[16] , and the soma membrane potential and currents of individual neurons [17]–[21] . To what extent these various power laws have the same origin , is currently not known [4] , [6] . In any case , it is the latter type of power law , i . e . , those observed in the PSDs of electrical recordings , which is the topic of the present paper . Ever since Hans Berger recorded the first human electroencephalogram ( EEG ) in 1924 [22] , its features have been under extensive study , especially since many of them are directly related to disease and to states of consciousness . In the last decades the frequency spectra of the EEG has , for example , attracted significant attention as the high-frequency part of the PSD in experiments with maximal frequencies typically in the range 30–100 Hz has often well fitted by a power laws with α typically in the range from 1 to 2 . 5 [7] , [8] . Such apparent power laws have not only been seen in macroscopic neural recordings such as EEG , ECoG and LFP , they also appear at the microscopic level , i . e . , in single-neuron recordings . PSDs of the subthreshold membrane potentials recorded in the somas of neurons often resemble a power law in their high-frequency ends ( 100–1000 Hz ) , typically with a larger exponent α ranging from 2 to 3 [17]–[21] . This particular power law seems to be very robust: it has been observed across species , brain regions and different experimental set-ups , such as cultured hippocampal layer V neurons [17] , pyramidal layer IV–V neurons from rat neocortex in vitro [19] , [20] , and neocortical neurons from cat visual cortex in vivo [18] , [21] . At present , the origin , or origins , of these macroscopic and microscopic power laws seen in PSDs of neural recordings are actively debated [4] , [6] . Lack of sufficient statistical support have questioned the validity of identified power-law behaviors , and as a rule of thumb , it has been suggested that a candidate power law should exhibit an approximately linear relationship in a log-log plot over at least two orders of magnitude [1] . Further , a mechanistic explanation of how the power laws arise from the underlying dynamics should ideally be provided [1] . In the present paper we show through a combination of analytical and numerical investigations how power laws in the high-frequency tail of PSDs naturally can arise in neural systems from noise sources homogeneously distributed throughout neuronal membranes . We further show that the mechanism behind microscopic ( soma potential , soma current ) power laws will also lead to power laws in the single-neuron contribution ( current-dipole moment ) to the EEG . Moreover , we demonstrate that if all single-neuron contributions to the recorded EEG signal exhibit the same power law , the EEG signal will also exhibit this power law . We find that for different measurement modalities different power-law exponents naturally follow from the well-established , biophysical cable properties of the neuronal membranes: the soma potential will be more low-pass filtered than the corresponding current-dipole moment determining the single-neuron contribution to the EEG [23] , [24] , and as a consequence , the power-law exponent α will be larger for the soma potential than for the single-neuron contribution to the EEG [25] ( see illustration in Fig . 1 ) . When comparing with experimental data , we further find that for the special case when uncorrelated and homogeneously distributed membrane-current sources themselves exhibit power laws in their PSD , the theory predicts power-law exponents α in accordance with experimental observations for the microscopic measures , i . e . , the soma current and soma potential . The experimental situation is much less clear for the EEG signal where frequency spectra presently is limited upwards to 100 Hz . However , we note that under the assumption that such single-neuron sources dominate the high-frequency part of the EEG signal , the theoretical predictions are also compatible with the power-law-like behavior so far observed experimentally . Both synaptic noise and intrinsic channel noise will in general contribute to the observed noise spectra , cf . Fig . 1 . While our theory per se is indifferent to the detailed membrane mechanism providing the noisy current , our findings suggests that the dominant noise source underlying the observed high-frequency power laws seen in PSDs may be channel noise: prevalent theories for synaptic currents are difficult to reconcile with a power law in the high-frequency tail of power spectra , while potassium ion channels with such noise spectra indeed have been observed [26] . Note that this does not imply that channel noise in general dominates synaptic noise in electrophysiological power spectra: it only suggests that the high-frequency power-law part , which in the in vivo situation typically represents a tiny fraction of the overall noise power , is dominated by channel noise . Through the pioneering work by Wilfred Rall half a century ago [27] , [28] the ball and stick neuron model was established as a key model for the study of the signal processing properties of neurons . An important advantage is the model's analytical tractability , and this is exploited in the present study . We first demonstrate the relevance of this simplified model in the present context by numerical comparisons with results from a morphologically reconstructed multicompartmental pyramidal neuron model . Then we derive analytical power-law expressions for the various types of electrophysiological measurements . While a single current input onto a dendrite does not give rise to power laws , we here show that power laws naturally arise for the case with homogeneously distributed inputs across the dendrite and the soma [29] , see Fig . 1 . For this situation we show that the ball and stick neuron model acts as a power-law filter for high frequencies , i . e . , the transfer function from the PSD of the input membrane currents , , to the PSD of the output ( soma potential , soma current , or current-dipole moment setting up the EEG ) , , is described by a power law: . Notably the analytically derived power-law exponents α for these transfer functions are seen to be different for the different measurement modalities . The analytical expressions further reveal the dependence of the PSDs on single-neuron features such as the correlation of input currents , dendritic length and diameter , soma diameter and membrane impedance . The theory presented here also contributes to -theory in general [30]: it illustrates that a basic physics equation , the cable equation , can act as a power-law filter for high frequencies when the underlying model has spatially distributed input . Furthermore , α may have any half-numbered value between 1/2 and 3 , depending on the physical measure ( some potential , soma current , single-neuron contribution to the EEG ) under consideration , and the coherence of the input currents . Intuitively , the emergence of the power-law spectra can be understood as a result of a superposition of simple low-pass filters with a wide range of cutoff frequencies due to position-dependent dendritic filtering of the spatially extended neuron [23] , . This is in accordance with the orginal idea of Schottky from 1926 [32] that the shot-noise observed in vacuum tubes by Johnson could be understood by the combined action of a continuous distribution of ‘exponential relaxation processes’ [33] . The paper is organized as follows: In the next section we derive analytical expressions for the soma potential , soma current and current-dipole moment for the ball and stick neuron for the case with noisy current inputs impinging on the soma ‘ball’ and homogeneously on the dendritic stick . While these derivations are cumbersome , the final results are transparent: power laws are observed for all measurement modalities in the high-frequency limit . In Results we first demonstrate by means of numerical simulations the qualitative similarity of the power-law behaviors between the ball and stick model and a biophysically detailed pyramidal neuron . We then go on to analytically identify the set of power-law exponents for the various measurement modalities both in the case of uncorrelated and correlated current inputs . While the derived power laws strictly speaking refer to the functional form of PSDs in the high-frequency limit ( Eq . 1 ) , the purported power laws in neural data have typically been observed for frequencies less than a few hundred hertz . Our model study implies that the true high-frequency limit is not achieved at these frequencies . However , in our ball and stick model , quasi-linear relationships can still be observed in the characteristic PSD log-log plots for the experimentally relevant frequency range . These apparent power laws typically have smaller power-law exponents than their respective asymptotic values . The numerical values of these exponents will depend on details in the neuron model , but the ball and stick model has a very limited parameter space: it is fully specified by four parameters , a dimensionless frequency , the dimensionless stick length , the ratio between the soma and infinite-stick conductances , and the ratio between the somatic and dendritic current density . This allows for a comprehensive investigations of the apparent power-law exponents in terms of the neuron parameters , which we pursue next . To facilitate comparison with experiments we round off the Results section exploring how PSDs , and in particular apparent power laws , depend on relevant biophysical parameters . In the Discussion we then compare our model findings with experiments and speculate on the biophysical origin of the membrane currents underlying the observed PSD power laws .
For a cylinder with a constant diameter d the cable equation is given by ( 2 ) with the length constant and the time constant . , and denote the specific membrane resistance , the specific membrane capacitance and the inner resistivity , respectively , and have dimensions , and . Lower-case letters are used to describe the electrical properties per unit length of the cable: , and , with units , and . For convenience , the specific membrane conductance , , will also be used , see Table 1 for a list of symbols . With dimensionless variables , and , the cable equation , Eq . 2 , can be expressed ( 3 ) Due to linearity , each frequency component of the input signal can be treated individually . For this , it is convenient to express the membrane potential in a complex ( boldface notation ) form , ( 4 ) where is a complex number containing the amplitude and phase of the signal , and the dimensionless frequency is defined as . The complex potentials are related to the measurable potential through the Fourier components of the potential , ( 5 ) where is the direct current ( DC ) potential . The cable equation can then be simplified to ( 6 ) where , see [23] , [31] . The general solution to Eq . 6 can be expressed as ( 7 ) The expression for the axial current is given by ( 8 ) and is applied at the boundaries to find the specific solutions for the ball and stick neuron . In complex notation and with dimensionless variables this can be expressed as ( 9 ) where is the infinite-stick conductance . Similarly , the transmembrane current density ( including both leak currents and capacitive currents ) is given by ( 10 ) with its complex counterpart , ( 11 ) The ball and stick neuron [27] consists of a dendritic stick attached to a single-compartment soma , see Fig . 3A . Here we envision the stick to be a long and thin cylinder with diameter d and length l . The membrane area of the soma is set to be , corresponding to the surface area of a sphere with diameter , or equivalently , the side area of a cylindrical box with diameter and height . The solution of the cable equation for a ball and stick neuron with a single input current at an arbitrary dendritic position is found by solving the cable equation separately for the neural compartment proximal to the input current and the neural compartment distal to the input current , These solutions are then connected through a common voltage boundary condition at the connection point . For the proximal part of the stick , Ohm's law in combination with the lumped soma admittance gives the boundary condition at the somatic site , and for the distal part of the stick , a sealed-end boundary is applied at the far end . In this configuration the boundary condition acts as the driving force of the system . The potential can , however , also be related to a corresponding input current through the input impedance , i . e . , . Above we derived transfer functions for the ball and stick neuron , connecting current input at an arbitrary position on the neuron to the various measurement modalities , i . e . , the current-dipole moment ( ) , the soma potential ( ) and the soma current ( ) . We will now derive expressions for the PSDs when the ball and stick neuron is bombarded with multiple inputs assuming that all input currents have the same PSD and a pairwise coherence [37] . The PSDs can then be divided into separate terms for uncorrelated ( ) and fully correlated ( ) input . The PSD , , of the output can for the case of multiple current inputs be expressed as ( 39 ) where is the PSD of the input currents , is their coherence and is the transfer function between the PSD of the input and the PSD of the output . The complex conjugate is denoted by the asterisk . We now assume the first of the input currents to be positioned at the soma compartment , and the rest of the input to be spread homogeneously across the dendritic stick . The transfer function for the soma compartment , , is the same for all somatic inputs , for , while the input transfer function for the dendritic stick is position dependent , for . The PSD transfer function can then be expressed ( 40 ) To allow for analytical extraction of power laws , we next convert the sums into integrals . By assuming uniform current-input density ( per membrane area ) in the dendritic stick ( given by ) , it follows that the axial density of current inputs is . In the continuum limit ( ) we thus have ( 41 ) where the last factor comes from the conversion to dimensionless lengths . The PSD transfer function , , in Eq . 40 can then be split into three parts , ( 42 ) where ( 43 ) is the PSD transfer function for uncorrelated input at the soma compartment , ( 44 ) is the PSD transfer function for uncorrelated input distributed throughout the dendritic stick , and ( 45 ) is the PSD transfer function for correlated input distributed both across the dendritic stick and onto the soma . We have now derived ( i ) a general expressions for the PSD transfer function expressed by the general , single-input transfer functions and , and ( ii ) specific analytical expressions for the single-input transfer functions for the dipole moment , the soma potential and the soma current . We will next combine these results and analytically derive specific PSD transfer functions for the dipole moment , the soma potential and the soma current for distributed input . For convenience we here summarize the results , now solely in terms of dimensionless variables ( except for the amplitudes ) , i . e . , , , , and ( see Table 2 ) . The general expression for the PSD transfer functions reads: ( 79 ) where represents the contributions from uncorrelated current inputs , represents the contributions from correlated inputs , and is the pairwise coherence function . The contributions from uncorrelated input currents are in turn given as sums over contributions from somatic and dendritic inputs , i . e . , ( 80 ) The contribution to the PSD transfer functions for correlated input currents are given by ( 81 ) ( 82 ) ( 83 ) with the squared norm of given by Eq . 49 , and and defined by Eqs . 50 and 51 , respectively . The contributions from uncorrelated dendritic inputs are: ( 84 ) ( 85 ) ( 86 ) In the special case with input to soma only , the PSD transfer functions are the same for uncorrelated ( Eq . 43 ) and correlated input ( Eq . 45 ) , the only difference being the amplitudes , ( implies that the input is onto soma only . ) The corresponding PSD transfer functions from uncorrelated somatic input thus become ( 87 ) ( 88 ) ( 89 ) In an infinite , homogenous , isotropic Ohmic medium with conductivity , the extracellular potential recorded at a given position far away from a single-neuron current dipole is given by [25] , [38] . ( 90 ) where designates the spatial position of the current dipole , is the magnitude of the current-dipole moment , and is the angle between the dipole moment vector and the position vector . An important feature is that all time dependence of the single-neuron contribution to the potential lies in so that factorizes as ( 91 ) For the electrical potential recorded at an EEG electrode , the forward model in Eq . 90 is no longer applicable due to different electrical conductivities of neural tissue , dura matter , scull and scalp . Analytical expressions analogous to Eq . 90 can still be derived under certain circumstances such as with three-shell or four-shell concentric spherical head models ( see Nunez and Srinivasan [38] , Appendix G ) , but the key observation for the present argument is that the single-neuron contribution to the EEG will still factorize , i . e . , where here is an unspecified function . The compound EEG signal from a set of single-neuron current dipoles is now given by ( 92 ) where the index runs over all single-neuron current dipoles . For each Fourier component ( frequency ) we now have ( 93 ) For the special case where the different single-neuron current dipoles moments are uncorrelated we find that the power spectral density of the EEG is of the form [39] ( 94 ) ( We have here introduced the notation ‘UC’ , i . e . capitalized , to highlight the difference between the present assumption of uncorrelated single-neuron current dipoles and the separate assumption of uncorrelated membrane currents onto individual neurons in the above sections . ) If the single-neuron current dipoles have the same power-law behavior in a particular frequency range , i . e . , , it follows directly that the EEG signal will inherit this power-law behavior: ( 95 ) where determines the PSD amplitude , but not the slope . The inheritance of the single-neuron power-law behavior also applies to the case of correlated sources , provided that the pairwise coherences are frequency independent . By similar reasoning as above we then find ( 96 ) Analogous expressions for the PSD for the EEG can also be derived when both correlated and uncorrelated single-neuron current dipoles contribute , but we do not pursue this here; see Lindén et al . [37] and Leski et al . [39] for more details . The NEURON simulation environment [35] with the supplied Python interface [36] was used to simulate a layer-V pyramidal neuron from cat visual cortex [34] . The main motivation for pursuing this was to allow for a direct numerical comparison with results from the ball and stick neuron to probe similarities and differences , see Fig . 2 . In addition , NEURON was also used on the ball and stick neuron model to verify consistency with the analytical results above . Both the layer-V pyramidal neuron and the ball and stick neuron had a purely passive membrane , with specific membrane resistance , specific axial resistivity m , and specific membrane capacitance F/m . Simulations were performed with a time resolution of 0 . 0625 ms , and resulting data used for analysis had a time resolution of 0 . 25 ms . All simulations were run for a time period of 1200 ms and the first 200 ms were removed from the subsequent analysis to avoid transient upstart effects in the simulations . The digital cell reconstruction of the layer-V pyramidal neuron was downloaded from ModelDB ( http://senselab . med . yale . edu/ ) , and the axon compartments were removed . To ensure sufficient numerical precision compartmentalization was done so that no dendritic compartment was larger than 1/30th of the electrotonic length at 100 Hz ( using the function lambda_f ( 100 ) in NEURON ) , which resulted in 3214 compartments . The soma was modeled as a single compartment . The ball and stick neuron was modeled with a total of 201 segments , one segment was the iso-potential soma segment with length and diameter , and 200 segments belonged to the attached dendritic stick of length 1 mm and diameter . Simulations were performed with the same white-noise current trace injected into each compartment separately . The white-noise input current was constructed as a sum of sinusoidal currents [24] ( 97 ) where represents a random phase for each frequency contribution . Due to linearity of the cable equation , the contributions of individual current inputs could be combined to compute the PSD of the soma potential , the soma current and the dipole moment resulting from current injection into all compartments . In correspondence with Eq . 39 , the summation of the contributions from the input currents of different segments with membrane areas was done differently for uncorrelated and correlated input currents . The uncorrelated PSDs , , were computed according to ( 98 ) while the correlated PSDs , , were computed according to ( 99 ) Here , denotes the Fourier components of the signal ( either soma potential , soma current or dipole moment due to input in one segment ) , the product gives the total number of input currents into one segment , and the density represents for dendritic input and for somatic input . The total dipole moment was in the numerical computations assumed to equal the dipole moment in one direction only: the direction along the stick for the ball and stick model , and the direction along the apical dendrite for the pyramidal neuron model , both denoted as the -component , . For the pyramidal neuron this is an approximation as the dipole moment also will have components in the lateral directions . However , the prominent ‘open-field’ asymmetry of the pyramidal neuron in the vertical direction suggests that this is a reasonable approximation when predicting contributions to the EEG signal . The current-dipole moment is then given by ( 100 ) where is the transmembrane current of compartment , and is the corresponding -position .
To establish the relevance of using the simple ball and stick neuron to investigate the biophysical origin of power laws , we compare in Fig . 2 the normalized power spectral densities ( PSDs ) of the transmembrane soma current ( row 1 ) , the current-dipole moment ( row 2 ) , and the soma potential ( row 3 ) of this model ( column 1 ) with the corresponding results for a biophysically detailed layer-V pyramidal neuron ( column 2 ) ; the rightmost column gives a direct comparison of PSDs . Both neuron models have a purely passive membrane and receive spatially distributed current input . As described in the Models section , the PSD of the single-neuron contribution to the EEG will be proportional to the PSD of the neuronal current-dipole moment given the observation that the extracellular medium , dura matter , scull and scalp appear to be purely ohmic [24] , [38] . We here stick to the term ‘current-dipole moment’ even if the term ‘single-neuron contribution to the EEG’ could equally be used . A first striking observation is that unlike single-input PSDs ( thin gray lines in Fig . 2 ) , the PSDs resulting from numerous , homogeneously distributed input currents ( thick lines ) have a linear or quasi-linear appearance for high frequencies in these log-log plots , resembling power laws . This is seen both when the numerous current inputs are correlated ( green thick lines ) and uncorrelated ( blue thick lines ) . We also observe that the decay in the PSD with increasing frequency is strongest for the soma potential , somewhat smaller for the current-dipole moment , and smallest for the soma current . This is reflected in the power-law exponents estimated at 1000 Hz from these PSDs , see legend in Fig . 2 . Here we observe that is largest for the soma potential ( bottom row ) and smallest for the soma current ( top row ) . In the example in Fig . 2 we have assumed constant input current densities across the neurons , i . e . , . For this special case , correlated current input will , at all times , change the membrane charge density equally across the neuron , and as a consequence the neuron will be iso-potential . In this case the axial current within the neuron will be zero , and likewise the net membrane current ( with the capacitive current included ) for any compartment , including the soma . As a consequence the current-dipole moment vanishes , and the model can effectively be collapsed to an equivalent single-compartment neuron . For the soma current and dipole moment we thus only show results for uncorrelated inputs in Fig . 2 . However , correlated current input will still drive the soma potential ( green curves in columns 1 and 2 ) . Here we observe that the exponent is smaller for uncorrelated input than for correlated input both for the ball and stick neuron and for the pyramidal neuron . The results above pertain to the situation with white-noise current inputs , i . e . , flat-band PSDs . However , the results are easily generalized to the case with current inputs with other PSDs . Since our neuron models are passive and thus linear , the PSDs simply multiply . This is illustrated in column 3 of Fig . 2 which shows how our PSDs for uncorrelated input change with varying PSDs of the current input , . The blue curves correspond to white-noise input and are identical to the blue curves in column 2 . The pink and brown curves illustrate the case of pink ( ) and Brownian ( ) input , respectively . Since the PSDs multiply , the power-law exponent of the input noise simply adds to the exponent . Thus , the pink and Brownian input increase the slope with and , respectively , compared to white-noise input . Even though the dendritic structure of the reconstructed pyramidal neuron is very different from the ball and stick neuron in that it has both a highly branched structure and a varying diameter along its neural sections ( tapering ) , both models seem to produce linear or quasi-linear high-frequency PSDs in the log-log representation . Also the power-law exponents are found to be fairly similar . This implies that the ball and stick neuron model captures salient power-law properties of the more biophysically detailed neuron model , and motivates our detailed analytical investigation of the power-law properties of the ball and stick neuron following next . In the Models section above we derived analytical expressions for the PSD transfer functions of the soma current ( ) , current-dipole moment ( ) and soma potential ( ) for the ball and stick neuron for spatially distributed input currents . The resulting transfer functions , summarized in Eqs . 79–89 , were of the form ( 101 ) where and represent the contributions from uncorrelated somatic and dendritic inputs , respectively , and represents the contribution from correlated inputs . is the pairwise coherence of the current inputs , all assumed to have the same PSDs ( ) . These mathematical expressions are quite cumbersome , but they are dramatically simplified in the high-frequency limit , , in which the dominant power can be found analytically by a series expansion of the mathematical expressions for the transfer functions in Eqs . 81-89 . The expressions for the PSD transfer functions contain terms which are both polynomial and superpolynomial ( i . e . , including exponentials/exponentially decaying functions ) with respect to frequency . As these superpolynomial terms will dominate the polynomial terms in the high-frequency limit , it follows from Eq . 49 that for high frequencies the absolute square of the denominator can be approximated by ( 102 ) where terms decaying exponentially to zero with increasing frequency have been set to zero . The frequency dependence is through and , see Eqs . 50 and 51 . Note that since . In the high-frequency limit the PSD transfer functions Eqs . 81–89 become ( 103 ) ( 104 ) ( 105 ) ( 106 ) ( 107 ) ( 108 ) where the amplitudes are found in Table 2 . When the PSDs expressed in Eqs . 103-107 are expanded reciprocally for high frequencies , i . e . , , we get ( 109 ) ( 110 ) ( 111 ) ( 112 ) ( 113 ) ( 114 ) where is the dimensionless relative density , , and , with and denoting the somatic and dendritic diameter , respectively , and denoting the dendritic length constant . The expansions were done in Mathematica ( version 7 . 0 ) , and a list of parameters used throughout the present paper is given in Table 1 ( along with the default numerical values used in the numerical investigations in later Results sections ) . In Eqs . 109–114 terms which are exponentially decaying to zero for large have been approximated to zero . Note that Eq . 114 does not apply in the special case of no somatic input , , for which the series expansion gives ( 115 ) The corresponding high frequency expansions of the PSD transfer functions for uncorrelated somatic input , , are not shown , as these expressions are identical to the corresponding transfer functions for correlated input into the soma only , ( i . e . , equal to Eqs . 110 , 112 and 114 with ) . Eqs . 109–115 show that , due to position-dependent frequency filtering of the numerous inputs spread across the membrane ( cf . Fig . 3B ) , all PSD transfer functions express asymptotic high-frequency power laws . Moreover , these genuine ‘infinite-frequency’ power-law exponents , denoted , span every half power from ( for , Eq . 109 ) to ( for , Eq . 115 ) for the different transfer functions . The results are summarized in Table 2 . To obtain the power-law exponents in the general case with contributions from both correlated and uncorrelated current inputs , we need to compare the different terms in the general expression for in Eq . 101 . With different leading power-law exponents in their asymptotic expressions , the term with the lowest exponent will always dominate for sufficiently high frequencies . From Table 2 we see that for all three quantities of interest , i . e . , , and , the lowest exponent always comes from contributions from uncorrelated inputs . Note that the correlated term in Eq . 101 also involves a frequency-dependent coherence term , but to the extent it modifies the PSD , it will likely add an additional low-pass filtering effect [39] and , if anything , increase the power-law exponent . If we assume that the coherence is constant with respect to frequency we identify the following asymptotic exponents ( i . e . , with ‘all’ types of possible input ) for , and : Note that these power-law exponents are unchanged as long as uncorrelated activity is distributed both onto the soma and the dendrite , but will increase to and if no uncorrelated input are present on the dendrite . Similarly , without input onto soma , the asymptotic value will change for the soma potential PSD: it becomes if uncorrelated input is uniformly distributed on the dendrite , and if the dendritic input is correlated . Detailed inspection of the power-law slopes for the ball and stick model in Fig . 2 and comparison with the power-law exponents listed in Table 2 reveal that although the curves might look linearly decaying in the log-log plot for high frequencies , the expressed exponents are still deviating from their high-frequency values , even at 1000 Hz . As experimental power laws have been claimed for much lower frequencies than this , we now go on to investigate apparent PSD power laws for lower frequencies . For this it is convenient to define a low-frequency ( lf ) regime , an intermediate-frequency ( if ) regime and a high-frequency ( hf ) regime , as illustrated in Fig . 3C . The transition frequencies between the regimes are given by the frequencies at which is and of , respectively . The log-log decay rates of the PSD transfer functions can be defined for any frequency by defining the slope as the negative log-log derivative of the PSD transfer functions , ( 116 ) In Figs . 4 , 5 , and 6 we show color plots of for the soma current , current-dipole moment , and soma potential , respectively , both for cases with uncorrelated and correlated inputs . The depicted results are found by numerically evaluating Eq . 116 based on the expressions for listed in Eqs . 81–89 . Note that since our model is linear , the log-log derivative is independent of the amplitude . Thus , with either completely correlated or completely uncorrelated input , the dimensionless parameters , , and span the whole parameter space of the model . The 2D color plots in Figs . 4–6 depict as function of and for three different values of the electronic length ( = 0 . 25 , 1 , and 4 ) , i . e . , spanning the situations from a very short dendritic stick ( ) to a very long stick ( ) . Electrotonic lengths greater than produced plots that were indistinguishable by eye from the plots for . The thin black contour line denotes the transition between the low- and intermediate-frequency regimes ( ) , whereas the thick black contour line denotes the transition between the intermediate- and high-frequency regimes ( ) . The 2D color plots in Figs . 4–6 depicting the slopes of the PSDs of the transfer functions , give a comprehensive overview of the power-law properties of the ball and stick model as they are given in terms of the three key dimensionless parameters , , and . To get an additional view of how the model predictions depend on biophysical model parameters , we plot in Figs . 7 and 8 PSDs , denoted , for a range of model parameters for the soma current , current-dipole moment and soma potential when the neuron receives homogeneous white-noise current input across the dendrite and/or the soma . We focus on biophysical parameters that may vary significantly from neuron to neuron: the dendritic stick length , the specific membrane resistance , the dendritic stick diameter , and the soma diameter . The specific membrane resistance may not only vary between neurons , but also between different network states for the same neuron [40] , [41] . To predict PSDs of the various measurements , and not just PSDs of the transfer functions , we also need to specify numerical values for the current-input densities and ( and not only the ratio ) , as well as the magnitude of the PSDs of the current inputs . These choices will only affect the magnitudes of the predicted PSDs , not the power-law slopes . As the numerical values of the high-frequency slopes predicted by the present work suggest that channel noise from intrinsic membrane conductances rather than synaptic noise dominates the observed apparent high-frequency power laws in experiments ( see Discussion ) , we gear our choice of parameters towards intrinsic channel noise . We first assume the input densities and ( when they are non-zero ) to be 2 m , in agreement with measurements of the density of the large conductance calcium-dependent potassium ( BK ) channel [42] . Next we assume the magnitude of PSD of the white-noise current input to be = const = 1 fA2/Hz . This choice for gives magnitudes of predicted PDSs of the soma potential , assuming uncorrelated current inputs , in rough agreement with what was observed in the in vitro neural culture study of [17] , i . e . , about 10−3–10−2 mV2/Hz for low frequencies . Note that the shape of the PSDs , and thus estimated power-law exponents , are independent of the choice of current-noise amplitude . Figs . 7 and 8 show PSDs for uncorrelated and correlated input currents , respectively . A first observation is that the predicted PSD magnitudes are typically orders of magnitude larger for correlated inputs , than for uncorrelated inputs . With the present choice of parameters , the cases with correlated inputs predict PSDs for the soma potential and soma current much larger than what is seen in in vitro experiments [17] , [19] , [20] . A second observation is that variations in the dendritic stick length ( first column in Figs . 7–8 ) and membrane resistance ( second column ) typically have little effect on the PSDs at high frequencies , but may significantly affect the cut-off frequencies , i . e . , the frequency where the PSD kinks downwards . This may be somewhat counterintuitive , especially that the PSDs for the current-dipole moment are independent of stick length as one could think that a longer stick gives a larger dipole moment . For the ball and stick neuron , however , this is not so: input currents injected far away from both boundaries ( ends ) of a long stick will not contribute to any net dipole moment , as the input current will return symmetrically on both sides of the injection point and thus form a quadrupole moment . This symmetry is broken near the ends of the stick: for uncorrelated input a local dipole is created at each endpoint; for correlated input the dendrite will be iso-potential near the distal end of the stick , while a local dipole will arise at the somatic end if . Note though that this is expected to be different for neurons with realistic dendritic morphology , since the dendritic cables typically are quite asymmetric due to branching and tapering . The effects of varying the dendritic stick diameter and soma diameter are quite different ( cf . , two rightmost columns in Figs . 7–8 ) . Here both the magnitudes and the slopes of the high-frequency parts are seen to be significantly affected . On the other hand , the cut-off frequency is seen to be little affected when varying the soma diameter , in particular for the current-dipole ( ) and soma potential ( ) PSDs . ( Note that for the case with homogeneous correlated input , ( row 4 in Fig . 8 ) , the ball and stick model is effectively reduced to a single-compartment neuron for which the PSD is independent of and . ) In Figs . 4–6 regions in the log-log slope plots were observed to have positive double derivatives , i . e . , concave curvature . The effect was particularly prevalent for the soma potential transfer function in the case of short dendritic sticks ( ) with dominant current input to the soma . This feature is also seen in the corresponding ‘soma-input’ curves ( bottom rows of Figs . 7–8 ) , also for non-compact sticks , i . e . , for the default value = 1 mm ( = 1 ) .
In the present work we have taken advantage of the analytical tractability of the ball and stick neuron model [27] to obtain general expressions for the power spectral density ( PSD ) transfer functions for a set of measures of neural activity: the somatic membrane current , the current-dipole moment ( corresponding to the single-neuron EEG contribution ) , and the soma potential . With homogeneously distributed input currents both onto the dendritic stick and with the same , or another current density , onto the soma we find that all three PSD transfer functions , relating the PSDs of the measurements to the PSDs of the noisy inputs currents , express asymptotic high-frequency power laws . The corresponding power-law exponents are analytically identified as for the somatic membrane current , for the current-dipole moment , and for the soma potential . These power-law exponents are found for arbitrary combinations of uncorrelated and correlated noisy input current ( as long as both the dendrites and the soma receive some uncorrelated input currents ) . The significance of this finding goes beyond neuroscience as it demonstrates how power laws with a wide range of values for the power-law exponent may arise from a simple , linear physics equation [30] . We find here that the cable equation describing the electrical properties of membranes , transfers white-noise current input into ‘colored’ -noise where may have any half-numbered value within the interval from to 3 for the different measurement modalities . Intuitively , the physical underpinning of these novel power laws is the superposition of numerous low-pass filtered contributions with different cut-off frequencies ( i . e . , different time constants ) [32] , [33] due to the different spatial positions of the various current inputs along the neuron . ( Note , however , that power laws with integer coeffients ( 1 and 2 ) also are obtained with purely somatic input; cf . Table 2 . ) As our model system is linear , the results directly generalize to any colored input noise , i . e . , transferring spectra of input currents to output spectra . Our ball and stick model expressions for the PSDs cover all frequencies , not just the high frequencies where the power-law behavior is seen . When comparing with results from neural recordings , one could thus envision to compare model results with experimental results across the entire frequency spectrum . However , the experimental spectra will generally be superpositions of contributions from numerous sources , both from synapses [41] and from ion channels [17] . These various types of input currents will in general have different PSDs , i . e . , different . A full-spectra comparison with our theory is thus not possible without specific assumptions about the types and weights of the various noise contributions , information which is presently not available from experiments . However , the presence of power-law behavior at high frequencies implies that a single noise process ( or several noise processes with identical power-law exponents ) dominates the others in this frequency range . In the following we first discuss apparent power laws observed in the soma potential and soma current in vitro [17] , [19] , [20] . Next , we discuss apparent power laws seen in vivo , both in the soma potential [18] , [21] , [43] and , briefly , in the EEG [7] . Here synaptic noise is expected to provide almost all of the noise variance , but our results suggest that the power law at the high-energy tail of the spectrum nevertheless may be due to ion-channel noise . Power laws have also been reported in recordings of extracellular potentials inside ( local field potential; LFP ) and at the surface of cortex ( electrocorticography; ECoG ) . However , the reported power-law exponents vary a lot , with 's between 1 and 3 for LFPs [13]–[16] and between 2 and 4 for ECoG signals [9]–[12] , [50] . From a modeling perspective the single-neuron contribution to putative power-law exponents for these signals is more difficult as , unlike the EEG signal , the single-neuron contributions are not determined only by the current-dipole moment: dominant contributions to these signals will in general also come from neurons close to the electrode ( typically on the order of hundred or a few hundred micrometers [37] ) , so close that the far-field dipole approximation relating the current-dipole moment directly to the contributed extracellular potential [25] is not applicable [37] . A point to note , however , is that it may very well be that power laws observed in the LFP or ECoG are dominated by other current sources than the power laws observed in the EEG spectra: As observed in [37] , [39] ( see also [51] ) the LFP recorded in a cortical column receiving correlated synaptic inputs can be very strong , and it is thus at least in principle conceivable that power laws in the LFP may stem from synaptic inputs from neurons surrounding the electrode , whereas the EEG signal , which picks up contributions from a much larger cortical area , may be dominated by uncorrelated noise from ion channels . Further , the soma potential and soma current of each single neuron may also still be dominated by uncorrelated channel noise , even if the LFP is dominated by correlated synaptic activity . This is because correlated synaptic inputs onto a population of neurons add up constructively in the LFP , whereas the uncorrelated inputs do not [37] , [39] . For single-neuron measures such as the soma potential and soma current there will be no such population effects , and the uncorrelated inputs may more easily dominate the power spectra . As a final comment it is interesting to note that in the only reported study we are aware of for the frequency range 300-3000 Hz , the PSD of the LFP exhibited a power law with a fitted exponent of = 1 . 1 [15] . This is very close to what would be predicted if the LFP was dominated by the soma current from uncorrelated ( pink ) noise sources: In Table 2 we see that the ‘infinite-frequency’ power-law exponent for the transfer function from dendritic current inputs to soma current is . With a pink ( ) PSD of the input noise current , the ‘infinite-frequency’ prediction for the soma current exponent will thus be 1 . 5 . This is already fairly close to the experimental observation of 1 . 1 . Further , from Fig . 4 it follows that the apparent power-law coefficient for the transfer-function power law may be somewhat smaller than 0 . 5 in the frequency range of interest , suggesting that the agreement between experiments and model predictions assuming uncorrelated noise may be even better . If so , it may be that the LFP power spectra are dominated by synaptic inputs for frequencies below a few hundred hertz ( with rapidly decaying LFP contributions with increasing frequency , i . e . , higher power-law exponents in accordance with [13] , [14] , [16] ) , while uncorrelated inputs , and thus power laws with smaller exponents , dominate at higher frequencies . In the present analysis we have modeled the membranes of somas and dendrites as simple passive linear ( RC ) circuit elements . This implies a strictly linear response to the current inputs , allowing for the present frequency-resolved ( Fourier ) analysis . However , the same kind of analysis can be done for active dendritic membrane conductances , at least close to the resting potential of the neuron: In the so called quasi-active membrane models , the active conductances are linearized and modeled by a combination of resistors , capacitors and inductors [52] , [53] . These extra circuit elements will change the PSD . For example , the inductor typically introduces a resonance in the system . In Koch [53] the impedance for this ‘quasi-active’ membrane was however found to coincide with the impedance for a purely passive membrane for frequencies above 200 Hz , implying that the predicted high frequency power laws will be about the same . This is in accordance with experimental results from neocortical slices , where blocking of sodium channels were shown mainly to affect the soma potential PSD for frequencies below 2 Hz [19] . Nevertheless , the investigation of the role of active conductance on PSDs is a topic deserving further investigations . Here we modeled the noise-generating membrane mechanism as a simple current , i . e . , , making the system fully linear . As a ( non-linear ) alternative , these noise currents could have been modeled as conductance-based currents , i . e . , where is the conductance , and is the channel reversal potential . In the case of potassium channels , will typically be around -80 mV . However , when exploring the situation when the membrane potential is not too close to the channel reversal potential , we observed in simulations the same high-frequency power-law behavior for conductance-based and current-based noise-current models ( results not shown ) . That these two models give the same power law can be understood as follows: In the conductance-based case the channel current has two terms , i . e . , . The conductance is here dependent on the incoming spike trains , but not on the membrane potential . The first term involves a product of and , while the second term has the same mathematical form as the current-based noise model . Since the potential membrane potential always will be low-pass filtered compared to the input , the linear term is expected to dominate the product for high frequencies . If so , it follows that the linear term will determine the power-law behavior , and that the power-law behavior will be the same as for the current-based model . A key conclusion from the present work is that the power-law predictions from our models are in close agreement with experimental findings for the soma potential and the soma current provided the transmembrane current sources are assumed to be ( i ) homogeneously distributed throughout the whole neuron , ( ii ) uncorrelated , and ( iii ) have a pink ( ) noise distribution . It should be stressed that we do not argue against synaptic noise being a major component underlying neural noise spectra; the importance of synaptic inputs in setting the noise level has been clearly demonstrated , for example by the large difference in membrane potential fluctuation between in vivo and in vitro preparations [41] , [43] . We rather suggest that the power-law behavior seen at the high-frequency end of these noise spectra may be dominated by intrinsic channel noise , not synaptic noise . We also speculate that potassium channels with inherent noisy current with PSDs following a distribution in the relevant frequency range , underlie the observed high-frequency power laws , and the slow voltage- and calcium-activated BK channel , reported to have a very large channel conductance [47] , is suggested as a main contributor [17] . If future experiments indeed confirm that the BK channel is a dominant source of membrane noise , this may have direct implication of the understanding several pathologies . Not only has the BK channel been implicated as a source of increased neural excitability [54] and epilepsy [55] , but also disorders such as schizophrenia [56] , autism and mental retardation [57] have been linked to the BK channel through a decrease in its expression [58] .
|
The common observation of power laws in nature and society , that is , quantities or probabilities that follow distributions , has for long intrigued scientists . In the brain , power laws in the power spectral density ( PSD ) have been reported in electrophysiological recordings , both at the microscopic ( single-neuron recordings ) and macroscopic ( EEG ) levels . We here demonstrate a possible origin of such power laws in the basic biophysical properties of neurons , that is , in the standard cable-equation description of neuronal membranes . Taking advantage of the mathematical tractability of the so called ball and stick neuron model , we demonstrate analytically that high-frequency power laws in key experimental neural measures will arise naturally when the noise sources are evenly distributed across the neuronal membrane . Comparison with available data further suggests that the apparent high-frequency power laws observed in experiments may stem from uncorrelated current sources , presumably intrinsic ion channels , which are homogeneously distributed across the neural membranes and themselves exhibit pink ( ) noise distributions . The significance of this finding goes beyond neuroscience as it demonstrates how power laws power-law exponents α may arise from a simple , linear physics equation .
|
[
"Abstract",
"Introduction",
"Models",
"Results",
"Discussion"
] |
[
"physics",
"computational",
"neuroscience",
"single",
"neuron",
"function",
"biophysics",
"theory",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"biophysics",
"neuroscience",
"biophysical",
"simulations"
] |
2014
|
Power Laws from Linear Neuronal Cable Theory: Power Spectral Densities of the Soma Potential, Soma Membrane Current and Single-Neuron Contribution to the EEG
|
Newborn granule cells become functionally integrated into the synaptic circuitry of the adult dentate gyrus after a morphological and electrophysiological maturation process . The molecular mechanisms by which immature neurons and the neurites extending from them find their appropriate position and target area remain largely unknown . Here we show that single-cell–specific knockdown of cyclin-dependent kinase 5 ( cdk5 ) activity in newborn cells using a retrovirus-based strategy leads to aberrant growth of dendritic processes , which is associated with an altered migration pattern of newborn cells . Even though spine formation and maturation are reduced in cdk5-deficient cells , aberrant dendrites form ectopic synapses onto hilar neurons . These observations identify cdk5 to be critically involved in the maturation and dendrite extension of newborn neurons in the course of adult neurogenesis . The data presented here also suggest a mechanistic dissociation between accurate dendritic targeting and subsequent synapse formation .
New granule cells are added into the dentate circuitry throughout life , as neurogenic neural stem and progenitor cells ( NPCs ) persist in the adult hippocampus [1] . Hippocampal NPCs give rise to only one neuronal subtype , excitatory dentate granule cells , in contrast to the other neurogenic area of the adult brain , the subventricular zone ( SVZ ) /olfactory bulb , where several subtypes of inhibitory neurons are born [2] . Dentate granule cells are highly polarized , glutamatergic neurons that receive their main excitatory input onto dendrites extending into the molecular layer ( ML ) and that send out axons targeting pyramidal cells in area CA3 , as well as inhibitory basket cells , interneurons , and excitatory mossy cells [3–5] . Despite growing knowledge regarding the cellular and molecular mechanisms of fate choice instruction of NPCs in vivo [6–9] and the description of several developmental steps during the integration , selection , and maturation process of adult-generated neurons [10–12] , the regulatory genes required for neuronal maturation and neurite pathfinding of newborn granule cells remain largely unknown . It seems plausible that there is a certain degree of conservation between the molecular pathways used during embryonic and early postnatal development , and the integration of new granule cells in the adult brain [9 , 13 , 14] . However , an obvious difference between embryonic and early postnatal development in one case and the integration of newborn granule cells into the preexisting dentate circuitry in the other is that adult neurogenesis is a heterogeneous and dynamic process , with cells in all stages of maturation at any given time [15] . For example , previously suggested models of axonal growth that predict a concerted repulsion by chemorepellents expressed at a certain time postnatally are difficult to apply to the mechanisms of pathfinding and integration of new neurons in the mature hippocampus [16–18] . Given that alterations in hippocampal neurogenesis might be key components in hippocampus-associated neurological diseases , such as major depression [19] , Alzheimer disease [20] , and epilepsy [21] , understanding the molecular mechanisms underlying neuronal migration , neurite extension , and pathfinding of newborn neurons would seem important to gain further insights into neurological disease . By analyzing quantitative trait loci derived from recombinant inbred mice [22] , we have previously identified several gene loci that may harbor genes critically involved in the process of adult neurogenesis . One of the candidate loci was a region on mouse chromosome 5 containing the cyclin-dependent kinase 5 ( cdk5 ) gene . Cdk5 is a highly versatile kinase that requires association with its regulatory partner , p35 , for activation [23] , and plays a pivotal role in a variety of neurobiological processes , such as neuronal migration , neurite extension , dendritic pathfinding , homeostatic synaptic plasticity , neuronal degeneration , dopamine signaling , and learning and memory [24–34] . Here , we used a retrovirus-based approach for a cell-type–specific knockdown of cdk5 activity in newly generated granule cells born in the adult hippocampus . We show that cdk5 is critically involved in migration , dendritic pathfinding , and neuronal maturation of newborn granule cells . Thus , our findings identify a molecular pathway that governs the accurate spatial integration of newborn neurons in the adult dentate gyrus .
We first analyzed the expression profile of cdk5 in the course of neuronal differentiation of NPCs isolated from rat hippocampus and mouse whole brain . Consistent with a specific role for cdk5 in neuronal cells , we found that , in proliferating rat and mouse NPCs , cdk5 mRNA and protein were expressed at low levels but were robustly up-regulated upon induction of neuronal differentiation ( Figure S1A and unpublished data ) . To analyze the functional involvement of cdk5 in NPC proliferation and/or neuronal differentiation , we generated stably transduced lines of adult mouse- and rat-derived NPCs overexpressing the wild-type cdk5 protein or a well-characterized kinase-deficient cdk5 mutant that acts as a dominant-negative ( DNcdk5 , [23] ) . Overexpression of cdk5 ( either alone or with its coactivator , p35; [23] ) or inhibition with DNcdk5 had no significant effects ( p > 0 . 2 ) on the proliferation of adult rat NPCs ( Figure 1A–1E ) or mouse NPCs ( unpublished data ) , as measured by the number of bromo-deoxy-uridine ( BrdU ) -incorporating cells after a 1-h pulse . Likewise , gain- or loss-of-function of cdk5 activity did not substantially alter the neuronal differentiation capacity in rat ( Figures 1F and 2A–2C ) or mouse NPCs ( unpublished data ) , because the number of MAP2ab-labeled neuronal cells was not significantly different between control cells and cells expressing cdk5 or DNcdk5 ( p > 0 . 08 ) after induction of neuronal differentiation . Since we had previously shown that hippocampal astrocytes instruct NPCs to adopt a neuronal fate [8] , we sought to confirm these findings in cocultures of rat NPCs with hippocampal astrocytes . Cells overexpressing cdk5 or DNcdk5 were indistinguishable from control cells , and they acquired a neuronal morphology ( Figure 2D–2F ) . Thus , we conclude that cdk5 function is not critical for in vitro NPC proliferation or differentiation . Based on these observations , we reasoned that cdk5 might function in advanced stages of neuronal maturation and integration , which might be revealed only when NPCs are embedded in their physiological niche . We therefore examined the role of cdk5 in newborn neurons within the adult dentate gyrus . Cdk5 mRNA is expressed throughout the hippocampus , with strong expression in the adult dentate area ( Figure S1B ) . We overexpressed cdk5 or DNcdk5 in newborn cells using a retroviral strategy that delivers a transgene , along with green fluorescent protein ( GFP ) as a label , to dividing NPCs and their progeny [5 , 35] . Four weeks after transduction with control virus expressing GFP alone , newborn cells consistently showed a highly polarized morphology , with a single apical dendrite branching in the outer parts of the dentate granule cell layer ( GCL ) and extending through the ML ( Figure 3A ) . Overexpression of cdk5 or its coactivator p35 ( unpublished data ) did not substantially change neuronal morphology 4 wk after virus injection ( Figure 3B and 3D ) . Importantly , newborn control cells or cdk5-overexpressing cells never extended dendritic processes towards the hilus . In striking contrast , 51 . 3 ± 9 . 8% of newborn neurons over-expressing DNcdk5 , and thus inhibiting cdk5 kinase activity , lost the polarity typical of dentate granule cells , and extended dendrites along the GCL or even towards the hilus when analyzed 4 wk after viral transduction ( Figure 3C and 3E ) . Confirming overexpression of retrovirus-expressed genes , we visualized both cdk5 and DNcdk5 ( recognized by the cdk5 antibody ) using immunohistochemistry . Cdk5 ( Figure 3D ) and DNcdk5 ( Figure 3E ) showed high expression levels in cell somata and dendritic processes extending from newborn cells 4 wk after viral injection , highlighting the robust and cell-type–specific expression of the transgenes in newborn cells . To confirm the effects of decreasing cdk5 , we phenocopied the effect of functional cdk5 deficiency on the dendritic morphology of newborn granule cells by small interfering RNA ( siRNA ) -mediated reduction in cdk5 levels , using a retrovirally expressed small hairpin RNA ( shRNA ) directed to the cdk5 message ( Figure S2 ) . Despite their dramatic morphological abnormality , newborn neurons overexpressing DNcdk5 expressed the granule cell-specific marker , Prox-1 , and the Ca2+-binding protein of mature granule cells , calbindin , as did CAG-GFP–labeled control cells ( Figure 4 ) , indicating that no change in neuronal fate had occurred . Morphological changes induced by DNcdk5 overexpression occurred early during the neurogenic process: as early as 7 d after viral transduction , the orientation of dendritic processes was altered ( Figure 5A ) . Given the morphological alterations of newborn granule cells in vivo , we next analyzed whether cdk5 might regulate not only dendritic targeting , but also the complexity of dendrites . We traced control , cdk5-overexpressing , and DNcdk5-overexpressing cells and measured the total dendritic length and number of branching points 4 wk after virus injections . Since the penetrance of the DNcdk5 phenotype ( i . e . , extending aberrantly targeted dendrites ) was approximately 50% , we grouped DNcdk5-expressing newborn granule cells into cells with ML-targeted dendrites ( DNcdk5 ) and cells whose dendrites failed to extend into the ML ( BasalDNcdk5 ) . Cdk5 inhibition with DNcdk5 reduced the number of dendritic branching points and total dendritic length 4 wk after retroviral injection irrespective of the position of dendrites within the dentate area , whereas retroviral overexpression of cdk5 had no effect on dendritic architecture ( Figure 5B and 5C ) . Dendritic length of newborn cells expressing DNcdk5 was already reduced 2 wk after virus injection , suggesting that impaired growth , rather than enhanced dendritic pruning , is the underlying cause for dendrite length reduction ( Con 232 ± 63 . 1 μm , DNcdk5 117 . 9 ± 39 . 3 μm , and BasalDNcdk5 108 . 0 ± 46 . 4 μm; p < 0 . 05 ) . These data show that cdk5 is critical for the development of proper dendritic architecture . Following up on previous reports analyzing the function of cdk5 in dendritic maturation and spine formation in vitro or during embryonic and early postnatal development [36–38] , we next asked whether cdk5 was not only involved in dendritic targeting and complexity , but might also be critical for spine formation and maturation . Therefore , we analyzed the density and shape of spines from control cells and cdk5-overexpressing or DNcdk5-overexpressing cells . Again , DNcdk5-expressing newborn granule cells were grouped into cells with ML-targeted dendrites ( DNcdk5 ) and cells whose dendrites failed to extend into the ML ( BasalDNcdk5 ) . We found no difference in spine density or number of mushroom spines between control and cdk5-overexpressing cells ( Figure 6A and 6B ) . The number of spines extending from correctly targeted DNcdk5-expressing cells was also not different from control numbers . However , the density of spines on dendrites extending from DNcdk5-overexpressing cells that failed to extend dendrites into the ML ( BasalDNcdk5 ) was reduced compared to control cells ( Figure 6 ) . Interestingly , both correctly and aberrantly targeted dendrites from DNcdk5-expressing cells showed reduced numbers of mushroom spines compared to newborn cells labeled with a control virus . Despite aberrant dendritic targeting , dendrites of DNcdk5-expressing newborn granule cells extending into the hilus were decorated with dendritic spines ( Figures 6 and 7A ) . The existence of spines suggested that incorrectly targeted dendrites became synaptically integrated into the dentate circuitry despite their aberrant localization . In support of this notion , we consistently found close apposition of the presynaptic protein , synapsin , with spines of dendrites extending from DNcdk5-expressing cells that were located in the hilus , similar to dendritic spines in the ML extending from control cells and cdk5-overexpressing cells ( Figure S3 ) . To obtain independent evidence of synapse formation , we analyzed spines of DNcdk5-expressing newborn granule cells extending aberrant dendrites at the ultrastructural level [3] . We used serial section electron microscopy and identified spines from DNcdk5-expressing hilar dendrites that appeared to form bona fide synapses ( Figure 7B and 7C ) . Cdk5 inhibition thus led to aberrant neurite extension and ectopic synapse formation . We next analyzed the migration pattern of newborn granule cells with attenuated cdk5 activity . In contrast to the SVZ of the lateral ventricles , from where newborn neurons migrate via the rostral migratory stream into the olfactory bulb , there is no clear evidence for a migrational route of newborn granule cells in the dentate gyrus . Previous studies have shown that the vast majority of adult-generated cells stay within the inner third of the GCL [39 , 40] . To analyze the position of newborn granule cells in more detail , we measured the position of newborn cells in relation to neighboring granule cells 4 wk after viral injection ( Figure 8A and 8B ) . Newborn cells labeled with a control virus migrated on average 3 . 83 ± 0 . 66 μm deep into the GCL ( Figure 8B and 8C ) . Strikingly , newborn granule cells expressing DNcdk5 and extending aberrant dendrites were consistently located below the GCL ( −3 . 60 ± 0 . 74 μm , p < 0 . 01 ) . In contrast , the fraction of DNcdk5-expressing cells that had dendrites growing toward the ML showed some migration into the GCL ( 0 . 28 ± 0 . 36 μm ) , suggesting that impaired migration is associated with aberrant neurite extension ( Figure 8B and 8C ) . It has been shown previously that the survival of newborn granule cells in the adult dentate gyrus is an activity-dependent selection process [10] . Given this finding , we next asked if the survival of newborn granule cells deficient in cdk5-kinase activity is altered . We analyzed the number of retrovirally labeled newborn granule cells at several time points after intrahippocampal coinjections of either the DNcdk5-overexpressing retrovirus or the GFP-expressing control virus , together with another control retrovirus encoding the red fluorescent protein ( RFP ) . This strategy allowed us to analyze the dynamics in the number of newborn cells , largely irrespective of variations in virus titer and injection sites . We found that , 2 and 4 wk postinjection , the number of cdk5 activity-deficient neurons was not affected ( p > 0 . 2 and p > 0 . 09 , respectively ) , and only a small decrease in surviving DNcdk5-expressing granule cells with aberrant dendrites appeared 8 wk after viral transduction ( p < 0 . 05 , Figure 9A ) . However , even 1 y after virus injection , aberrant DNcdk5-overexpressing granule cells still existed in the dentate area ( Figure 9B ) , suggesting that , despite a moderately increased elimination rate of DNcdk5-overexpressing granule cells , a substantial portion of DNcdk5-expressing cells survived for extended periods of time . To determine whether the positioning or aberrant growth of dendrites extending from DNcdk5-expressing cells predicted neuronal survival , we again measured the relative position into the GCL of newborn cells compared to neighboring cells . Newborn cells labeled with a control virus were significantly deeper in the GCL at 8 wk compared to their position 4 wk postinjection ( 5 . 76 ± 0 . 96 μm , p < 0 . 05 ) . Newborn granule cells expressing DNcdk5 and extending aberrant dendrites remained below the GCL ( −2 . 91 ± 0 . 59 μm ) .
We here used a cell-type–specific , retrovirus-based strategy to characterize the function of cdk5 in the context of adult hippocampal neurogenesis . A substantial fraction of cdk5-deficient newborn granule cells failed to extend dendritic processes towards the appropriate target zone of the ML and instead formed aberrant synaptic connections in the hilar area . The findings presented here identify a critical role for cdk5 in the migration and morphological maturation of newborn granule cells within the adult dentate gyrus and suggest a mechanistic dissociation between dendritic targeting and subsequent synaptic integration . Previous work has established that newborn granule cells become synaptically and functionally integrated into the dentate circuitry [5 , 41 , 42] . Approximately 7-d-old newborn neurons start to extend dendritic processes toward the ML , where their dendrites ramify [4] . Four weeks after birth , newborn neurons show the highly polarized morphology typical of granule cells [4] . GABA-mediated depolarization , reelin signaling , Disc1 activity , and neurotrophic factors , such as brain-derived neurotrophic factor ( BDNF ) and their downstream signals , have been implicated in neurite extension and migrational behavior of newborn granule cells [13 , 43–46] . However , the signals that regulate the correct positional extension of newly formed neurites are unknown . We here show that a substantial fraction of newborn granule cells deficient in cdk5 activity ( using dominant-negative and shRNA-mediated knockdown of cdk5 ) fail to send dendrites toward the ML , but extend neurites into the hilar region or along the GCL , suggesting a failure of the dendrites of newborn cells to find their appropriate target region in the ML . This finding is consistent with earlier observations that cdk5 is required for proper dendritic development of pyramidal neurons during embryonic cortical development [34] . Furthermore , genetic deletion of p35 , which is a neuron-specific regulatory subunit of cdk5 required for cdk5 activity [23] , leads to the formation of aberrant dendrites extending from granule cells born during embryonic and early postnatal development [47–49] . These previously described phenotypes of cdk5/p35 ablation support our results indicating that cdk5 is critically involved in proper dendrite extension from granule cells born in the adult dentate gyrus . Interestingly , status epilepticus ( SE ) , which is the most commonly used rodent model for temporal lobe epilepsy , induces , among other effects on neurogenesis , the formation of basal dendrites that become integrated into the dentate circuitry , which is not observed under physiological conditions [50–52] . The abnormal extension of dendrites after SE indicates a vulnerable phase for the initiation of aberrant growth of neurites extending from granule cells born in the adult hippocampus . Nevertheless , it remains unknown whether the effects of SE on newborn neurons are cell autonomous or due rather to seizure-associated changes in the dentate area niche . Our results presented here show that aberrant dendrites can develop with inhibition of cdk5 activity in a normal wild-type niche because only newborn cells are genetically modified by retroviral vectors . These results also point to a cell-autonomous effect of cdk5 on neuronal maturation . Notably , not only the position , but also the length , of dendrites extending from cdk5-deficient newborn granule cells was altered . In principle , there are two possibilities to explain this observation: ( 1 ) reduced dendritic growth is secondary to the abnormal position of dendrites , or ( 2 ) attenuation of cdk5 activity itself impairs dendritic growth directly . Our finding that dendrite length was reduced not only in hilar dendrites , but also in dendritic processes that extended correctly from DNcdk5-expressing granule cells supports previous reports that cdk5 is involved in BDNF-stimulated dendritic growth [37] , a mechanism that might also regulate dendrite extension in the context of adult neurogenesis [53 , 54] . Dendritic length of cdk5-deficient cells was not only reduced at 4 wk , but also at 2 wk after virus injection , suggesting impaired growth and not altered dendritic pruning as the underlying cause for reduced dendritic complexity . In addition , future studies will have to address whether altered presynaptic input in response to cdk5-deficiency might contribute to the dendritic phenotype observed in DNcdk5-expressing cells . Strikingly , aberrant dendrites extending from DNcdk5-expressing granule cells were covered with dendritic spines , which are the common sites for glutamatergic synapses within the adult brain . Our results showing spine extension and synapse formation using light and electron microscopy indicate synaptic integration of aberrant dendrites into the preexisting circuitry . Formation of the first dendritic spines occurs approximately 16 d after newborn granule cells are born [4] . Interestingly , the tips of filopodia , thin processes that presumably become stabilized and transform into spines , are found in close vicinity to preexisting axonal boutons that already synapsed on other granule cells more frequently than random chance would predict [3 , 55] . Furthermore , axodendritic synapses of newborn granule cells seem to synapse preferentially onto preexisting boutons , suggesting a competitive mechanism of synapse formation and integration within the ML [3] . At this time , it remains unclear whether ectopic synapses formed on aberrant cdk5-deficient dendrites undergo similar steps as nascent synapses under normal conditions . The data presented here provide evidence of a functional dissociation between dendritic targeting and subsequent synapse formation . The finding that hilar dendrites extending from DNcdk5-expressing granule cells form synapses suggests that dendritic filopodia/spines might be sufficient to induce axodendritic synapse formation within the adult brain . However , it is also possible that ectopic synapses mature in a similar way to newborn granule cells , with correctly targeted dendrites forming , at least transiently , multiple synapse boutons [3] . Interestingly , the number of spines was only reduced in aberrant dendrites , whereas spine density in dendrites extending from DNcdk5-expressing granule cells that targeted the ML were not reduced compared to controls . This finding suggests that cdk5 inhibition does not , per se , reduce the number of spines . However , cdk5 inhibition impaired maturation of dendritic spines into mushroom spines in aberrant as well as in correctly targeted dendrites , indicating an involvement of cdk5 in spine maturation . Notably , overexpression of cdk5 itself affected neither spine number nor shape of newborn granule cells . Together with previous reports that showed that hyperactivation of cdk5 by p25 enhances spine formation in CA1 pyramidal neurons [56] , our findings show that cdk5 overexpression alone is not sufficient to increase the number of spines in newborn granule cells in the adult dentate gyrus but that cdk5 is required , irrespective of dendritic localization for spine maturation . What might be the underlying cause of the aberrant extension of dendrites arising from DNcdk5-expressing granule cells ? In accordance with previous reports [39 , 43] , we found that newborn granule cells migrated short distances into the GCL under control conditions . Neuronal migration was impaired in DNcdk5-expressing granule cells that were extending aberrant dendrites compared to newborn DNcdk5-expressing cells with correctly targeted dendrites , indicating that impaired migration was associated with aberrant dendritic growth . There is no “hard measure” as to when migration of a newborn cell has “failed” ( as there was also a small population of control cells that did not migrate into the GCL ) . This is clearly different from dendritic orientation , which is completely consistent in the group of newborn granule cells of the control groups: all of them extend a dendrite toward the ML . Therefore , the functional association between failed migration and aberrant dendrite extension remains unknown . Notably , cdk5 has a well-characterized function during neuronal migration . One interesting target protein of cdk5 in the context of adult neurogenesis is the microtubule-associated protein , doublecortin ( DCX , [57 , 58] ) . Since DCX is transiently expressed in newborn granule cells [59 , 60] and we found that DCX and the cdk5-coactivator p35 coimmunoprecipitate in lysates of the adult hippocampus ( unpublished data ) , the morphological changes caused by cdk5 inhibition might be associated with reduced phosphorylation of DCX . Alternatively , pathfinding of apical dendrites from newborn granule cells could be impaired by cdk5 inhibition . Indeed , previous reports have shown that cdk5 is involved in the semaphorin-3A–dependent extension of apical dendrites of glutamatergic principal neurons during cortical development [30 , 61] . Future studies will have to determine the downstream targets and upstream signals leading to altered migration and aberrant dendrite formation of DNcdk5-expressing , newborn granule cells . This search is complicated by the fact that the effects of cdk5 inhibition on migration and dendritic development in the context of adult hippocampal neurogenesis can best be adequately tested in vivo , within the context of the dentate area niche , as we found no substantial effects of cdk5 inhibition on cell proliferation and neuronal differentiation when we used NPCs in vitro . The fact that aberrant dendrites are capable of forming synapses also has potential implications for transplantation-based approaches to treat neurological disease: our results indicate that aberrant neurites receive sufficient synaptic input leading to integration and long-term survival , which appears to require an activity-dependent mechanism in the adult hippocampus [10] . Thus , aberrant integration of transplanted cells that can be observed in subsets of transplanted neurogenic cells [62–65] might attenuate graft function or alter neuronal connectivity within the host brain . Interestingly , our experiments indicate that migration of newborn cells into the GCL is not a prerequisite for neuronal survival , as surviving DNcdk5-expressing cells consistently remained below the GCL even 8 wk after viral labeling . The data presented here demonstrate a pivotal role for cdk5 in the neuronal maturation process of newborn granule cells , improving our understanding of molecular mechanisms that govern developmental steps from dividing NPCs to fully integrated newborn granule cells in the course of adult hippocampal neurogenesis .
The in situ probe was isolated from a mouse whole-brain cDNA library using PCR primers against the 3′-UTR of cdk5 ( corresponding to bp 1 , 066–1 , 667 ) and subcloned into pCRII vector ( Invitrogen ) . CMV-driven expression constructs for cdk5 , DNcdk5 , and p35 , and fusion construct cdk5-GFP were gifts from L . H . Tsai ( Massachusetts Institute of Technology , Cambridge , Massachusetts ) . Expression constructs for cdk5 , DNcdk5 , and p35 were subcloned into a retroviral vector driving the transgene under the chicken β-actin ( CAG ) promoter and containing an IRES-GFP ( gift from I . Verma , Salk Institute , La Jolla , California ) . The shRNA against cdk5 was cloned into pAmbion Silencer 2 . 0 for in vitro experiments . The target sequence was GCTGTACTCCACGTCCATC . For in vivo retroviral shRNA experiments , the shRNA was cloned into a vector containing an U6-promoter multiple cloning site and CAG promoter-driven GFP ( gift from V . M . Sandler , Boston , Massachusetts ) . Control viruses were CAG-driven GFP and CAG-driven RFP viruses ( RFP construct was a gift from R . Y . Tsien , University of California San Diego , La Jolla , California ) . Retroviruses were produced as described earlier [4] . Titers ranged between 1–5 × 107 colony-forming units ( cfu ) /ml . The rat and mouse NPCs used in this study have been described earlier [66] . For reverse transcriptase ( RT ) -PCR experiments , RNA was isolated using RNeasy ( Qiagen ) and cDNA was generated using the Superscript system ( Invitrogen ) . Primer sequences for cdk5 , p35 , and GAPDH are available upon request . To produce stable transduced cell lines expressing control virus , cdk5-expressing , and DNcdk5-expressing virus , NPCs were transduced with the respective retrovirus and fluorescence activated cell sorting ( FACS ) was performed 4 d after transduction to yield pure virus-transduced cell populations . Electroporation of NPCs with the p35 expression construct was performed as described earlier [6] . Electroporated cells were identified by coelectroporation with a CAG-RFP construct . Proliferation assays were performed with the addition of 1 μM bromo-deoxy-uridine ( BrdU , Sigma-Aldrich ) 1 h before fixation in FGF-2 ( 20 ng/ml ) . For differentiation studies , 1 μM retinoic acid ( RA ) and 5 μM forskolin ( FSK ) were added to the culture medium after the withdrawal of the mitogens . Cells were fixed 4 d later . The coculture experiments were performed as described before [8] , plating approximately 2 × 103 control , cdk5-expressing , or DNcdk5-expressing cells per square centimeter on a confluent astrocytic layer under serum-free conditions . Cells were fixed 12 d after plating . To quantify proliferation and differentiation , 500–1 , 000 cells were analyzed for labeling with BrdU or the neuronal marker MAP2ab . The investigator was blinded for the experimental conditions , and every experiment was performed with at least three biological replicates . Tissue and cells were fixed and processed for immunostaining as described earlier [41] . Primary antibodies used were rat α-BrdU ( Harlan Seralab ) , mouse α-MAP2ab ( Sigma ) , rabbit α-GFP ( Molecular Probes ) , chicken α-GFP ( Aves ) , rabbit α-Prox-1 ( Chemicon ) , mouse α-Calbindin ( Swant ) , rabbit α-synapsin ( Calbiochem ) , mouse α-NeuN ( Chemicon ) , rabbit α-cdk5 ( Santa Cruz ) , rabbit α-p35 ( Santa Cruz ) , and goat α-DCX ( Santa Cruz ) . Secondary antibodies were obtained from Jackson Laboratory . Protein was isolated from 293T cells for siRNA testing 48 h after transfection of the respective constructs with Lipofectamine 2000 ( Invitrogen ) , protein from rat NPCs was isolated under proliferating conditions and at the times indicated after addition of RA and FSK using RIPA buffer supplemented with protease inhibitors ( Roche Complete EDTA-free supplemented with 0 . 2 mM 4- ( 2-aminoethyl ) benzenesulfonyl fluoride hydrochloride ) . Proteins were separated by electrophoresis on 12% polyacrylamide gels , transferred to PVDF or nylon membranes , and probed overnight with rabbit α-Cdk5 ( Santa Cruz Biotechnology ) or rabbit α-p35 ( Santa Cruz Biotechnology ) antibody together with a mouse α-GAPDH ( HyTest ) antibody . α-Rabbit HRP-conjugated secondary antibodies ( Pierce and Jackson ) and α-mouse AP-conjugated ( Promega ) antibody were used at 1:5 , 000 dilution . Bands were detected by enhanced chemiluminescence . In situ hybridization was performed as described earlier [67] using 601-bp-long sense and antisense 35S-labeled riboprobes against cdk5 . Sections were counterstained with bisbenzimide ( Sigma ) . Sense probes detected no signal beyond unspecific background . All animal procedures were performed in accordance with the protocols approved by the Animal Care Use Committee of the Salk Institute for Biological Studies . All mice in this study were 8- to 10-wk-old female C57Bl/6 mice . Mice were stereotactically injected with 1 μl of the cdk5- or DNcdk5-expressing viruses into the right dentate gyrus; the left side was always injected with a GFP-expressing control virus . Coordinates used were −2 a/p , ±1 . 5 m/l , from Bregma and −2 . 3 d/v from skull . For all experimental time points ( 1 wk , 4 wk , and 12 mo ) , the group sizes were n > 6 . For survival studies , 0 . 75 μl of RFP-expressing control virus was coinjected with 0 . 75 μl of control GFP- , cdk5-expressing , or DNcdk5-expressing virus . Animals were killed 2 , 4 and 8 wk later ( n = 5 per group ) . We counted all virus-labeled cells using 1-in-6 series and formed a ratio between GFP- and RFP-expressing cells , allowing the estimation of the relative survival levels [10] . For shRNA experiments , six mice were injected with shRNA-expressing retrovirus into the right dentate gyrus , and control virus ( with an empty U6-cassette ) into the left dentate gyrus . Animals were killed 4 wk after viral injection . To analyze the position of newborn granule cells in relation to their neighboring cells , confocal images were acquired at the maximum extension of the soma of GFP-labeled cells . From these images , the center of the DAPI-labeled nucleus of the two adjacent cells to the GFP-labeled neuron were connected and a perpendicular line was drawn from the center of the DAPI-labeled nucleus of the GFP-expressing cells ( see scheme in Figure 8A ) . Distances were measured using ImageJ . Analyses were performed 4 wk and 8 wk after viral injection . DNcdk5-expressing cells were grouped into cells that extended an apical dendrite toward the ML ( DNcdk5 ) and cells that extended dendritic processes toward the hilus ( BasalDNcdk5 ) . For each condition , 15 randomly picked cells per mouse ( n = 4 ) were analyzed . Analyses of the dendritic structure were performed as described earlier [4] . Seven to 16 cells per experimental group were analyzed 2 and 4 wk after viral injection . The number and shape of spines were analyzed using at least six dendritic segments per animal from five mice per group 4 wk after retroviral injections , as previously described [4] . Mice were transcardially perfused with 4% paraformaldehyde in 0 . 1 M phosphate buffer ( PB ) ( pH 7 . 4 ) at room temperature for 10 min . Fifteen hours after the perfusion was stopped , the brain was removed and postfixed for 48 h in 4% paraformaldehyde . Fifty-micrometer horizontal vibratome sections were then cut , cryoprotected in 2% glycerol and 20% DMSO in 0 . 1 M PB for 20 min , and freeze-thawed eight times in liquid nitrogen . After a treatment in 0 . 3% hydrogen peroxide ( 5 × 5 min ) and three washes of 10 min in PB + 0 . 5% bovine serum albumin ( BSA-C , Aurion ) , slices were incubated overnight in the primary antibody ( rabbit α-GFP , Chemicon ) in PB + 0 . 1% BSA-C at 4 °C . After washing in PB + 0 . 1% BSA-C , the sections were incubated for 4 h at room temperature in biotinylated secondary antibody ( goat α-rabbit , Jackson Laboratories ) . To reveal this labeling , slices were incubated for 2 h in avidin biotin peroxidase complex ( ABC Elite , Vector Laboratories ) , followed by 3 , 3′-diaminobenzidine tetrachloride ( Vector Laboratories Kit ) for 10 to 20 min . The sections were then postfixed overnight in 2 . 5% glutaraldehyde , washed in 0 . 1 M PB , postfixed in osmium tetroxide for 1 h , dehydrated , and embedded in epoxy resin . Serial sections were cut at 40-nm thickness and collected on single-slot grids , then contrasted by incubating for 35 min in 5% uranyl acetate solution , followed by 25 min in Reynolds solution . Serial images of the labeled structures were then collected with a digital camera ( MegaView III , SIS ) mounted on a JEOL 100 CXII transmission electron microscope , at a 19 . 000× magnification , at a filament voltage of 80 kV . All statistical analyses were performed using Statview 5 . 0 . 1 . for Mac . For all comparisons , ANOVA was performed followed by the Fisher post hoc test , when appropriate . Differences were considered statistically significant at p < 0 . 05 .
|
Neural stem cells divide and generate new neurons throughout life in the mammalian hippocampus . After a distinct maturation process , newborn neurons become functionally integrated into the preexisting circuitry and appear to participate in hippocampal function , which is critically involved in certain forms of learning and memory . However , the molecular mechanisms by which new neurons find their position and project to their appropriate target area remain largely unknown . We here show that cell-type–specific reduction of cyclin-dependent kinase 5 ( cdk5 ) activity in newborn neurons results in impaired neuronal migration and leads to the extension of incorrectly targeted neuronal processes . Strikingly , ectopic processes extending from newborn cells synaptically integrate , suggesting a dissociation between accurate targeting of processes extending from newborn neurons and subsequent synapse formation , which might have important implications for the restorative use of neural stem cells in neurological disease .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"neuroscience"
] |
2008
|
Cdk5 Regulates Accurate Maturation of Newborn Granule Cells in the Adult Hippocampus
|
The human intestinal microbiota is essential to the health of the host and plays a role in nutrition , development , metabolism , pathogen resistance , and regulation of immune responses . Antibiotics may disrupt these coevolved interactions , leading to acute or chronic disease in some individuals . Our understanding of antibiotic-associated disturbance of the microbiota has been limited by the poor sensitivity , inadequate resolution , and significant cost of current research methods . The use of pyrosequencing technology to generate large numbers of 16S rDNA sequence tags circumvents these limitations and has been shown to reveal previously unexplored aspects of the “rare biosphere . ” We investigated the distal gut bacterial communities of three healthy humans before and after treatment with ciprofloxacin , obtaining more than 7 , 000 full-length rRNA sequences and over 900 , 000 pyrosequencing reads from two hypervariable regions of the rRNA gene . A companion paper in PLoS Genetics ( see Huse et al . , doi: 10 . 1371/journal . pgen . 1000255 ) shows that the taxonomic information obtained with these methods is concordant . Pyrosequencing of the V6 and V3 variable regions identified 3 , 300–5 , 700 taxa that collectively accounted for over 99% of the variable region sequence tags that could be obtained from these samples . Ciprofloxacin treatment influenced the abundance of about a third of the bacterial taxa in the gut , decreasing the taxonomic richness , diversity , and evenness of the community . However , the magnitude of this effect varied among individuals , and some taxa showed interindividual variation in the response to ciprofloxacin . While differences of community composition between individuals were the largest source of variability between samples , we found that two unrelated individuals shared a surprising degree of community similarity . In all three individuals , the taxonomic composition of the community closely resembled its pretreatment state by 4 weeks after the end of treatment , but several taxa failed to recover within 6 months . These pervasive effects of ciprofloxacin on community composition contrast with the reports by participants of normal intestinal function and with prior assumptions of only modest effects of ciprofloxacin on the intestinal microbiota . These observations support the hypothesis of functional redundancy in the human gut microbiota . The rapid return to the pretreatment community composition is indicative of factors promoting community resilience , the nature of which deserves future investigation .
Specialized microbial communities inhabit the skin , mucosal surfaces , and gastrointestinal tract of humans ( and other vertebrates ) from birth until death , with by far the largest populations in the colon [1 , 2] . Humans rely on their native microbiota for nutrition and resistance to colonization by pathogens [3–6]; furthermore , recent discoveries have shown that symbiotic microbes make essential contributions to the development , metabolism , and immune response of the host [7–10] . Co-evolved , beneficial , human–microbe interactions can be altered by many aspects of a modern lifestyle , including urbanization , global travel , and dietary changes [1] , but in particular by antibiotics [11] . The acute effects of antibiotic treatment on the native gut microbiota range from self-limiting “functional” diarrhea to life-threatening pseudomembranous colitis [12 , 13] . The long-term consequences of such perturbations for the human–microbial symbiosis are more difficult to discern , but chronic conditions such as asthma and atopic disease have been associated with childhood antibiotic use and an altered intestinal microbiota [14–16] . Because many chemical transformations in the gut are mediated by specific microbial populations [17] , with implications for cancer [18 , 19] and obesity [20 , 21] , among other conditions [22] , changes in the composition of the gut microbiota could have important but undiscovered health effects . An approximate return to pretreatment conditions often ( but not always ) occurs within days or weeks after cessation of antibiotic treatment , as assessed by subjective judgments of bowel function and characterizations of overall community composition using techniques with low phylogenetic resolution [23–25] . However , the effects of a single course of antibiotics on specific microbial populations in vivo can persist for years [26–28] . Overall , the duration and extent of antibiotic-induced disturbance throughout the intestinal microbiota remains poorly characterized , particularly at the species and strain level where the diversity of the community is greatest [2 , 29] , and we lack the information to compare these disturbances to normal temporal variation in community composition . The diversity and abundance of the human microbiota , and of the gut community in particular , poses a challenge for researchers investigating changes in community composition over time . The laborious cultivation-based techniques that were the mainstay of microbiology for a century have revealed only a minority of the species that inhabit the human colon [30 , 31] . Over the past decade , cultivation-independent molecular techniques , particularly those based on the small subunit ribosomal RNA ( 16S rRNA ) gene , have given us a broader and less biased view of the gut microbiota [32 , 33] . Full-length 16S rRNA sequences offer the highest possible degree of taxonomic resolution using this gene , but the cost of dideoxy Sanger sequencing limits our ability to survey the less-abundant members of this diverse community . Studies comparing changes in the gut microbiota over time or between treatments have often used alternative molecular techniques that are more rapid and less expensive than sequencing , but that offer lower taxonomic resolution and reveal only the more abundant members of the community [23–25 , 34–37] . These approaches can be focused more narrowly on particular taxonomic groups , which facilitates the investigation of less-abundant taxa , but at the expense of a broader view of community composition that might reveal important microbial interactions . Pyrosequencing ameliorates some of these constraints by generating a much larger amount of genetic sequence data at a lower cost [38] . Both with this approach and with the established approach of clone library sequencing , DNA is extracted from a sample , and PCR primers complementary to conserved regions of the 16S rRNA are used to amplify the intervening variable sequence . The diversity and relative abundance of 16S rRNA sequence variants in the pool of amplicons is analyzed as a proxy for the diversity and relative abundance of the microbial populations in the sample . Because the gene sequence-based recognition of uncultivated microbial populations is not equivalent to traditional taxonomic classification , terms such as “species” or “strain” are not appropriate . Instead , the populations inferred to exist on the basis of sequence data are referred to as operational taxonomic units ( OTUs ) , which can be defined in various ways and at different levels of resolution . Both the clone library and pyrosequencing approaches can be affected by biased PCR amplification of microbial populations in the sample , although the problem is reduced for the shorter pyrosequencing amplicons [39] . The pyrosequencing approach also avoids the cloning bias of the earlier technique by attaching amplicons individually to beads and physically separating the beads into picoliter-scale wells on a specialized plate . Less phylogenetic information is available from a single pyrosequencing read ( at most , ∼230 informative bases with recent technology ) than from near full-length 16S rRNA gene sequence ( ∼1 , 400 bases with commonly used primers ) , but reads that span particular variable regions of the gene are highly informative [40 , 41] . On the other hand , well over 200 , 000 short 16S rRNA sequence reads , which we refer to as tags , can be obtained in a single run of the Genome Sequencer FLX System . In comparison , two runs of a state-of-the-art capillary Sanger sequencer are required to obtain at most 384 full-length 16S rRNA sequences . Sogin and his co-workers have demonstrated the power of pyrosequencing by amplifying the V6 variable region of 16S rRNA from marine deep water and hydrothermal vent samples , analyzing over 900 , 000 bacterial and archaeal tags from the vent system and revealing greater taxonomic richness ( over 20 , 000 OTUs observed at 3% sequence divergence ) than has previously been reported for any microbial habitat [39 , 42 , 43] . Roesch et al . followed a similar strategy using the V9 region to analyze North American soil samples [40 , 44] , and a group led by Knight have focused on a longer tag that includes the V2 region in macaque gut samples and in a mixture of five diverse microbial habitats [45 , 46] . Computer simulations by Liu et al . [43] , Wang et al . [41] , and Sundquist et al . [47] have indicated that pyrosequencing tags from different regions of the gene will vary in their utility for the distinct tasks of revealing microbial diversity and performing taxonomic classification , both of which contribute to making informative comparisons between complex microbial communities . As with the established approach of generating data from clone libraries , the actual performance of tag pyrosequencing for the goals of a particular study will depend on the region of the 16S rRNA gene analyzed , the PCR primers used , and the composition of the microbial communities in the samples . In the present study , we used pyrosequencing tags spanning the V6 and V3 regions as well as full-length 16S rRNA sequences to analyze the bacterial community composition in a time series of stool samples obtained from three healthy adults before , during , and after a short course of the antibiotic ciprofloxacin ( Cp ) . A companion paper [48] reports a high degree of similarity in the community composition inferred from the two variable regions and from full-length sequences . Here we describe the diversity of human intestinal bacteria more completely and with more precision than has previously been possible . Cp has an extensive and individualized effect on the composition of these communities , but most members of the community returned to their pretreatment abundance within weeks .
Healthy adults who had not taken any antibiotics within the previous year were recruited to donate stool samples before , during , and after a short course of ciprofloxacin ( Cp ) ( 500 mg twice a day for 5 d ) , typical of the treatment prescribed , e . g . , for an uncomplicated urinary tract infection . Samples of approximately 5 g were collected in sterile plastic containers by the participants themselves and immediately stored in home freezers until brought to the laboratory ( within several days ) for storage at −80 °C . From a larger number of samples collected , we chose five time points spanning an 8-mo interval from each of three participants for analyses using full-length 16S rRNA clone libraries and pyrosequencing of the V6 variable region of 16S rRNA . An additional three samples from one of the participants were analyzed along with the original 15 samples using V3 tag pyrosequencing . Participant characteristics and the sampling and analysis regime are shown in Table 1 . For each participant , samples 1 and 2 ( 1 , 2a , 2b , and 2c for individual A ) were collected prior to Cp treatment , sample 3 ( 3a and 3b for individual A ) was the Cp-associated sample taken during or immediately after treatment , and samples 4 and 5 were more distant post-treatment samples . Informed consent was obtained from participants , and the study protocol was approved by the Administrative Panel for Medical Research on Human Subjects ( Institutional Review Board ) of Stanford University . A subsample of approximately 200 mg of frozen stool was added to a 2 . 0-ml screwcap vial containing glass beads of 1 mm , 0 . 5 mm , and 0 . 1 mm diameter ( BioSpec Products ) , and kept on ice until the addition of 1 . 4-ml ASL buffer from the QIAamp DNA Stool Mini Kit ( Qiagen ) . Samples were immediately subjected to beadbeating ( 45 s , setting 4 ) using a FastPrep machine ( Bio 101 ) prior to the initial incubation for heat and chemical lysis at 95 °C for 7 min . Subsequent steps of the DNA extraction followed the QIAamp kit protocol . DNA from two separate extractions of each sample were pooled for molecular analysis . Extraction controls lacking fecal material but otherwise treated identically were carried through the entire procedure including PCR amplification ( no bands visible ) and attempted cloning of PCR amplicons , which was unsuccessful . To amplify near full-length 16S rRNA sequences , 2 ml of extracted DNA served as the template in 20-μl reactions containing 1 unit AmpliTaq polymerase ( Applied Biosystems ) , 1x Buffer II , 1 . 5mM MgCl2 , 20mM tetramethylammonium chloride , 5% DMSO , 0 . 02% Triton X-100 , 100 nM of each dNTP , and 20 nM each of forward and reverse primers ( Integrated DNA Technologies ) . The forward primer was a mixture of 90% bacterial primer 8F ( 5′-AGAGTTTGATCMTGGCTCAG-3′ ) and 10% 8F-Bif targeting Bifidobacteria ( 5′-AGGGTTCGATTCTGGCTCAG-3′ ) ; it was paired with the three domain reverse primer 1391R ( 5′-GACGGGCGGTGTGTRCA-3′ ) . Thermocycling involved a 5-min denaturation at 95 °C followed by 15–22 cycles of 94 °C ( 30 s ) , 55 °C ( 30 s ) , and 72 °C ( 90 s ) , with a final extension at 72 °C for 8 min . The fewest possible cycles were used to produce a faint band of the expected size under UV illumination after electrophoresis of 5-μl reaction product in a 1% agarose minigel containing 1 mM ethidium bromide . Two 4-cycle , 50-μl reconditioning PCR reactions were performed per sample to eliminate heteroduplexes [49] , with 5-μl aliquots of the initial PCR product mixture as the template and other PCR conditions unchanged . Products of the two reconditioning PCR reactions per sample were combined , purified using QIAquick PCR purification columns ( Qiagen ) , and sent to the J . Craig Venter Institute ( Rockville , MD ) for cloning and automated bidirectional dideoxy sequencing . Forward and reverse sequencing reads were assembled using the Phred/Phrap/Consed suite of programs with default parameters [50–52]; assembled sequences were aligned via the NAST algorithm [53] at the Greengenes Website ( http://greengenes . lbl . gov/ ) [54] . Aligned sequences were checked for chimeras using version 2 of the Greengenes implementation of Bellerophon , with 99% similarity to a trusted sequence over 1 , 250 nucleotides as the threshold to bypass checking , and rejecting both putative and subthreshold chimeras regardless of the divergence ratio . Of 10 , 062 assembled sequences , 203 did not result in a NAST alignment that included both the V3 and V6 regions , and 2 , 651 were excluded as possible chimeras , leaving 7 , 208 full-length sequences for analysis . Using the less-stringent default settings in Bellerophon to identify chimeras ( divergence ratio > 1 . 1 ) would have excluded 766 fewer sequences from the analysis , but the exclusion of these sequences did not significantly affect the inferred taxonomic composition of the community ( Text S1 ) . The taxonomic affiliation of full-length sequences was determined using the Classifier tool of the RDP , with an 80% bootstrap threshold [41] . The NAST-aligned full-length sequences were imported into ARB [55] and a genetic distance matrix calculated with the Olsen distance correction using the Hugenholtz version of the Lane mask ( included with Greengenes ARB database ) to exclude regions of questionable alignment; 1 , 241 columns were retained . The distance matrix was used with a 0 . 01 genetic distance , furthest-neighbor threshold in DOTUR version 1 . 53 [56] to designate 1 , 295 operational taxonomic units ( OTU0 . 01 ) . The full-length sequences have been deposited in GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/; accession numbers: EU761594–EU768801 ) . Amplicon libraries for 16S rRNA V6 region pyrosequencing were generated as recommended by 454 Life Sciences , using the same DNA extracts as for full-length sequencing , with 50 ng of template DNA ( determined by a NanoDrop 1000 spectrophotometer , NanoDrop Technologies ) in 50-μl , 30-cycle reactions . The PCR reagents and thermocycling parameters were those suggested in the protocol , except that the annealing temperature was reduced from 57 °C to 55 °C . Each PAGE-purified fusion primer ( Integrated DNA Technologies ) consisted of the 454 platform A or B linker/primer sequence , a trinucleotide key ( or barcode ) that was unique for each sample , and sequence complementary to a conserved region flanking the V6 variable region of the 16S rRNA gene . The forward primer was an equimolar mixture of two nondegenerate oligonucleotides 5'-BxxxCAACGCGAAGAACCTTACC-3′ and 5'-BxxxATACGCGAGGAACCTTACC-3′ , where B represents the B linker ( 5′-GCCTTGCCAGCCCGCTCAG-3′ ) , xxx represents the sample identification key , and the remaining nucleotides correspond to positions 967–985 of the 16S rRNA gene ( Escherichia coli numbering ) . The degenerate reverse primer was 5'-AxxxCGACARCCATGCASCACCT-3′ , including the A linker ( 5′-GCCTCCCTCGCGCCATCAG-3′ ) , sample key , and nucleotides corresponding to E . coli positions 1 , 064–1 , 046 . Two or three separate amplification reactions for each sample were pooled and purified as recommended using Ampure beads ( Agencourt ) . The amplicon length and concentration in the libraries was estimated using the BioAnalyzer microfluidics device ( Agilent ) , and an equimolar mix of all 15 V6 amplicon libraries was used to prepare both A-linked and B-linked pyrosequencing beads ( i . e . , for bidirectional reads of the amplicons ) via emulsion PCR using the recommended kit and protocol ( 454 Life Sciences ) . An equal mixture of A-linked and B-linked beads was loaded onto both regions of a Picotiter plate ( estimated 825 , 000 beads/region ) . Pyrosequencing on the Genome Sequencer FLX System ( Roche ) at the Stanford Genome Technology Center resulted in 490 , 881 reads which passed the length and quality criteria of the machine software . Preparation of amplicon libraries for the V3 region used the same DNA extracts as for the full-length and V6 sequences , except that new DNA extracts were prepared for samples 2c and 3b from individual A , and additional samples 2a , 2b , and 3a from individual A were included , as described above and in Table 1 . V3 amplicon libraries were prepared as for the V6 region , using a single 50-μl PCR reaction per sample and with the following PAGE-purified primers ( Integrated DNA Technologies ) : 5'-BxxxxxACTCCTACGGGAGGCAGCAG-3′ ( B linker , pentamer sample identification key , forward primer at E . coli positions 338–357 ) and 5'-AxxxxxTTACCGCGGCTGCTGGCAC-3′ ( A linker , pentamer key , reverse primer at E . coli positions 533–515 ) . Preparation of A-linked and B-linked beads and pyrosequencing were carried out as described above , with an estimated 900 , 000 beads loaded in each region of the Picotiter plate , which resulted in 593 , 088 pyrosequencing reads passing machine quality filters . Pyrosequencing reads of both the V6 and V3 regions were passed through additional quality filters to reduce the overall error rate [43] . Any reads containing one or more ambiguous nucleotides and reads shorter than 50 nucleotides were discarded . The expected sample keys and primer sequences were trimmed from the proximal and distal ends of the reads , and those lacking an expected sample key and primer sequence at either end were examined further . In many of these cases , a rare variant of the expected sequence was identified and trimmed from the read . The origin of at least some of these rare primer variants as oligonucleotide synthesis errors was suggested by the frequent appearance of several copies of a particular rare variant in a single sample , while no examples of that variant appeared in any other sample . Because unique primers ( differing in the sample key and synthesized separately ) were used to generate the amplicon libraries from each sample , a synthesis error would be likely to affect only a single sample . Reads lacking a recognizable primer sequence ( including rare variants ) at either end , reads covering only a portion of the variable region , and reads that could not be unambiguously assigned to a sample were discarded; in addition , 18 reads from the V3 region with identity or near identity to eukaryotic rRNA sequences were discarded . The resulting datasets contained 441 , 894 V6 tags and 490 , 699 V3 tags , which were the basis of all subsequent analyses . The V6 and V3 reference databases ( V6 RefDB , V3 RefDB ) are composed of publically available , high quality , full-length 16S rRNA sequences from Silva release 92 ( http://www . arb-silva . de/ ) [57] with taxonomic classifications obtained from the RDP Classifier [41] ( minimum 80% bootstrap score ) as described in [48] . The V6 region is defined as E . coli positions 986-1045 corresponding to the sequence between primers 967F and 1046R . The V3 region is defined as E . coli positions 358–514 , corresponding to the sequence between primers 338F and 515R . Pyrosequencing tags were assigned the taxonomic classification of the most similar reference sequence or sequences in the V6RefDB or V3RefDB , based on a multiple sequence alignment . In cases where a tag was equidistant to multiple reference sequences , the tag was classified to the level of the most resolved taxon shared by at least two-thirds of the reference sequences nearest that tag . For example , a tag with exact matches to multiple reference sequences would be resolved to the genus level only if at least two-thirds of the sequences containing that tag were classified in the same genus with an 80% bootstrap score by the RDP . A reference sequence-based operational taxonomic unit ( refOTU ) was defined as containing all query tags sharing the same RefDB sequence ( or group of sequences ) as their nearest neighbor . The vast majority of human gut pyrosequencing tags obtained in the current study ( unlike marine tags [39] ) have an exact or near match in the corresponding RefDB ( Figure 1 ) , which simplifies the classification of tags and the definition of refOTUs for this habitat . In this study , we will refer to a specific refOTU as V3_Taxon refOTU_X , where Taxon is the most-resolved taxon name assigned to the refOTU , and X is the rank order of the refOTU within Taxon based on the total normalized abundance data from this study . The comparison of nearest database distance to tag abundance , the rank abundance plots , the rarefaction curves , and the diversity summary statistics were calculated using actual tag counts without normalizing the number of tags obtained per sample . Rarefaction curves for full-length sequence OTU0 . 01s were calculated on a per-tag basis using DOTUR version 1 . 53 [58]; rarefaction curves for variable region refOTUs were calculated per sample using EstimateS version 7 . 5 [59] . The Shannon diversity index H = –Σ pi ln ( pi ) and Shannon equitability index EH = H/ln ( S ) ( where pi is the proportion of the ith OTU and S is the total number of OTUs ) were calculated using spreadsheet software . The diversity summary statistics for Cp-associated samples were compared to the lower tail of values for other samples , assuming that summary statistics would be distributed normally when calculated for samples drawn repeatedly from an unchanging community . Comparisons of V3 refOTU abundance between samples were made after correcting for the unequal number of tags obtained from each sample . All tag abundance values within a sample were multiplied by the required factor so that total normalized tag abundance for all samples was 43 , 405 , which was the number of V3 tags obtained from sample C5 , the maximum for any sample . Normalization factors for the other samples ranged from 1 . 072 to 2 . 087 . Principal component analysis of normalized refOTU abundance used the prcomp function of the stats package of the R statistical language , version 2 . 2 . 1 ( http://www . r-project . org/ ) . Statistical assessment of variation in the abundance of individual refOTUs between participants or with respect to Cp was conducted with Edge software , version 1 . 1 . 291 [59] , applying both a 95% confidence level for individual taxa to declare significance , and a 10% false discovery rate ( FDR ) criterion for all taxa declared significant in a given comparison ( additional details in Text S1 ) .
From 15 stool samples ( five each from individuals A , B , and C ) , we obtained 7 , 208 near full-length 16S rRNA sequences by traditional PCR amplification , cloning , and capillary dideoxy sequencing . From the same set of 15 samples , we obtained 441 , 894 V6 pyrosequencing tags ( average length 59 ) , and from the same samples plus three additional samples from individual A ( 18 samples total ) , we obtained 490 , 699 V3 pyrosequencing tags ( average length 145 ) ( Table 2 ) . The number of unique tags or sequences revealed by each method followed the same rank order as the total number of tags or sequences ( V3 > V6 > full-length ) . However , the numerical disparity between pyrosequencing and capillary sequencing was smaller for unique tags or sequences than for all tags or sequences , reflecting the competing trends of higher throughput for tag pyrosequencing but greater resolution for full-length sequences . The V6 and V3 tags were grouped into 3 , 316 and 5 , 671 refOTUs , respectively , via comparison to reference databases ( see Methods ) ; 1 , 295 OTUs were identified among full-length sequences using a 1% genetic distance threshold after masking sequence positions of uncertain alignment ( designated as OTU0 . 01 ) . The number of unique genera or other most-resolved taxa above the genus level ( see Methods ) identified among sequences or tags was much lower , ranging from 56–130 , confirming conclusions based on earlier molecular studies that the diversity of human colonic bacteria is concentrated at the species and strain level [2 , 29] . The abundant pyrosequencing tags derived from these three participants were much more likely than were the rare tags to have been seen before , i . e . , to correspond perfectly to one or more of the full-length 16S rRNA sequences already available in the public databases . Even though only 9% of the unique V6 tags and 7% of the unique V3 tags had exact matches in the reference databases , these represented 90% of all V6 tags and 85% of all V3 tags , because nearly all the abundant tags had exact matches . Of the 100 most abundant V6 and V3 tags ( representing 75% and 80% of all tags , respectively ) , 97 V6 tags and 94 V3 tags had perfect matches in the reference databases . The abundant gut bacteria from these participants were less well represented by cultivated strains; only about half of the 100 most abundant tags ( 51 V6 tags and 55 V3 tags ) had perfect matches to 16S rRNA sequences derived from bacterial strains in culture ( Dataset S1 ) . The recently reported average accuracy per base of better than 99 . 75% for pyrosequencing of the V6 region [43] ( after applying additional filters to eliminate error-prone reads ) is lower than that of capillary dideoxy sequencing , although considerably better than the 96% accuracy initially reported [38] . Undoubtedly , some fraction of the unique tags in our data arose from pyrosequencing errors , but individual error products were rare relative to the frequency of the correctly sequenced tag from which they were derived . The most abundant V6 and V3 tags in our data ( exactly matching Bacteroides dorei sequences ) were observed 31 , 788 and 41 , 153 times , respectively , across all three participants . Assuming a uniform error rate per base of 0 . 25% over 59 or 145 nucleotides , most tags are expected to be error-free , and most error-containing tags are expected to have only a single error . There were 177 unique V6 tags in our dataset that differed from the B . dorei V6 tag at a single nucleotide position and that lacked an exact match to any sequence in the reference database . The most abundant of these potential error products occurred only 60 times , 0 . 19% the frequency of the error-free B . dorei V6 tag . Similarly , of 270 unique V3 tags differing from the B . dorei V3 tag at only one position and lacking a perfect database match , the most abundant occurred 367 times . However , we consider this to be a rare , genuine tag , because the distribution of this tag among samples differed significantly from that of the B . dorei tag , in contrast to the pattern expected for a pyrosequencing error product ( following paragraph; also Text S1 and Dataset S2 ) . The most abundant potential V3 error product derived from the B . dorei sequence occurred 226 times , 0 . 55% the frequency of the error-free tag . Sogin et al . suggested that using a reference database of public full-length 16S rRNA sequences to define taxa ( i . e . , refOTUs ) for pyrosequencing tags would result in a conservative estimate of taxonomic richness [39] . This procedure will also minimize the influence of pyrosequencing errors on comparisons of community composition , since such errors could never result in the definition of a novel ( but artificial ) taxon . Instead , most pyrosequencing error products will be counted appropriately as a member of the refOTU containing the nearly identical error-free tag from which it was derived . However , this OTU definition inherently limits the detectable taxonomic diversity to the sequences already present in the public databases , and it obscures genuine biodiversity whenever tags derived from distinct organisms are grouped into a single refOTU . We investigated this diversity-masking effect for the most abundant V6 and V3 taxa by calculating , for each sample separately , the abundance ratios of individual rare tags to the most common tag belonging to the same refOTU ( Text S1 and Dataset S2 ) . For a given rare tag , significant variation in this ratio across samples is not expected if the rare tag is a pyrosequencing error product , because the likelihood of a particular error is independent of the origin of the tag . On this basis , four of the 42 rare V6 tags and 33 of the 87 rare V3 tags that we investigated were not likely to be pyrosequencing error products derived from the most common tag in the same refOTU ( G test , p < 0 . 05 ) . Even though this approach cannot differentiate all error-free , rare tags from pyrosequencing error products , it is clear that analyzing pyrosequencing tags on the basis of refOTUs will obscure some genuine biodiversity . Nonetheless , we chose to accept this cost in order to ensure that pyrosequencing errors did not inflate the reported taxonomic richness , nor influence comparisons between communities . Because the vast majority of abundant tags had exact matches in the reference databases , the loss of taxonomic resolution in the current study would have affected mostly a subset of the rare taxa . The rank abundance curves for OTUs derived from V6 and V3 tags and full-length sequences ( Figure 2 ) show the extent to which tag pyrosequencing facilitated our exploration of the rare biosphere of the human gut . There was substantial agreement between the methods in the estimated relative abundance of specific taxa ranging from the phylum to genus level [48] . At a finer scale of resolution , the rank abundance curves had similar shapes for the V6 and V3 refOTUs and full-length OTU0 . 01s , with a small number of dominant taxa and a long tail of less abundant taxa ( Figure 2A ) . Tag pyrosequencing identified many more OTUs than were identified by a large , traditional 16S rRNA survey , but of perhaps greater importance , there are many OTUs for which pyrosequencing provided a more precise estimate of relative abundance , and an improved confidence of detection . For example , 869 of the 1 , 295 OTU0 . 01s detected among the cloned full-length sequences were singletons , represented by only a single sequence . The best estimate for the relative abundance of each singleton OTU0 . 01 in the pooled community is 1/7 , 208 or 1 . 4 × 10−4 , but the 95% confidence interval for this estimate ( assuming a Poisson distribution of detection events ) ranges over almost 2 orders of magnitude , 7 . 3 × 10−6–6 . 6 × 10−4 . There is only a 63% chance for a specific taxon at 1 . 4 × 10−4 relative abundance to be detected in a single sample of 7 , 208 clones , so the presence ( or absence ) of these singleton OTUs does not provide a robust basis for comparisons between samples . A refOTU that appeared 65 times among the V6 or V3 tags would have had the same estimated relative abundance as a singleton OTU0 . 01 in the clone libraries , but with a much narrower 95% confidence interval of 1 . 0 × 10−4–1 . 7 × 10−4 . There were 1 , 122 refOTUs in the pooled V6 libraries which occurred often enough ( ≥5 times ) to estimate that their probability of detection was at least 99%; the estimated relative abundance for these taxa ranged from 7 . 4 × 10−2 to 1 . 1 × 10−5 . The pooled V3 libraries had 1 , 759 refOTUs with at least 99% probability of detection , with relative abundance ranging over almost 4 orders of magnitude , from 9 . 3 × 10−2 to 1 . 0 × 10−5 Because the depth of pyrosequencing permits a high confidence of detection for many moderately rare and rare taxa , their presence or absence is informative for comparisons between samples . In contrast , in the pooled clone libraries , there were only 199 OTU0 . 01s ( relative abundance ranging from 4 . 8 × 10−2 to 6 . 8 × 10−4 ) , with at least a 99% estimated probability of detection . Rank abundance curves for the V6 and V3 taxa within each individual are shown in Figure 2B and 2C . A range of 1 , 673–1 , 745 V6 refOTUs and 2 , 592–3 , 319 V3 refOTUs were found per individual , providing minimal estimates of the number of bacterial strains present in one person over an 8 month interval . Considering all the sampling times together , between 539–552 V6 refOTUs and 680–810 V3 refOTUs were found in each individual at 99% confidence of detection or above , compared to 66–98 full-length OTU0 . 01 per individual . Rarefaction curves can be used to assess the degree of completion of a taxonomic survey , i . e . , how closely the observed taxonomic richness approaches the endpoint of all the taxa detectable with a particular method for that set of samples [60] . Figure 3 shows that the OTU0 . 01 rarefaction curve rises more steeply than the curves for V6 or V3 refOTUs , which reflects the greater taxonomic resolution of this typical OTU definition applied to full-length sequences , compared to the refOTU definitions for pyrosequencing tags . On the other hand , although this set of 7 , 208 sequences is among the largest datasets of full-length 16S rRNA sequences from the human microbiota ( or any environment ) , the rarefaction curves for V6 and V3 tag pyrosequencing eventually rise higher and display more curvature toward the horizontal than the OTU0 . 01 curve . These features show that a single run of the FLX sequencer targeting V6 or V3 tags from the human gut microbiota can reveal more taxa , and capture a larger proportion of the detectable taxa , than a more extensive effort directed toward full-length 16S rRNA clone sequencing . However , the tag-based surveys are by no means complete , as shown by the large number of singleton refOTUs in both the V6 and V3 data ( 1 , 361/3 , 316 and 2 , 306/5 , 671 , respectively ) . Nonparametric estimators of total refOTU richness [61–63] were 60–70% higher than the observed richness , in the range of 5 , 300–5 , 700 refOTUs for the V6 tags and 8 , 875–9 , 650 refOTUs for the V3 tags . These large discrepancies between observed and estimated richness , and the 20–35% increase in estimated richness over the last half of sampling ( Figure S1 ) indicate undersampling , a situation in which nonparametric estimators underestimate true richness [60 , 63] . The biological significance of the many rare , unobserved taxa implied by these estimates is unknown . While rarefaction and richness estimators consider survey completion from the perspective of identified and unidentified taxa , coverage considers completion from the perspective of individual tags or sequences . Good's coverage [64] , the estimated likelihood that a tag or sequence chosen from the sample at random will belong to an OTU that has already been identified in the dataset , was 99 . 2% and 99 . 5% for V6 and V3 data , respectively , in contrast to 88% for the full-length sequences . In other words , we expect that more than 100 additional tags would need to be sequenced in order to detect a new refOTU for either variable region , while on average less than ten new full-length sequences would suffice to detect a new OTU0 . 01 . The heatmap in Figure 4A shows the relative abundance across 18 samples of 1 , 450 V3 refOTUs with normalized total abundance of at least ten . ( Because normalization factors are as high as two for some samples , this corresponds to at least five observed tags , the threshold for 99% probability of detection . ) All the statistical comparisons between samples that follow ( e . g . , PCA analysis and the identification of taxa that vary in response to Cp treatment ) use the normalized abundance of these 1 , 450 taxa . Normalized V3 refOTU abundance within the three most abundant genera , Bacteroides , Faecalibacter , and Roseburia , are shown in greater detail in Figure 4B–4D , and the complete refOTU by sample abundance matrix is provided as Dataset S3 . The most striking features of Figure 4 are consistent with our current understanding of the human distal gut microbiota [1 , 2 , 29 , 30]: ( 1 ) Most members of the community belong to a small number of genera within the Bacteroidetes and Firmicutes phyla , with members of Proteobacteria and Actinobacteria occurring much less frequently . ( 2 ) The diversity of genera and subgeneric taxa is greater in the Firmicutes than Bacteroidetes . ( 3 ) The same genera tend to be abundant or rare across all participants and all time points during unperturbed states , but at a finer taxonomic scale , the gut microbiota is individualized—some refOTUs are abundant in only one or two individuals . Additional observations can be made from Figure 4 and Dataset S3 that are also consistent with previous data , although perhaps not as widely appreciated: ( 4 ) Temporal variability that is not associated with Cp treatment existed within each individual , ( especially individual A ) although it was not as evident as interindividual variation . ( 5 ) A total of ten phyla were found in the gut microbiota of these participants ( six–nine per individual ) , some of which were discovered only recently . The fifth most abundant phylum was Verrucomicrobia ( described in 1997 ) [65] , within which almost all tags were associated with the genus Akkermansia ( described in 2004 ) [66] . Akkermansia was of comparable abundance to the entire family Enterobacteriaceae including E . coli , the first described species of colonic bacteria . ( 6 ) In 2008 , abundant , uncharacterized bacterial taxa are still found in the human gut . The sixth most abundant refOTU was V3_Clostridiales_refOTU_2 , accounting for 3 . 5% of all V3 tags . The dominant tag in this refOTU exactly matched 203 cloned sequences in the V3 RefDB , but no sequences from cultivated organisms . Of these 203 sequences , 29 are shorter than 1 , 200 nucleotides or possibly chimeric; the remaining 174 sequences form a clade to the exclusion of any cultivated organisms . Pairwise genetic distances within this clade averaged 0 . 9% , while distances from sequences within the clade to the most similar sequences from cultivated strains averaged 10% , suggesting the presence of a novel species divergent from cultivated bacteria at roughly the family level ( Dataset S4 and Figure S2 ) . A dramatic effect of Cp on many taxa can also be seen in Figure 4 , although the effect is apparently stronger in individuals A and B than individual C . Ecological diversity statistics confirmed this impression , as shown in Figure 5 . Cp treatment significantly decreased taxonomic richness in the Cp-associated samples in individuals A and B ( p < 10−4 and p < 0 . 005 , respectively ) but not C ( p = 0 . 24 ) . Cp treatment decreased the Shannon diversity index H ( p < 10−5 in all cases ) and Shannon equitability index EH ( p < 10−5 for A and B , p < 0 . 05 for C ) in all three individuals , although the magnitude of the effect was smaller for C . The magnitude of the Cp effect on diversity is best conveyed by transforming the Shannon index to an effective number of species [67 , 68] , which is reduced 82% , 63% , and 36% in the Cp-associated samples of A , B , and C , respectively ( Figure S3 ) . None of the indices showed a detectable effect of Cp by 4 wk after treatment . Principal component analysis ( PCA ) of the log-transformed relative abundance of V3 refOTUs ( Figure 6 ) confirmed that the primary sources of variability between samples were inter-individual differences and Cp treatment . Differences between individual B and the others were captured largely by the first PCA axis , which accounted for 35 . 8% of all variability between samples . The Cp-associated samples separated from non-Cp samples of the same individual primarily along PCA axis 2 , which accounted for 18 . 6% of inter-sample variability . The observation that Cp-associated samples of A and B are closer to each other along axis 1 than the non-Cp samples indicated that some of the abundant taxa that differed between these individuals at other times became less abundant following Cp treatment . The third PCA axis ( 9 . 7% of variability ) appeared to be driven primarily by Cp-independent temporal variability in individual A , with the three samples collected in the one-week interval prior to Cp treatment clustering at one extreme . In the now-familiar context of interindividual variability in the human microbiota [30 , 69–71] , the similarity of the gut communities in individuals A and C was unexpected . None of the study participants are related , nor do they live or work together , and individuals B and C would appear to be the most similar to each other based on host traits such as gender , ethnicity , and geographic origin ( Table 1 ) . Hence , other factors ( e . g . , diet ) that were not investigated in this study may have an important influence on the similarity of the intestinal communities from individuals A and C . Log transformation of taxon relative abundance prior to PCA reduces the influence of abundant taxa relative to that of moderately abundant taxa; it has an impact intermediate between performing no transformation , which allows abundant taxa to dominate inter-sample variability , and scaling the data so that the variability of each taxon among samples is weighted equally . Not surprisingly , the first axes of PCA on the untransformed refOTU data explained a higher proportion of the total variability ( because a few taxa accounted for most of the variation ) ; in this case , the distinction between the Cp-associated and non-Cp samples of individual C was obscured within the temporal variability of non-Cp samples of individual A ( Figure S4A ) . On the other hand , with scaled PCA , the proportion of explained variability on the initial axes was lower , but the Cp-associated samples of both A and C were clearly separated in the same direction along axis 2 from a cluster containing all the non-Cp-associated samples from both individuals ( Figure S4B ) . As with the diversity statistics , PCA ( regardless of the abundance transformation ) did not reveal any Cp effect at 4 wk post-treatment that could be distinguished from other sources of temporal variability . A caveat to interpreting PCA with any weighting scheme is that each refOTU is treated as equally dissimilar to all others . Preliminary results from ordination techniques that considered a range of refOTU relatedness showed a smaller magnitude of Cp-associated variability relative to the magnitude of interindividual differences of community composition ( unpublished data ) . Over a third of the V3 refOTUs with normalized total abundance of ten or more ( 528/1 , 450 ) differed significantly in abundance between participants ( p < 0 . 05 , 2 . 9% FDR ) . The taxa that varied between subjects included 64 of the 100 most abundant refOTUs , i . e . , those representing 0 . 2% or more of all tags . An examination of these abundant taxa demonstrated the importance of maintaining high phylogenetic resolution during the analysis of tag pyrosequencing data ( Dataset S5 ) . The 64 abundant refOTUs that varied between subjects represent 23 different genera or other most-resolved taxa above the genus level . However , when these genera and more inclusive taxa were compared , only 11 of the 23 taxa were found to differ between individuals ( p < 0 . 05 ) . For example , 19 of the 100 most abundant V3 refOTUs belonged to the genus Bacteroides , and 18 of them differed significantly between individuals ( Figure 7 ) . However , the abundance of the Bacteroides genus as a whole did not differ significantly between individuals . We tested taxa for a significant response to Cp by looking for two patterns: either a difference between the relative abundance in Cp-associated samples versus all other samples ( pattern 1 ) , or a difference between the relative abundance in pre-Cp samples versus samples collected during and after treatment ( pattern 2 ) . A detectable Cp response in a particular taxon could be a direct effect of Cp activity , or it could be mediated by ecological interactions with other taxa . Because these data represent relative abundance , a taxon that maintains constant abundance in absolute terms as the total community abundance declines will represent an increasing proportion of the community . For comparisons across all individuals we expressed the refOTU abundance in a sample as the deviance of log abundance from the mean log abundance of the refOTU in that individual; this procedure compared changes in relative , not absolute abundance across individuals . Cp effects were found in 30% of refOTUs ( 442/1 , 450 ) when tested over all individuals , more according to pattern 1 ( 428 taxa , p < 0 . 05 , 7 . 7% FDR ) and fewer according to pattern 2 ( 46 taxa , p < 0 . 005 , 10% FDR; 32 taxa are significant according to both patterns ) ( Dataset S6 ) . The predominance of pattern 1 was expected based on the PCA results shown in Figure 6; Cp-associated samples differed from others within each subject , but no separation of pre- and post-Cp samples was evident on the primary PCA axes . A number of the abundant taxa found to vary significantly in response to Cp appeared to have different responses among the participants ( Figure 8 ) , which suggested that testing for a Cp response within each individual might reveal additional taxa that respond to Cp only in some subjects . Testing only within A , for example , revealed that 25% ( 314/1 , 260 ) of refOTUs varied in response to Cp ( pattern 1: 314 taxa significant , p < 0 . 05 , 6 . 9% FDR; pattern 2: 1 taxon significant , p < 104 , 1 . 5% FDR , this taxon was significant by pattern 1 as well ) ( Dataset S7 ) . Slightly less than half of these taxa ( 158/321 ) were found to be Cp responsive when testing across all individuals; the taxa found to be significant both within A and across all individuals were concentrated among the more abundant taxa . With data from only a single Cp-associated sample , we cannot yet conduct a similar comparison within individuals B and C , but the Cp response of the intestinal microbiota appears to be individualized , particularly for less abundant members of the community .
We applied the 16S rRNA tag pyrosequencing strategy of Sogin et al . [39] to characterize bacterial populations in the distal human gut , which constitute the predominant community of the human microbiota and one of the most densely populated microbial habitats on Earth . This approach revealed both the pervasive effects of an antibiotic that is considered to be relatively benign for the gut microbiota , and the resilience of the human gut microbiota following perturbation . As shown with marine , soil , and macaque gut samples [39 , 42 , 45] , massively parallel pyrosequencing permits the exploration of microbial diversity in complex communities to an unprecedented depth . In comparison to clone libraries and traditional dideoxy sequencing , tag pyrosequencing detects more taxa , and provides more-accurate estimates of the relative abundance of a large number of moderate- and low-abundance taxa . Such data will facilitate better ecological studies of the human microbiota , with immediate clinical application . For example , the composition of the microbiota at the start of antibiotic treatment may determine the likelihood of pathogen overgrowth and life-threatening antibiotic-associated diarrhea ( AAD ) [72 , 73]; investigation of this hypothesis will depend on comprehensive characterizations of the community over time , including both rare and abundant taxa . Of the four bacterial pathogens associated with severe AAD , tags matching Klebsiella oxytoca and Clostridium perfringens were detected rarely in the current study ( nine tags in two subjects and one tag , respectively ) ; tags matching Clostridium difficile or Staphylococcus aureus were not found . The rarity of these pathogens and the absence of severe AAD in the current study is not surprising , since only a small number of healthy young adults were involved . It would be of tremendous interest and practical value to conduct a similar study of stool samples collected over time from elderly patients in hospitals or long-term care facilities , which have higher frequencies of both antibiotics use and serious complications from such use . Cp is reported to have a lower rate of common gastrointestinal side effects than some other broad-spectrum antibiotics , and effects on gut microbial diversity ( as detected by a low resolution “fingerprinting” technique ) were less pronounced for Cp than for clindamycin or amoxicillin-clavulanic acid [25] . Nonetheless , the relative abundance levels of about 30% of the taxa in the gut were affected by Cp treatment when the comparison was made across all individuals , and additional taxa were found to respond to Cp within a single participant . Some of these changes may be direct effects of varying sensitivity to the antibiotic among the taxa comprising the gut microbiota . The refOTUs that increased in relative abundance following Cp treatment may represent taxa with intrinsic resistance to the antibiotic , strains that are typically Cp-sensitive but that had already acquired resistance prior to this study , or strains that developed Cp-resistance due to the current exposure . However , many of the changes in the community are likely to be explained by indirect effects , mediated by ecological interactions among taxa such as resource competition , cross-feeding , or the cooperative lysis of polymeric substrates [74 , 75] . Despite a pervasive disturbance of the gut microbiota , gut function remained normal as assessed subjectively by the participants , and the community composition in samples taken 4 wk after treatment were within the range of temporal variability of pretreatment samples . While the current study did not include samples showing the temporal features of the return of the community to its prior state , a pyrosequencing-based investigation of this rapid transition in a complex community is clearly of interest , and is continuing in our laboratory . The apparent continuity of gut function supports the hypothesis that the diversity of the microbial community provides functional redundancy [2] . At the same time , continuity in the predominant metabolic activity of the community , i . e . , hydrolysis and fermentation of polysaccharides , does not necessarily imply the continuity of more specialized activities [7 , 76] such as bile transformation [77] , immune modulation [9 , 78] , or pathogen resistance due to specific inhibition [79 , 80] . Yet , the rapid return of the gut community to its pretreatment state in each individual suggests that not all communities supporting similar functions are equivalent . A quantitative examination of the metabolic transformations and host interactions of the gut microbiota during perturbation will be necessary to assess functional stability . In addition , it will be important to examine the microbiota at different sites within the intestinal tract in order to appreciate possible local antibiotic effects that are not revealed in fecal specimens . Epidemiological studies associating antibiotic-induced changes in the gut microbiota with chronic diseases [15 , 81] suggest that some consequences of community change may not be evident immediately . An investigation into the factors responsible for community resilience is warranted . A mixture of selective forces intrinsic to the community ( e . g . , a competitive hierarchy based on relative growth rates and substrate affinity or interference mechanisms ) and imposed by the environment ( e . g . , composition of the diet and of host-derived substrates ) is likely to be involved , but nonselective forces such as re-colonization of the gut lumen from protected environments ( perhaps the mucosa ) must also be considered . The current study can be compared to that of Young and Schmidt , who investigated changing bacterial populations in stool samples from a single patient with self-resolving AAD subsequent to amoxicillin-clavulanic acid treatment [82] . These investigators noted the complete absence of sequences from Clostridium cluster XIVa on the fourth day of antibiotic treatment ( down from 20% on day 0 ) and a reduction from 33% to 15% of sequences affiliated with the genus Faecalibacterium; these two groups include the majority of butyrate-producing bacteria in the human gut [83] . ( Butyrate is the preferred energy source of colonocytes . ) In contrast , none of the participants in the current study experienced AAD . While Faecalibacterium refOTUs declined in all participants in response to Cp , each individual had other refOTUs with dominant tags exactly matching known butyrate-producing organisms that maintained or increased their abundance during Cp treatment ( e . g . , V3_Roseburia_refOTU_1 in individual B matches Butyrivibrio fibrisolvens and Roseburia intestinalis [84] , V3_Anaerostipes_refOTU_1 , in individuals A and C , matches the unnamed butyrate-producing strains SS2/1 and SSC/2 [85] ) . Although the community as a whole showed a substantial return to the pre-Cp composition within 4 wk of the end of Cp treatment , there were examples of taxa that were affected by Cp and did not recover . V3_Clostridiales_refOTU_23 was the 41st and 52nd most abundant taxon before Cp treatment in individuals A and C , respectively , and present in all pre-Cp samples ( means of 206 and 158 tags per sample ) ; it was completely absent from all samples after the start of Cp treatment . This refOTU includes perfect matches to clones obtained from the human gut and from swine manure lagoons , but none to a cultivated bacterium . The most similar cultivated species are Clostridium piliforme and C . colinum , both zoonotic pathogens . V3_Bilophila_refOTU_1 was the 82nd and 79th most abundant pre-Cp refOTU in individuals A and C ( means of 60 and 91 tags/sample ) ; it was reduced to an average of less than two tags per sample after Cp treatment began . V3_Bilophila_refOTU_1 responded differently in individual B , in whom it was absent from the Cp-associated sample but subsequently rebounded to pre-Cp levels . The dominant tag of this refOTU exactly matches Bilophila wadsworthia , a common gut microbe that is frequently isolated in cases of appendicitis and from intestinal and extraintestinal abscesses [86] . The clinical significance of these persistent Cp-induced changes is unknown , but some gut bacteria are known to have important health effects despite being present at moderate or low abundance . For example , Oxalobacter formigenes is the most important known oxalate degrader in the human gut , an activity associated with a reduced risk for the development of calcium oxalate kidney stones [22] . O . formigenes is found at low abundance in most children , but is absent from many adults , which is generally attributed to its susceptibility to many antibiotics [22] . Tags matching O . formigenes were not detected in the current study . Certain abundant taxa responded differently to Cp in different individuals , as shown in Figure 8 . This result is not surprising when we consider that antibiotic resistance , even high-level Cp resistance resulting from chromosomal mutations [87] , can be acquired far more rapidly than the rate of evolutionary change in the 16S rRNA gene . Although the participants had not taken any antibiotics in the year prior to the study , their earlier exposure to Cp or other antibiotics is unknown . Once established , antibiotic resistant strains can persist in the human gut for years in the absence of further antibiotic use [28] . Another potential explanation for divergent Cp responses of the same refOTU between individuals is the existence of indirect effects; the response may be mediated by ecological linkages that differ between individuals . As shown in the companion paper of Huse et al . [48] , the inferred bacterial community composition in these samples at taxonomic levels from phylum to genus was very similar , whether based on tags from the V6 or V3 variable regions or on full-length 16S rRNA sequences . All three techniques involve PCR amplification , and hence could fail to detect certain phylogenetic groups or lead to skewed levels of tag or sequence relative abundance with respect to the starting material [88 , 89] . However , the similarity of the results obtained from three distinct sets of primers suggests that the PCR bias of these primers ( and the cloning bias for full-length sequences ) for the bacterial taxa present in this habitat is minimal . The interpretation of these results as a measure of the relative abundance of cells in the gut habitat must still be made with caveats regarding differential cell lysis and the unequal numbers of rRNA genes per genome , as for any 16S rRNA-based technique [88 , 89] . Clearly , full-length sequences will provide the highest possible phylogenetic resolution for any genetic locus . However , pyrosequencing tags from either of two carefully chosen , short variable regions of the 16S rRNA gene have sufficient resolution to reveal taxonomic richness that exceeds any previously observed for samples of host-associated microbial communities , even though most diversity in these habitats is concentrated at the species and strain level [2 , 29] . This success occurred despite our use of reference databases to define OTUs , which necessarily constrained the resolution of tag pyrosequencing to the diversity already represented in public 16S rRNA sequence databases . A more accurate assessment of microbial diversity that accounts for the novel 16S rRNA variants discovered by pyrosequencing will depend on defining OTUs with reference to the tags themselves ( e . g . , using tag sequence divergence ) , with appropriate treatment of potential pyrosequencing errors . A high proportion of tags in the current study have an exact match in the database , due to the substantial sequencing effort that has already been directed to the human gut microbiota [7 , 30 , 90] . Hence , the constraint of defining OTUs with respect to a database is not as severe for this habitat as it may be for others [39 , 42] . However , the limitations of full-length 16S rRNA sequence surveys for discovering rare members of the gut biosphere are demonstrated by the contrast between the abundant tags in this study , almost all of which have an exact match in the database , and rare tags , which occur at a range of distances from their nearest database match ( Figure 1 ) . This study has confirmed the existence of more than 5 , 600 bacterial taxa in the human gut , exceeding earlier predictions made on the basis of nonparametric richness estimators and full-length 16S rRNA sequencing studies [29 , 30] , but the presence of singleton OTUs in our data leads to a still larger predicted richness for this habitat . As has long been appreciated , the highly uneven abundance distribution among taxa in this community ( Figure 1 , Dataset S3 ) hampers the identification of rare taxa via traditional sequencing approaches . However , coverage values exceeding 99% for our V6 and V3 data correspond to estimates that fewer than 1% of all detectable 16S rRNA tags in these samples belong to taxa that have not yet been detected . Individually , any of the rare , unobserved taxa in these samples are unlikely to exceed 0 . 001% of all 16S rRNA genes , the relative abundance at which we have roughly 99% confidence of detection . Because so many distinct reference tags are matched exactly by the tags obtained in this study , we conclude that a large number of bacterial taxa previously identified as gut inhabitants in a range of previous 16S rRNA sequencing studies can be present simultaneously or sequentially in a single individual , albeit at widely varying abundance . As suggested in a recent revision to the maxim of Baas Becking , “Everything may be everywhere , but not in equal amounts” [91] . This result challenges microbial ecologists to account for persistent differences in the composition of the microbiota between individuals , either by ongoing selective forces that vary among individuals such as diet and host genotype , or by factors such as founder effects and specific mutualistic interactions that stabilize community composition . However , any explanation will also have to account for the temporal variability shown by some taxa , and the resilience of the community to perturbations involving a significant fraction of its constituent populations .
|
The intestinal microbiota is essential to human health , with effects on nutrition , metabolism , pathogen resistance , and other processes . Antibiotics may disrupt these interactions and cause acute disease , as well as contribute to chronic health problems , although technical challenges have hampered research on this front . Several recent studies have characterized uncultured and complex microbial communities by applying a new , massively parallel technology to obtain hundreds of thousands of sequences of a specific variable region within the small subunit rRNA gene . These shorter sequences provide an indication of diversity . We used this technique to track changes in the intestinal microbiota of three healthy humans before and after treatment with the antibiotic ciprofloxacin , with high sensitivity and resolution , and without sacrificing breadth of coverage . Consistent with previous results , we found that the microbiota of these individuals was similar at the genus level , but interindividual differences were evident at finer scales . Ciprofloxacin reduced the diversity of the intestinal microbiota , with significant effects on about one-third of the bacterial taxa . Despite this pervasive disturbance , the membership of the communities had largely returned to the pretreatment state within 4 weeks .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases",
"ecology",
"microbiology"
] |
2008
|
The Pervasive Effects of an Antibiotic on the Human Gut Microbiota, as Revealed by Deep 16S rRNA Sequencing
|
In natural environments , bacteria often adhere to surfaces where they form complex multicellular communities . Surface adherence is determined by the biochemical composition of the cell envelope . We describe a novel regulatory mechanism by which the bacterium , Caulobacter crescentus , integrates cell cycle and nutritional signals to control development of an adhesive envelope structure known as the holdfast . Specifically , we have discovered a 68-residue protein inhibitor of holdfast development ( HfiA ) that directly targets a conserved glycolipid glycosyltransferase required for holdfast production ( HfsJ ) . Multiple cell cycle regulators associate with the hfiA and hfsJ promoters and control their expression , temporally constraining holdfast development to the late stages of G1 . HfiA further functions as part of a ‘nutritional override’ system that decouples holdfast development from the cell cycle in response to nutritional cues . This control mechanism can limit surface adhesion in nutritionally sub-optimal environments without affecting cell cycle progression . We conclude that post-translational regulation of cell envelope enzymes by small proteins like HfiA may provide a general means to modulate the surface properties of bacterial cells .
The majority of bacteria in the biosphere exist within surface-attached communities [1]–[3] that facilitate metabolic cooperation , sharing of genetic information , and protect cells against stress ( reviewed in [1] ) . Environmental signals including nutrient availability , pH , and ion concentrations influence surface community formation by modulating expression of adhesive cell envelope structures and extracellular polymers that determine surface attachment ( reviewed in [4] ) . The Gram negative bacterium , Caulobacter crescentus , thrives in dilute freshwater ecosystems and has the ability to permanently attach [5] , [6] to a chemically diverse range of surfaces [7]–[10] via a polysaccharide-rich , polar organelle known as the holdfast [9] , [11]–[13] . As organic polymers and ions concentrate on material surfaces in aquatic environments [14] , surface attachment likely provides C . crescentus a nutritional advantage . Given that holdfast surface attachment is permanent , C . crescentus should exhibit tight control over holdfast development to ensure that cells do not become perpetual residents of a poor environment . In this study , we have sought to elucidate the molecular regulatory determinants of holdfast development in C . crescentus . Elaboration of the holdfast adhesin in C . crescentus is cell-cycle-regulated , though it is not requisite for cell-cycle progression [8] , [15]–[17] . The cell cycle yields two cell types that are physiologically , morphologically and functionally distinct ( Figure 1A ) . The flagellated and motile swarmer cell provides this species a means for dispersal; this cell type is arrested in G1 and incapable of replication . In order to initiate growth and replication , the swarmer relinquishes motility and differentiates into a stalked cell . The stalked cell , specialized for nutrient uptake , grows and divides asymmetrically to generate a new swarmer cell upon division [8] , [18] . Development of the holdfast at the cell surface is temporally restricted to the late swarmer cell stage , where it emerges at the nascent stalked cell pole ( [15] , [17] , Figure 1A ) . However , the timing of holdfast emergence within this developmental window can be hastened at the post-translational level by physical contact of the flagellum with surfaces [19] . Once constructed , the holdfast is a permanent feature of the cell surface that is not shed or reassimilated . Premature holdfast development at the nascent swarmer pole prior to cell division would hinder dispersal of newborn swarmer cells . Thus cell-cycle control of holdfast biogenesis helps to ensure appropriate cell dispersal . We have previously observed that a two-component regulatory system composed of the soluble sensor histidine kinase , LovK , and the single domain receiver , LovR , regulates the Caulobacter general stress response [20] and modulates cell adhesion [21] . We sought to understand the mechanism of adhesion control and have discovered a novel inhibitor of holdfast development , hfiA , that is regulated downstream of lovK-lovR . A forward genetic screen for HfiA-insensitive mutants identified suppressing mutations in a glycosyltransferase gene required for holdfast development , which we name hfsJ . We demonstrate a physical interaction between HfiA and HfsJ , and that suppressing mutations in HfsJ attenuate the HfsJ-HfiA interaction . These results support a model in which HfiA inhibits holdfast development via direct interaction with an enzyme required for holdfast biosynthesis . Expression of hfiA is temporally regulated across the cell cycle , and is lowest during the period when the holdfast is elaborated at the cell surface . Multiple Caulobacter developmental regulators , CtrA , GcrA and StaR , physically occupy and control transcription from the hfiA promoter . The coordinate action of these regulators induces hfiA at the end of G1 , thus restricting holdfast formation to the swarmer cell . However , not every cell makes a holdfast; the probability of holdfast emergence at the single cell level depends on the nutritional composition of the growth medium and is inversely correlated with hfiA expression . Our data thus support a model in which holdfast development is controlled by cell cycle and nutritional input signals that are integrated at the promoter of hfiA . As a negative regulator of an enzyme required for holdfast production , HfiA functions as a checkpoint protein that ensures holdfast development occurs within the appropriate cell cycle window and nutritional conditions .
We previously observed that coordinate overexpression of lovK and lovR increases cell-cell adhesion , and deletion of lovK or lovR reduces adhesion [21] . To understand the genetic basis of this adhesion phenotype , we first tested if the holdfast is required for lovK-lovR-enhanced adhesion . We overexpressed lovK and lovR in a strain lacking hfsA , a gene required for holdfast synthesis [13] . In a wild-type background , overexpression of lovK and lovR results in large cell aggregates that are readily visible in the culture ( Figure 1B ) and accumulate in a ring along the culture tube wall ( Figure S1 ) . Strains lacking hfsA do not exhibit enhanced cell-cell adhesion or tube ring formation upon lovK-lovR overexpression . Thus , holdfast is required for the lovK-lovR-enhanced adhesion phenotype . We recently discovered that lovK and lovR are potent negative regulators of the general stress response ( GSR ) sigma factor σT and , concordantly , function as negative regulators of cell survival during osmotic and oxidative stress [20] . As σT controls expression of a number of genes involved in cell envelope function [20] , [22] , we hypothesized that lovK-lovR affects adhesion via σT-dependent modulation of the cell envelope . Our hypothesis predicts that cells lacking sigT should be hyper-adhesive . However , a ΔsigT null strain is not hyper-adhesive . Moreover , coordinate overexpression of lovK-lovR in ΔsigT background results in an equivalent cell adhesion phenotype as a strain expressing lovK-lovR in a wild-type background ( Figures 1B & S1 ) . These data demonstrate that lovK-lovR modulates adhesion independent of sigT , via a mechanism that requires holdfast development . We next sought to identify specific adhesion effector ( s ) regulated by LovK-LovR , independent of σT . We measured change in global transcript abundance upon lovK-lovR overexpression in ΔsigT and in wild-type genetic backgrounds . Only one transcript , cc_0817 ( ccna_00860 ) , exhibited differential steady-state levels in both experiments; showing a 2–3-fold reduction ( Figure 1C ) . A lacZ transcriptional fusion to the cc_0817 promoter confirmed that overexpression of lovK-lovR represses cc_0817 transcription ( Figure 1D ) . cc_0817 mRNA is among the top 10% of transcripts in the C . crescentus cell in terms of abundance , based on analysis of wild-type expression array data [20] . Thus modest fold changes in transcript abundance reflect large changes in absolute number of transcripts . Our finding that lovK-lovR overexpression results in decreased transcription of cc_0817 and increased holdfast-dependent adhesion suggested that cc_0817 functions downstream of lovK-lovR as an inhibitor of holdfast development . To test this hypothesis , we assayed the effect of cc_0817 deletion and overexpression on holdfast development . Holdfast development was monitored by incubating cells with fluorescently-labeled wheat germ agglutinin ( WGA-Alexa595 ) , a lectin that binds N-acetylglucosamine and marks holdfast at the cell surface [9] . In minimal defined medium with xylose as the carbon source , about 3% of wild-type cells display a holdfast ( Figure 2A&B ) . Overexpression of cc_0817 reduces the fraction of cells with a visible holdfast to near zero . lovK-lovR overexpression increases the fraction of cells with a holdfast ∼10-fold and cc_0817 overexpression in this background ( lovK-lovR++ cc_0817++ ) attenuates this effect ( Figure 2A&B ) . Conversely , deletion of cc_0817 results in elaboration of a holdfast on nearly every cell ( Figure 2A&B ) . Expression of cc_0817 from the ectopic xylX locus in a Δcc_0817 null background restores wild-type holdfast levels . These data support a model in which cc_0817 functions to inhibit holdfast development . We have named this gene holdfast inhibitor A , hfiA . hfiA was annotated as a 78 aa hypothetical protein [23]; the central portion of the putative protein contains a hydrophobic stretch of 35 amino acids . A search of the Pfam and Conserved Domain Databases with the primary sequence of HfiA revealed no conserved domains . Given the small predicted size of hfiA and the lack of functional clues in the sequence , we sought to validate the prediction that hfiA is translated into a protein , and to define the length of the predicted open reading frame . We identified two hfiA transcriptional starts by 5′ RACE . Seventy-five percent of the sequenced RACE products started at the third position of the predicted hfiA translational start codon and twenty-five percent started 126 bp upstream of the predicted translational start ( Figure 2C ) . These data suggested the hfiA start codon was annotated incorrectly . The first 14 codons of the annotated hfiA coding sequence include 4 additional NTG codons ( marked b , c , d , and e in Figure 2C ) that could potentially function as translation start sites . To test if putative codons a or b function as translation start sites , we expressed from the hfiA promoter a translational fusion between the first 7 predicted hfiA codons and lacZ . We engineered a second translational fusion that also included codons c , d and e fused to lacZ . Only the second fusion yielded β-galactosidase activity above background ( Figure S2 ) , demonstrating that hfiA is translated and that translation initiates from codons c , d or e . To identify the site of translation initiation we replaced the wild-type hfiA allele with mutant alleles in which codons c , d or e were mutated away from NTG , and thus could no longer function as translation start sites . We predicted that loss of the translation start would phenocopy the hyper-holdfast phenotype of the ΔhfiA null strain . However , no single codon mutant exhibited a null phenotype ( Figure S2 ) , suggesting hfiA translation can initiate at multiple sites . Furthermore , no double codon mutant exhibited a full hyper-holdfast phenotype . Only the strain bearing mutations in all three putative start codons ( c , d and e ) phenocopied the ΔhfiA null strain ( Figure S2 ) . Thus all three of these codons can likely function as sites of translation initiation , resulting in synthesis of 65–68 amino acid proteins ( ∼7 kDa ) . We have reannotated hfiA to reflect initiation at the ATG that was originally predicted to be codon 11 . We hypothesized that the small protein , HfiA , functions to directly inhibit a protein required for holdfast development . To test this hypothesis , we designed an unbiased genetic screen to identify adhesive mutants that continue to produce holdfast when hfiA is overexpressed ( Figure 3A ) . Several classes of adhesive mutants emerged from this screen: a ) Mutants in which hfiA overexpression was disrupted by lesions in the xylose-inducible promoter , in the hfiA coding sequence ( Table S1 & Figure S3 ) , or in the xylose transport system; b ) Mutants with increased surface adhesion but not increased holdfast including lesions in the gene encoding the S-layer protein or in genes involved in synthesis of lipopolysaccharide ( LPS ) , which attaches the S-layer protein to the cell [24]–[26]; c ) Mutants with an elevated number of holdfast-bearing cells in the presence of an intact hfiA overexpression system . We initially isolated two independent hfiA-suppressor strains , 256-39 and 261-15 , with strongly enhanced surface adhesion ( Figure 3C ) and a high fraction of cells bearing a holdfast . Whole genome sequencing of these suppressors revealed multiple mutations relative to the wild-type parent ( Table S1 ) . While each suppressor strain bore unique mutations , both 256-39 and 261-15 shared non-synonymous polymorphisms in gene cc_0095 ( ccna_00094 ) that resulted in a C260R substitution and C260R , W264R substitutions , respectively . We conducted additional enrichment screens and identified three other independent HfiA-supressors ( 256-112 , 256-177 and 256-185 ) that exhibit near wild-type surface adhesion when hfiA is overexpressed ( Figure 3C ) . Targeted sequencing of the cc_0095 locus in these strains revealed that each harbor mutations in the 3′ end of this gene which result in the following coding changes: L248R , a frame-shift after L266 and a duplication of F246-R254 respectively ( Table S1 , Figure 3B ) . The independent isolation of five unique lesions in the same region of cc_0095 strongly implicated this gene in hfiA-mediated control of holdfast development . CC_0095 is annotated as a UDP-N-acetyl-D-mannosaminuronic acid transferase and is related to E . coli WecG ( 29% identity/45% similarity ) and Bacillus subtilis TagA ( 27% identity/46% similarity ) glycosyltransferases . This protein is strongly classified as a WecG/TagA–family glycosyltransferase in the Conserved Domain Database ( E-value<e−84 ) [27] , though the glyco-substrates are difficult to predict from primary sequence . This family of enzymes is widely distributed in Gram-negative and Gram-positive bacteria . Within the Caulobacterales , all sequenced species that encode the holdfast synthesis gene cluster , hfsEFGHCBAD , also encode proteins that are 50–80% identical to CC_0095 . WecG/TagA–family enzymes catalyze the transfer of an activated nucleotide sugar to a glycosylated membrane phospholipid , undecaprenyl pyrophosphate ( Und-PP ) [27] . Execution of this chemistry is a critical early step in the biosynthesis of extracellular sugar polymers including the holdfast material [12] . To test if cc_0095 functions in holdfast development , we generated a strain carrying an in-frame deletion of this gene . Strains lacking cc_0095 do not develop holdfast ( Figure 3D ) and are completely defective in surface adhesion ( Figure 3C ) consistent with the defective biofilm phenotype reported for a cc_0095 transposon insertion mutant [28] . Expression of cc_0095 from an ectopic locus restores holdfast synthesis and surface adhesion to the null mutant ( Figure S4 ) . Neither E . coli wecG nor B . subtilis tagA complements the Δcc_0095 adhesion or holdfast defects , although expression of E . coli WecG alters C . crescentus morphology resulting in cells that are longer , thinner , and no longer curved ( Figure S4 ) . Genes required for holdfast synthesis have been assigned the names hfsA through hfsI [12] . We have named gene cc_0095 , hfsJ . We note that hfsJ prediction is 10 codons shorter in the C . crescentus NA1000 ( CCNA_00094 ) genome annotation compared to strain CB15 ( CC_0095 ) annotation . To experimentally define the start codon , we generated strains in which each predicted start codon was mutated . Only mutation of the predicted start annotated in NA1000 phenocopied the null , supporting the annotation of the shorter open reading frame ( Figure S4 ) ; this was used as the frame of reference for numbering the position of hfsJ mutations . A fluorescent protein fusion , HfsJ-venus , expressed from the native hfsJ promoter complements the holdfast null ΔhfsJ phenotype ( Figure S4 ) . In contrast to holdfast export and anchoring proteins , which are localized to the stalked pole [11] , [29] , HfsJ-venus is distributed throughout the cell ( Figure S4 ) . Western blot analysis on this strain with antibodies to GFP/venus indicates that the HfsJ-venus fusion is not cleaved; no degradation products were detected ( Figure S4 ) . Thus the fluorescence signal observed throughout the cell reflects the distribution of HfsJ-venus . Moreover , HfsJ-venus was detected only in the pellet fraction but not the soluble fraction of the cell lysate ( Figure S4 ) , supporting a model in which this protein is membrane associated . We cannot rule out the possibility that the fluorescent tag either alters the localization of HfsJ , or stabilizes it so that a localization site becomes saturated . The chemistry that HfsJ is predicted to execute ( modification of an inner membrane carrier glycolipid ) occurs in the cytoplasm . Given the rapid two-dimensional diffusion of lipids in the inner membrane , such a lipid-modifying enzyme need not be spatially restricted to produce a product that is utilized by the holdfast synthesis and export machinery located at the nascent stalked cell pole . To test if the hfsJ lesions identified in our genetic screen specifically suppress the holdfast inhibition function of hfiA , we constructed strains in which we replaced wild-type hfsJ with each of the mutant alleles . These hfsJ-mutant strains were transformed with either empty plasmid or the hfiA overexpression plasmid and assayed for surface adhesion and visible holdfast . Each of the hfsJ-mutant strains exhibits surface adhesion in the presence of the empty control plasmid ( Figure 3C , dark bars ) . Thus , the mutations in hfsJ do not compromise bulk adhesion . However , the holdfast in these strains do not stain as intensely as wild-type ( Figure 3D ) suggesting they may be smaller than wild type . While overexpression of hfiA nearly abolishes surface adhesion and holdfast development in wild-type cells ( Figure 3C&D ) , the hfsJ-mutant strains are largely insensitive to the effect of hfiA overexpression on surface adhesion and holdfast formation ( Figure 3C&D ) . To test if HfiA and HfsJ physically interact , we assayed whether the proteins co-purify by serial affinity chromatography . We cloned hfiA and hfsJ into a tandem E . coli expression plasmid , with N-terminal maltose binding protein ( MBP ) and His6 affinity tags , respectively ( Figure 4A ) . MBP-HfiA ( 51 kDa ) and His6-HfsJ ( 33 kDa ) co-eluted from amylose affinity resin ( Figure 4B ) . We observed an additional band , the size of the MBP tag alone ( 42 kDa ) , suggesting that the MBP-HfiA fusion is partially unstable . Eluate from the amylose resin was then bound to Ni2+ resin . Only two proteins of sizes corresponding to His6-HfsJ and MBP-HfiA co-eluted from the Ni2+ resin ( Figure 4B ) . We confirmed the identities of these proteins as His6-HfsJ and MBP-HfiA by mass spectrometry . As a control , we confirmed that MBP-HfiA does not bind to Ni2+ resin , nor does His6-HfsJ bind to amylose resin , nor is the interaction mediated by the MBP domain ( Figure S5 ) . Together these data provide strong support for a direct physical interaction between HfsJ and HfiA . To provide additional support for a physical interaction between HfiA and HfsJ , we preformed a bacterial two-hybrid assay [30] . Co-expression of T25-hfsJ and T18-hfiA fusions results in blue colonies when grown on medium containing X-gal ( Figure 4C ) , and significant β-galactosidase activity when grown in liquid ( Figure 4D ) , demonstrating that HfsJ and HfiA interact and bring together the split T25/T18 adenylyl cyclase domains . Reconstitution of adenylyl cyclase activity required both fusions; neither fusion alone was sufficient . Importantly , none of the HfsJ mutant alleles interact with HfiA sufficiently to yield a positive result in this assay ( Figure 4C&D ) . These data support a model in which HfiA functions to inhibit holdfast development through direct interaction with HfsJ , a putative glycosyltransferase required for holdfast development . Holdfast development is temporally regulated across the cell cycle [8] , [15] , [17] , [19]: the holdfast is elaborated at the flagellated pole of the swarmer cell , before or during the swarmer-to-stalked cell transition ( Figure 1A ) . Profiles of C . crescentus gene expression throughout the cell cycle reveal that transcription of the holdfast inhibitor hfiA is cell-cycle regulated , with a minimum at the period of holdfast development ( Figure 5A ) , [31]–[33] ) . These results are consistent with a model in which cell-cycle regulation of hifA expression determines the developmental window for holdfast biogenesis . C . crescentus cell cycle progression is controlled via dynamic interplay between a number of developmental regulatory proteins ( reviewed by [34]–[36] ) . Three known developmental regulators , CtrA , GcrA , and StaR directly control hfiA expression . These proteins are introduced briefly here: a ) CtrA , an essential response regulator with a DNA-binding output domain [37] , is a ‘master developmental regulator’ that directly or indirectly controls transcription of ∼25% of the C . crescentus cell cycle regulated genes [32] , [37] . b ) GcrA is critical developmental regulator required for efficient growth that forms a feedback control loop with CtrA [38] , [39] . c ) StaR is a non-essential developmental regulator that controls stalk biogenesis [40] . Transcription of these genes is temporally regulated across the cell cycle ( Figure 5A; [31]–[33] ) . StaR , CtrA and GcrA physically occupy the chromosomal region immediately upstream of hfiA ( Figure 5B; Tables S2 , S3 , S4 , S5 , and [41] ) . To test whether these proteins affect hfiA expression , we assayed transcription from a PhfiA-lacZ transcriptional fusion in ctrA , gcrA , or staR mutant backgrounds . Cells lacking ctrA and gcrA are severely compromised and have developmental defects thus , we used temperature sensitive alleles [37] , [42] or depletion strains [38] to evaluate the effects of protein loss on hfiA transcription . At restrictive temperatures , strains bearing ctrA ts alleles have 2-fold less PhfiA-lacZ activity than wild type; PhfiA-lacZ activity is reduced ∼25% upon GcrA depletion ( Figure 5C ) . We conclude that CtrA and GcrA are transcriptional activators of hfiA ( Figure 5G ) . Overexpression of staR from a xylose-inducible promoter reduced PhfiA-lacZ activity by ∼70% , deletion of staR enhanced PhfiA-lacZ activity by ∼20% ( Figure 5C ) . These data provide evidence that StaR represses hfiA transcription . We note that our experiments with unsynchronized populations will mask the amplitude of temporally-restricted transcriptional change . Indeed , endogenous StaR is expected to affect only a subset of cells in the population at any given time . Conversely , overexpression of staR from an inducible promoter affects PhfiA in all of the cells at any given time . Like many genes involved in holdfast biogenesis , transcription of hfsJ is also cell cycle regulated ( Figure 5A; [31]–[33] ) . Sequences corresponding to the hfsJ locus are enriched by immunoprecipitation of CtrA , but not GcrA or StaR ( Figure 5D ) . Transcription from the hfsJ promoter is diminished in a ctrA temperature sensitive ( ctrAts ) mutant ( Figure 5E ) ; we conclude that CtrA is a direct activator of hfsJ transcription ( Figure 5H ) . We identified CtrA binding sites upstream of both hfiA and hfsJ ( CTCttaaAGCTTTCtaaaCCT , 92 bp , p = 1 . 0e-04 and ATActtaGCGGGATttaaCCA , 66 bp , p = 6 . 3e-07 respectively ) . We next asked whether the activity StaR on hfiA transcription affects holdfast development . Both ctrA and gcrA are essential , and mutant strains have pleiotropic defects that confound assessment of holdfast development . The function of StaR was initially investigated in a holdfast deficient genetic background; thus no holdfast phenotype was reported [40] . As StaR is a repressor of hfiA , we predicted that overexpression of StaR should result in an increase in holdfast development . Indeed , staR overexpression results in a dramatic enhancement of visible holdfast ( Figure 5F ) . LovR is a single domain response regulator lacking a DNA-binding output domain [21] , thus the effects of LovK-LovR on hfiA transcription must be indirect . We investigated whether inhibition of hfiA transcription by lovK-lovR is dependent on CtrA , GcrA , or StaR . Deletion of staR had no effect on inhibition of hfiA transcription by lovK-lovR ( Figure 5C ) . We further demonstrated that lovK-lovR does not affect the occupancy of StaR at the hfiA locus ( Figure S6 ) . GcrA and CtrA regulated genes are not differentially regulated in lovK-lovR transcriptional profiling experiments [20] suggesting that the activity of GcrA or CtrA is not perturbed by LovK-LovR . Similarly , transcription from known GcrA and CtrA regulated promoters is not affected when lovK and lovR are coordinately overexpressed ( Figure S6 ) . We conclude that LovK and LovR affect hfiA transcription via a mechanism that is independent of StaR , GcrA , and CtrA . The cell cycle expression profile of hfiA is coordinated with the timing of holdfast development . However , not every cell makes a holdfast . What determines whether a cell will elaborate a holdfast ? Transcriptional profiling experiments suggest that culture environment affects hfiA transcription; in mid-log phase , cells grown in minimal medium ( M2X ) have 1 . 6 times more hfiA transcript than cells grown in complex medium ( PYE ) [43] . While this relative difference in hfiA transcript level is not large , hfiA is a highly expressed gene . Thus , the absolute difference in transcript levels is large . To test whether culture environment affects probability of holdfast development , we grew wild-type C . crescentus CB15 cells in either complex ( PYE ) or minimal defined medium ( M2X ) and quantified the fraction cells with visible holdfast . The culture environment has a dramatic impact on the probability that a cell displays a holdfast . In PYE medium , approximately 80% of cells develop a holdfast ( Figure 6A ) compared to 1–3% of cells in M2X . We extended this analysis beyond these two standard growth media by analyzing both the probability of holdfast development and hfiA transcription from the PhfiA-lacZ reporter in a series of minimal media supplemented with increasing amounts of peptone , from 0 . 0005% to 0 . 1% . In this panel , we observed an approximately 2-fold change in activity from the hfiA promoter ( Figure 6B ) . The probability of holdfast development in the population shows an inverse linear correlation with hfiA promoter activity ( r2 = 0 . 99; Figure 6B ) . Cells cultured with little or no peptone exhibit the highest hfiA transcription and the lowest fraction of cells with holdfast . Increasing peptone concentration results in both decreased hfiA transcription and an increased fraction of cells with a holdfast . These data are consistent with a model in which modest relative changes in hfiA expression can have a large effect on holdfast development . To test whether hfiA is required for regulated differences in holdfast between growth media , we evaluated holdfast development in the ΔhfiA null mutant . Regardless of the composition of the growth medium , ΔhfiA mutants elaborate a holdfast on nearly every cell ( Figure 6C ) ; only a small portion of swarmer cells can be found without holdfast . Transcription from the hfsJ promoter does not change between these growth conditions ( Figure S7 ) . We conclude that the capacity of a cell to elaborate a holdfast is controlled by the expression of the holdfast regulator , hfiA , and not by a change in expression of the holdfast synthesis gene hfsJ . Finally , we tested if nutrient-dependent regulation of PhfiA and holdfast development requires the LovK-LovR sensory system . We measured the number of holdfast in strains lacking the lovKR locus . In minimal medium only small fraction of ΔlovKR cells display a holdfast . Upon supplementation with 0 . 1% peptone , the majority of ΔlovKR cells exhibit a holdfast ( Figure 6D ) . Similarly , in a ΔlovKR background , the PhfiA-lacZ transcriptional reporter is reduced upon supplementation with 0 . 1% peptone ( Figure 6D ) . Together these results indicate that LovK and LovR are not required for nutrient-dependent control of holdfast and suggest an additional , unknown regulator of hfiA .
Holdfast adhesin development in C . crescentus is regulated by the developmental state of the cell and by the culture environment . This surface organelle emerges at the flagellated pole during the late swarmer cell stage ( [8] , [15]–[17] , [19] , Figure 1 ) . We have discovered a novel small protein , HfiA , whose expression is developmentally and nutritionally regulated , and which functions as a potent inhibitor of holdfast . We demonstrate that the predicted glycosyltransferase , HfsJ , is a required component of the holdfast development machinery and that residues at the C-terminus of HfsJ mediate a direct interaction with HfiA . We propose a model in which HfiA functions as a cell cycle and nutritional checkpoint protein that prevents inappropriate holdfast development via post-translational inhibition of HfsJ ( Figure 7 ) . Notably , the dynamic range of hfiA transcriptional control ( ∼2-fold ) is modest at the population level compared to the dynamic range of holdfast probability ( ∼2-log ) . One prediction from this observation is that the binding affinity and cellular concentrations of HfiA and HfsJ are tuned such that this regulatory system is responsive to small changes rather than robust to large changes . This predication is consistent with a highly responsive and sensitive regulatory system . HfsJ has strong similarity to WecG/TagA–family glycosyltransferases ( E-value<e−84 ) [27] . Enzymes in this family are known to catalyze the transfer of a nucleotide diphosphate ( NDP ) -activated sugar to monoglycosylated Und-PP ( i . e . lipid I ) [44] , [45] . The product of this reaction is the phosphoglycolipid , Und-PP-disaccharide ( i . e . lipid II ) . Varied forms of lipid II are precursors for extracellular polysaccharide structures in bacteria including lipopolysaccharide , wall teichoic acid , capsular polysaccharide , and holdfast . In C . crescentus , there is apparent redundancy in the holdfast synthesis enzymes predicted to catalyze formation of lipid I [12] . Our genetic data suggest that HfsJ is solely responsible for the production of holdfast lipid II . B . subtilis and Staphylococcus aureus TagA catalyze a specific transformation of lipid I to lipid II that commits the phosphoglycolipid for wall teichoic acid biosynthesis [46] . By analogy , we predict that HfsJ commits its phosphoglycolipid substrate to holdfast biosynthesis . Post-translational regulation of such a “gate-keeping” enzyme would enable specific control of lipid I commitment to holdfast development ( Figure 7 ) . Temporally staggered cell-cycle transcription of hfiA and hfsJ correlates with the developmental timing of holdfast synthesis . Key developmental regulators physically interact with and regulate both of these genes , directly tying holdfast development to the core cell cycle control network . The master cell cycle regulator , CtrA , activates transcription of both hfiA and hfsJ . The methylation-responsive transcriptional regulator , GcrA [41] , and the developmental regulator , StaR , provide additional layers of direct hfiA regulation . The activities of these regulators on hfiA , but not hfsJ can account for the temporal shift in hfsJ and hfiA expression . Consistent with the observation of hfs gene transcripts in late pre-divisional and swarmer cells ( [31]–[33] , Figure 5 ) , swarmer cells are born preloaded with all the proteins required to synthesize a holdfast [17] , [19] . staR is activated in swarmer cells ( [31]–[33] , [47] , Figure 5 ) and as a functional repressor of hfiA , presumably drives the decrease in hfiA transcription prior to holdfast development . Decreased expression of the holdfast inhibitor is predicted to permit holdfast production as cells approach the swarmer-to-stalk cell transition . Upon this transition , cells accumulate GcrA [38] , which can initiate activation of hfiA expression , but not hfsJ . As the cell cycle progresses , de novo synthesis and activation of CtrA [37] should reinforce expression of the inhibitor and also activate hfsJ in preparation for the next generation swarmer cell . Notably , hfsJ is among the last holdfast synthesis genes to be transcriptionally activated; it does not reach maximum transcription until just prior or coincident with cell division ( Figure 5 ) . This delayed expression provides two intuitive mechanisms that should restrict premature holdfast synthesis . First , HfsJ is essential for holdfast biosynthesis , thus a preassembled machine will not be functional in the predivisional cell until hfsJ is expressed , just around the time of cell division . Second , peak hfiA expression precedes that of its target . A pool of accumulated inhibitor should block activity of nascent HfsJ in the late predivisional cell . Together these features ensure that the motile swarmer cells are not born with a holdfast and are able to fulfill a dispersal role . What , then , relieves HfiA inhibition so that holdfast development can progress in the swarmer cell ? One possibility is that HfiA is inherently unstable and rapidly degraded by cellular proteases . Indeed , proteins optimized for regulatory flexibility tend to have short half-lives [48] . If HfiA is unstable , high synthesis rates would be necessary to maintain an appreciable steady state concentration in the cell . In this scenario , the initial concentration of HfiA in the swarmer cell ( where hfiA is not transcribed ) could serve as a timer for the initiation of holdfast synthesis . Several efforts by our research group to quantify HfiA protein levels in the cell have been unsuccessful . These negative results provide indirect support for the hypothesis that HfiA is an unstable polypeptide , though we still seek direct experimental support for this hypothesis . Alternatively , post-translational modification could affect HfiA stability or its binding affinity with HfsJ . Surface contact-dependent perturbation of the flagellum [19] could serve as a signal for HfiA modification or degradation . Another possibility is that cyclic-di-GMP ( cdG ) could serve as a second messenger that directly or indirectly inactivates HfiA or activates HfsJ . In many systems cdG serves as a developmental cue signaling the transition from motile to non-motile states [49] . Indeed the activity of the diguanylate cyclase , PleD , is cell cycle regulated and activated during the swarmer to stalk transition [50]; C . crescentus cells lacking pleD are delayed in holdfast development [17] . Thus , it is reasonable to predict that cdG may play a role in control of the HfiA-HfsJ adhesion checkpoint . While the developmental circuitry of the cell directly controls hfiA expression , environmental signals provide an additional regulatory input that can override developmental control . A mixed population of cells grown in carbon replete minimal defined medium have 60% more hfiA transcript than cells grown in complex medium [43]; this correlates with the observed frequency of holdfast-bearing cells in a population ( i . e . cells grown in minimal medium rarely elaborate a holdfast while the majority cells grown in complex medium possess holdfast ) ( Figure 6 ) . Moreover , supplementation of minimal defined medium with peptone modulates both hfiA expression over a two-fold range and the probability of holdfast development over a 2-log range ( Figure 6 ) . A similar correlation is observed upon overexpression of the LovK-LovR two-component sensory system , which results in hfiA repression and increased probability of holdfast development ( Figures 1 & 2 ) . Notably , the nutrient-dependent control of hfiA transcription and holdfast development is independent of the LovK-LovR sensory system . The exact regulatory connection between hfiA transcription and either LovK-LovR signaling or the metabolic state of the cell remains unclear . Indeed , the data presented here speak to existence of at least one additional direct regulator of hfiA , as the repressive effect of LovK-LovR on hfiA transcription is necessarily indirect and also independent of CtrA , GcrA , and StaR . Given the permanence of the cellular decision to adhere to a surface , it is not surprising that environmental and nutritional stimuli influence hfiA expression and holdfast adhesin development . Our study provides evidence that multiple developmental and environmental signals are integrated at the promoter of hfiA , which encodes a novel , small protein inhibitor of the required holdfast synthesis enzyme , HfsJ . C . crescentus employs a multi-level regulatory system that ensures proper timing of holdfast development , and safeguards against permanent cell adherence in a sub-optimal environment .
Standard cloning methods , strain construction techniques and growth conditions were employed and are detailed in the Text S1 . Strains and primers used are in Table S6 . Cells were imaged with a DM5000 microscope ( Leica ) in phase contrast and fluorescence modes using a HCX PL APO 63×/1 . 4na Ph3 objective . Fluorescent samples were excited with an external mercury halide bulb in an EL6000 lamp ( Leica ) . Standard filter sets were used to detect WGA-Alexa594 ( Chroma set 41043 ) and the fluorescent protein , Venus ( Chroma set 41028 ) . Images were captured using an Orca-ER digital camera ( Hamamatsu ) controlled by Image-Pro ( Media Cybernetics , Inc . ) . To visualize holdfast , 100–500 µl of cells were incubated for 10–15 minutes with 50 µg/ml Wheat Germ Agglutinin , Alexa Fluor 594 Conjugate ( Life Technologies , Molecular Probes ) , diluted with 1 ml water or media , collected by centrifugation for 3 minutes ( 14 , 000 g ) , and resuspended in 20–50 µl . For quantitative analyses , overnight cultures were diluted to an approximate OD660 of 0 . 00005 so that after 16–18 hours of growth the culture density was between 0 . 05 and 0 . 1 OD660 . This approach minimized cell-cell adhesion and rosette formation and ensured that all cells were “born” into nutritionally replete conditions . Global transcriptional profiling of ΔsigT xylX::pMT585 vanR::pMT528 ( EV ) and ΔsigT xylX::pMT585-lovR vanR::pMT528-lovK cultures were conducted as in [20] . β-galactosidase activity from promoter-lacZ fusions was measured colorimetrically [51] as in [20] . Transcription start sites were identified by mapping 5′ ends of mRNA using the FirstChoice RLM-RACE kit ( Life Technologies , Ambion ) following the manufacturer's protocol . The RNA template was extracted from log phase cells grown in M2X medium using Trizol ( Life Technologies , Invitrogen ) . 5′GTCGGTCGTGCGCATAGT and 5′GATCTTCGAGCGGCGAAA primers were used as hfiA specific primers . Mid-log phase cells grown in PYE were cross-linked in 10 mM sodium phosphate ( pH 7 . 6 ) and 1% formaldehyde at room temperature for 10 min and on ice for 30 min thereafter , washed three times in phosphate buffered saline ( PBS ) and lysed in a Ready-Lyse lysozyme solution ( Epicentre , Madison , WI ) according to the manufacturer's instructions . Lysates were sonicated ( Sonifier Cell Disruptor B-30; Branson ) on ice using 10 bursts of 20 sec at output level 4 . 5 to shear DNA fragments to an average length of 300–500 bp and cleared by centrifugation for 2 minutes ( 14 , 000 rpm , 4°C ) . Lysates were normalized by protein content , diluted to 1 mL using ChIP buffer ( 0 . 01% SDS , 1 . 1% Triton X-100 , 1 . 2 mM EDTA , 16 . 7 mM Tris-HCl ( pH 8 . 1 ) , 167 mM NaCl plus protease inhibitors ( Roche , Switzerland ) and pre-cleared with 80 µL of protein-A agarose ( Roche , Switzerland ) and 100 µg BSA . Ten percent of the supernatant was removed and used as total chromatin input DNA . Polyclonal antibodies to StaR or CtrA were added to the remains of the supernatant ( 1∶1 , 000 dilution ) , incubated overnight at 4°C with 80 µL of protein-A agarose beads pre-saturated with BSA , washed once with low salt buffer ( 0 . 1% SDS , 1% Triton X-100 , 2 mM EDTA , 20 mM Tris-HCl ( pH 8 . 1 ) , 150 mM NaCl ) , high salt buffer ( same as previous with 500 mM NaCl ) and LiCl buffer ( 0 . 25 M LiCl , 1% NP-40 , 1% sodium deoxycholate , 1 mM EDTA , 10 mM Tris-HCl ( pH 8 . 1 ) ) and twice with TE buffer ( 10 mM Tris-HCl ( pH 8 . 1 ) and 1 mM EDTA ) . The protein•DNA complexes were eluted in 500 µL freshly prepared elution buffer ( 1% SDS , 0 . 1 M NaHCO3 ) , supplemented with NaCl to a final concentration of 300 mM and incubated overnight at 65°C to reverse the crosslinks . The samples were treated with 2 µg of Proteinase K for 2 h at 45°C in 40 mM EDTA and 40 mM Tris-HCl ( pH 6 . 5 ) . DNA was extracted using phenol∶chloroform∶isoamyl alcohol ( 25∶24∶1 ) , ethanol-precipitated using 20 µg of glycogen as carrier and resuspended in 100 µL of water . Illumina Genome Analyzer IIx or HiSeq 2000 runs of barcoded ChIP-Seq libraries yielded several million reads that were mapped to C . crescentus NA1000 ( NC_011916 ) using the ELAND alignment algorithm ( services provided by Fasteris SA , Switzerland ) . Analysis of sequences is described in Text S1 . The goal of this screen was to identify mutants that are insensitive to HfiA and can develop holdfast even when hfiA is overexpressed . Our strategy to enrich the population with holdfast+ mutants is conceptually the opposite of that used by [7] to enrich a population with holdfast null mutants by removing holdfast bearing cells with cheese cloth . Here unattached cells are removed by aspiration and surface attached cells are allowed to populate the culture . FC1935 and FC1936 , strains that overexpress hfiA from either the xylose-inducible promoter on a mid-copy replicating plasmid or from the xylose promoter integrated at two independent sites on the chromosome , respectively , were used as the parental strains . Enrichment was carried out in complex medium , which promotes holdfast development in wild-type cells , supplemented with xylose to induce overexpression of hfiA . Explicit mutagenesis was unnecessary; spontaneous mutants arise in the course of the enrichment . Starter cultures inoculated from freshly grown colonies into PYE supplemented with 0 . 15% xylose and appropriate antibiotics were diluted 1∶100 in 1 ml of fresh medium in each well of a sterile 24-well polystyrene plate with a lid . The lid was sealed with a strip of AeraSeal air-permeable sealing film ( Excel Scientific ) to prevent evaporation . Plates were incubated with gentle shaking ( 155 rpm ) at 30°C overnight . The culture medium was removed by aspiration . Unattached cells were thoroughly washed away with a stream of sterile water expelled from a 20 cc syringe through a 22 G needle . The inoculum of attached cells was allowed to regrow to saturation in 1 ml of fresh medium under the same growth conditions . Washing and regrowth in fresh medium were repeated . After the first wash , the wells appear clear and regrowth requires 36–48 hours , but after 3–4 rounds of washing the wells are cloudy with attached cells and the outgrowth saturates in less than 18 hours . The culture was serially diluted and plated on solid medium to isolate single colonies . Isolated colonies were subjected to several secondary screens . First , the Pxyl-hfiA region of the overexpression plasmids were amplified and re-sequenced to eliminate mutants in the overexpression system . Second , surface attachment to polystyrene was measured with crystal violet staining ( see below ) to confirm enhanced adhesion capacity of each isolate . Third , cells were grown in minimal medium with xylose as the sole carbon source ( M2X ) to ensure a functional xylose transport system . Fourth , WGA-Alexa594 binding was used to assess holdfast development . Genomic DNA was isolated from individual suppressor strains using guanidium thiocyanate [52] . Bar-coded Next Generation Sequencing libraries were pooled and sequenced ( 50-bp single end reads ) using the SOLiD 5500 xl sequencing platform ( Applied Biosystems , Life Technologies ) generating an average of 12 million reads per library . The Functional Genomics Facility at the University of Chicago provided sequencing services . Sequences were processed through an automated analysis pipeline by University of Chicago Center for Research Informatics . The analysis pipeline is described in more detail in Text S1 . Cells were grown from 5 ul of an overnight starter culture in 1 ml of fresh growth medium in 24-well sterile polystryrene plates with lids . Lids were sealed with AeraSeal air-permeable sealing film ( Excel Scientific ) and plates were incubated with gentle shaking ( 155 rpm ) at 30°C for 24 hours . Culture medium and planktonic cells were removed by aspiration and washed away with tap water . Surface attached cells were measured by crystal violet staining using a protocol similar to those outlined by [53] , [54] . Briefly , wells were incubated with 1 ml 0 . 01% crystal violet for 5 minutes with gentle shaking then washed with tap water to remove unbound stain . Bound stain was extracted with 1 . 5 ml 95% ethanol while gently shaking for 5–10 minutes . Extracted stain was diluted 1∶6 and the optical density at 575 nm was measured spectroscopically . A 50 ml overnight culture ( 30°C , 220 rpm ) was used to inoculate 500 ml of LB broth supplemented with appropriate antibiotics and 0 . 2% glucose to repress expression of endogenous maltose degradation genes . When the culture reached OD660∼0 . 8 , 0 . 5 mM of IPTG was added to induce expression . After 2 hours ( 30°C , 220 rpm ) , cells were harvested by centrifugation for 20 min ( 12 , 000 g , 4°C ) and the pellet was resuspended in 5 ml of Tris-NaCl buffer ( 10 mM Tris pH 7 . 4 , 150 mM NaCl ) supplemented with 10 mM imidazole , 5 µg/ml of DNase I and PMSF . Cells were disrupted by three passages in a French pressure cell , and cell debris was removed by centrifugation for 20 min ( 25 , 000 g , 4°C ) . The supernatant was then mixed with 500 µl of Amylose resin ( New England Biolabs ) pre-equilibrated with Tris-NaCl buffer , which allowed for binding of MBP domains . The resin was thoroughly washed with the Tris-NaCl buffer and bound proteins were eluted with 500 µl of Tris-NaCl buffer supplemented with 20 mM maltose . The eluted proteins were mixed with 100 µl of Ni2+-NTA Sepharose affinity resin ( GE Life Sciences ) pre-equilibrated with Tris-NaCl buffer to allow binding of His-tagged proteins . Two stringent washing steps were performed with Tris-NaCl buffer containing 10 mM or 75 mM imidazole followed by elution with 100 µl of 1 M imidazole Tris-NaCL buffer . To monitor proteins bound and eluted from each resin , samples were separated by electrophoresis on a 14% SDS-PAGE gel and stained with Coomassie . Polyacrylamide fragments containing purified proteins were excised and sent to the Pan Facility at Stanford University , Palo Alto , CA for Mass Mapping to confirm protein identity . In a similar way , the reverse experiment ( purification first with Ni2+ Sepharose affinity resin and then with Amylose resin ) was also performed . Based on the system developed by [30] , plasmid bearing fusions to either the T18 or T25 domains of adenylate cyclase were co-transformed into the adenylate cyclase null strain , BTH101 , by electroporation . The outgrowth was serially diluted and plated on LB-agar containing Amp100 , Kan50 , X-gal ( 40 µg/ml ) and IPTG ( 0 . 5 mM ) . The color of single colonies from each transformation was evaluated after 48 hours growth at 30°C . Two colonies from each strain were repatched on identical medium for side-by-side comparisons .
|
Bacteria predominantly exist within surface-attached communities that facilitate metabolic cooperation , sharing of genetic information , and protect cells against stress . The freshwater bacterium , Caulobacter crescentus elaborates an adhesive structure known as the holdfast , which enables surface attachment . We have discovered a novel mechanism that controls holdfast development in response to cell cycle and environmental cues . This regulatory mechanism involves a small protein inhibitor , HfiA , which targets a conserved holdfast synthesis enzyme and ensures that the holdfast is produced at the appropriate stage of cell development and under the appropriate environmental conditions . To our knowledge , the regulatory system we report here is unprecedented , and provides a mechanism for integrative control of bacterial cell adhesion in response to cell cycle and environmental signals .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"bacteriology",
"gene",
"regulation",
"microbiology",
"bacterial",
"biochemistry",
"gene",
"function",
"prokaryotic",
"models",
"model",
"organisms",
"molecular",
"genetics",
"microbial",
"growth",
"and",
"development",
"microbial",
"physiology",
"cell",
"adhesion",
"caulobacter",
"crescentus",
"biology",
"bacterial",
"physiology",
"bacterial",
"biofilms",
"genetic",
"screens",
"gene",
"identification",
"and",
"analysis",
"genetics",
"molecular",
"cell",
"biology"
] |
2014
|
A Cell Cycle and Nutritional Checkpoint Controlling Bacterial Surface Adhesion
|
The Fanconi Anemia ( FA ) pathway is important for repairing interstrand crosslinks ( ICLs ) between the Watson-Crick strands of the DNA double helix . An initial and essential stage in the repair process is the detection of the ICL . Here , we report the identification of UHRF2 , a paralogue of UHRF1 , as an ICL sensor protein . UHRF2 is recruited to ICLs in the genome within seconds of their appearance . We show that UHRF2 cooperates with UHRF1 , to ensure recruitment of FANCD2 to ICLs . A direct protein-protein interaction is formed between UHRF1 and UHRF2 , and between either UHRF1 and UHRF2 , and FANCD2 . Importantly , we demonstrate that the essential monoubiquitination of FANCD2 is stimulated by UHRF1/UHRF2 . The stimulation is mediating by a retention of FANCD2 on chromatin , allowing for its monoubiquitination by the FA core complex . Taken together , we uncover a mechanism of ICL sensing by UHRF2 , leading to FANCD2 recruitment and retention at ICLs , in turn facilitating activation of FANCD2 by monoubiquitination .
DNA interstrand crosslinks ( ICLs ) can obstruct the critical processes of transcription and replication [1] . If unrepaired , an ICL can perturb accurate chromosome segregation during mitosis when replication has not been completed due to the crosslink , this can additionally lead to replication fork collapse causing double strand DNA breaks . For these reasons , repair of ICLs is required to preserve genomic stability . In humans , the Fanconi anemia ( FA ) pathway is critical to ICL repair . FA is a congenital disorder , which leads to bone marrow failure and predisposition to various cancers . FA patients are sensitive to ICLs due to defects in the FA pathway [2] . At least 22 FA proteins operate together to repair an ICL , assisted by several other key DNA repair factors , which are not known to be bona fide FA proteins [3–5] . The complicated repair process relies on multiple DNA repair pathways coordinated through the FA pathway . The repair process entails recognition of the damage , multiple incisions by nucleases , translesion synthesis ( TLS ) , nucleotide excision repair ( NER ) and homologous recombination ( HR ) [6] . A key step in initiation of repair requires recognition of the ICL . We previously identified UHRF1 as an ICL sensor protein [7] . We showed that UHRF1 interacts directly with ICLs in vitro and in vivo and that its recruitment in vivo is functionally important to initiate the repair reaction . UHRF1 has also been shown to function with BRCA1 in the double strand break repair pathway choice [8] . UHRF2 , a paralogue of UHRF1 , is structurally similar to UHRF1 , with five recognizable domains . UHRF2 contains a ubiquitin-like domain ( UBL ) , Tandem-Tudor domain ( TTD ) , Plant Homeodomain ( PHD ) , SET and RING associated domain ( SRA ) , and Really Interesting New Gene domain ( RING ) . The UBL domain resides in the N-terminus of the protein and structurally resembles ubiquitin . This structural similarity is evident when comparing the UBL of UHRF1 to Ubiquitin , as superimposition of the structures yields an RMSD value of 0 . 52Å [9] . TTD and PHD domains have been shown to work together in the recognition of H3K9me2/3 histone marks [10] , while the SRA domain , which is unique to UHRF1 and UHRF2 in humans [11] is a DNA binding domain , shown to recognize hemimethylated DNA [12] . Lastly , the RING domain of UHRF2 is an E3 ligase . UHRF2 has been shown to possess autoubiquitinaton activity and is capable of ubiquitinating PCNP , however , there is not much known about potential other substrates of the UHRF2 E3 ligase [13] . UHRF2 is not very well characterized , however it has been shown to be important in cell cycle progression and can cause G1 phase cell cycle arrest through its interaction with the Cdk2–cyclin E complex [14 , 15] . UHRF2 has been implicated in genome maintenance as it has been shown to interact with PCNA through a PIP box motif [16] , additionally , UHRF2 was shown to be recruited to DNA damage sites , primarily through its TTD-PHD and SRA domains [17] . Here we report the identification of UHRF2 as a sensor for ICLs in humans . We demonstrate that UHRF2 functionally cooperates with UHRF1 in the ICL repair process , and that both proteins physically interact . Importantly , we show that UHRF1 and UHRF2 interact directly with FANCD2 , in vitro and in vivo , and that UHRF1/UHRF2 facilitate the recruitment and retention of FANCD2 to ICLs , providing a mechanistic insight into how UHRF1 and UHRF2 mediate the early stages of ICL repair .
We previously devised a purification scheme to identify proteins interacting with ICL-containing DNA [7] . Using this scheme , we excised a band from an SDS-PAGE gel containing such proteins , and identified all proteins in the band by mass spectrometry [7] . We identified a total of two proteins , namely the paralogues UHRF1 and UHRF2 , by 76 and 11 peptides , respectively ( Fig 1A and [7] ) . We previously found UHRF1 to interact directly with ICLs both in vitro and in vivo , therefore we decided to investigate whether UHRF2 might possess similar properties . We expressed full-length UHRF2 in Sf9 insect cells and purified the protein to homogeneity ( Fig 1B ) . We then designed a DNA probe containing a single ICL in the center ( Fig 1C ) , and radiolabeled it by 32P . Using electrophoretic mobility shift assay ( EMSA ) we then assessed the ability of UHRF2 to interact directly with an ICL . As expected , we observed a weak complex formed with the control DNA molecule , and a stronger complex formed with the probe containing a central ICL ( Fig 1D , lanes 1–2 ) . In good agreement , we observed the same trend when assaying the DNA-binding properties of UHRF1 ( Fig 1D , lanes 3–4 ) . Both the UHRF1 and UHRF2 protein-DNA complexes could be supershifted using specific antibodies , confirming their identities ( Fig 1D , lanes 5–6 ) . Given the clear result in vitro , we then sought to test whether UHRF2 is recruited to ICLs in vivo . We previously described an experimental system where we can observe the live recruitment of fluorophore-fused proteins to ICLs using live-cell imaging [7] . The system is based on stable expression of the protein of interest as a fusion protein with a fluorophore . Cells are then incubated with the psoralen analog TMP ( trimethylpsoralen ) , and placed in a spinning disc microscope . While cells are being imaged , a local region of the nucleus is irradiated with a laser beam , causing the formation of ICLs in that region . Cells can then be monitored over time , allowing for observation and quantification of recruitment of the protein of interest to ICLs . Using this method , we analyzed the recruitment of UHRF2-EGFP to ICLs in HeLa cells . We observed recruitment of UHRF2-EGFP already 30 seconds after the introduction of the ICLs , kinetics similar to that of UHRF1 , and the strength of the signal increased over the following minutes ( Fig 1E ) . Taken together , UHRF2 displays distinct characteristics of an ICL sensor protein , namely direct interaction with ICLs in vitro and rapid recruitment to ICLs in vivo . Since UHRF2 is recruited to ICLs , we speculated that the protein might be involved in ICL repair . To test this directly , we evaluated the ability of cells to respond to ICLs after depleting the UHRF2 protein . To this end , we disrupted the UHRF2 gene in HeLa cells using CRISPR/Cas9 genome editing , leading to a complete depletion of endogenous UHRF2 protein ( S1A–S1C Fig ) . HeLa and HeLa UHRF2 -/- cells were then subjected to a clonogenic survival assay , where cells are treated with increasing concentrations of MMC , and survival is assessed by counting the number of colonies formed . As expected , depletion of UHRF2 resulted in sensitivity to MMC ( Figs 2A and S2A ) . We also tested the sensitivity towards ICLs formed by TMP . Again , UHRF2 deficient cells were sensitive compared to control cells ( Fig 2B ) . We assessed the sensitivity of UHRF2 deficient cells to other types of DNA damage and replication stress , and observed no sensitivity towards UVC and HU ( hydroxyurea ) ( Fig 2C and 2D ) . UHRF2 contains 5 unique domains , the UBL , PHD , TTD , SRA and RING domains ( Fig 3A ) . To gain more mechanistic insight into how UHRF2 is recruited to ICLs , we decided to abrogate each of the 5 domains one at a time , and then assess the recruitment of the resulting mutant proteins in vivo using live-cell imaging . EGFP-tagged versions of UHRF2 containing each of the 5 deletions were then stably expressed in UHRF2 -/- cells ( S2B Fig ) . Wildtype UHRF2-EGFP was recruited normally , showing quick recruitment to ICLs ( Fig 3B ) . Deletion of the UBL , PHD or RING domains caused no or minor reduction of recruitment , while deletion of the TTD domains caused a mild reduction of recruitment ( Fig 3B ) . In contrast , deleting the SRA domain completely abrogated the recruitment to ICLs ( Fig 3B ) . Since the SRA domain appears critical for recruitment of UHRF2 to ICLs , we next sought to test whether this and the other domains are also functionally important for the ICL repair function of UHRF2 . The five deletion versions of UHRF2 , where each of the domains were abrogated one at a time , were expressed in UHRF2 -/- cells , and the resulting cell lines were assessed for sensitivity to ICLs . In good agreement with our live-cell imaging data , deletion of the SRA domain completely abrogated the function of UHRF2 in response to ICLs , whereas deletion of the TTD domain caused a mild reduction of function , and deletion of UBL , PHD and RING domains caused no reduction of function ( Fig 3C ) . Given the clear functional importance of UHRF2 in ICL repair , and the previously suggested interplay between UHRF1 and FANCD2 [7] , we speculated that UHRF2 might function by recruiting FANCD2 to ICLs . To test this directly , we stably expressed mCherry-tagged FANCD2 in HeLa cells and in HeLa UHRF2 -/- cells ( S2C Fig ) , and assessed the ability of mCherry-FANCD2 to be recruited to ICLs in vivo . As expected , mCherry-FANCD2 was recruited normally in control cells ( Fig 4A ) . However , the recruitment was reduced when UHRF2 was depleted . We then assessed the recruitment of mCherry-FANCD2 upon UHRF1 reduction using shRNA ( S2C Fig ) , and also observed reduced recruitment of mCherry-FANCD2 to ICLs ( Fig 4A ) . Cellular depletion of UHRF1 and UHRF2 simultaneously led to strong reduction in recruitment of mCherry-FANCD2 ( Fig 4A ) . In good agreement with these data , we observed the same phenotype of reduced FANCD2 recruitment when cellular levels of UHRF1 and UHRF2 were reduced using only shRNA ( S3A and S3B Fig ) . Recruitment of FANCD2 to ICLs has traditionally been assessed via its formation of nuclear foci . Therefore , to test our conclusions , we asked whether the ability of FANCD2 to form such foci in response to either MMC or TMP/UVA , depends on the activities of UHRF1 and UHRF2 . We assessed the FANCD2 foci formation upon depletion of UHRF1 , UHRF2 , or both . The formation of FANCD2 foci was markedly reduced upon depletion of either or both of the proteins ( S3C and S3D Fig ) , strengthening the conclusions based on our experiments using localized introduction of ICLs ( Fig 4A ) . Monoubiquitination of FANCD2 on lysine 561 is required for its activation and accumulation on ICLs in vivo . Therefore , we hypothesized that UHRF1 and UHRF2 might affect the monoubiquitination of FANCD2 . To check this directly , we depleted UHRF1 , UHRF2 or both UHRF1 and UHRF2 in HeLa cells ( S4A Fig ) and determined the ability of the resulting cells to monoubiquitinate FANCD2 . We treated the cells with TMP and UVA and monitored the amount of monoubiquitinated FANCD2 over time . We observed a clear increase in the amount of monoubiquitinated FANCD2 after 3 and 6 hours ( Fig 4B , lanes 1–3 , S4B and S4C Fig ) . In contrast , monoubiquitination was reduced upon depletion of UHRF1 and/or UHRF2 ( Fig 4B , lanes 4–12 , S4B and S4C Fig ) . FACS analysis determined that a change in the population of S-phase cells upon reduction of UHRF1/2 did not cause this phenotype ( S4D Fig ) . On the other hand , an accumulation of G2/M cells in UHRF1/2 depleted cells was observed , a phenotype typically observed in cells with a disrupted FA pathway . To further test the conclusion that UHRF1/2 stimulates FANCD2 recruitment and monoubiquitination , we assessed the recruitment of FANCD2 to ICLs in cells depleted of endogenous UHRF1 , and complemented with a mutant of UHRF1 where the SRA domain has been deleted . This mutant protein of UHRF1 is not recruited to ICLs itself . In these cells , the recruitment of FANCD2 is abrogated , underscoring a functional relationship between UHRF1/2 and FANCD2 recruitment ( S5A Fig ) . Both UHRF1 and UHRF2 are E3 ligases . It could therefore be a possibility that these enzymes directly monoubiquitinate FANCD2 . To test this possibility , we assessed the monoubiquitination of FANCD2 in vitro , using purified components . While FANCL , the E3 ligase for FANCD2 , provided robust monoubiquitination of FANCD2 , neither of the UHRF1 and UHRF2 enzymes possessed such activity in vitro , suggesting that these enzymes are not E3 ligases for FANCD2 ( S5B–S5E Fig ) . Given the observed contribution by both UHRF1 and UHRF2 towards recruitment of FANCD2 , we speculated that the two proteins might form a physical interaction not described in the literature . To test this directly , we expressed either Flag-HA-UHRF1 or Flag-HA-UHRF2 proteins in HeLa cells , and assessed the co-immunoprecipitation of the two proteins with endogenous UHRF2 and UHRF1 , respectively , post TMP/UVA treatment . As predicted , immunoprecipitates of Flag-HA-UHRF1 contained UHRF2 , and immunoprecipitates of Flag-HA-UHRF2 contained UHRF1 ( Fig 4C ) . To further characterize the interaction between these two proteins we turned to in vitro co-immunoprecipitation , utilizing purified recombinant UHRF1 domain deletion mutants and purified recombinant UHRF2 . We identified the domain of UHRF1 that is responsible for the interaction with UHRF2 as the SRA domain ( Figs 4D and S5F ) . The interaction between UHRF1 and UHRF2 prompted us to determine whether this translates to a functional genetic relationship between the two genes . We depleted UHRF1 and UHRF2 by shRNA and CRISPR/Cas9 , respectively ( S4A Fig ) . Survival of the resulting cell lines in response to MMC was subsequently assessed . We observed a sensitivity resulting from reduction of UHRF1 and UHRF2 by knockdown or knockout , respectively , as expected . Importantly , in good agreement with the biochemical data , we did not observe a significant further sensitization of UHRF2 -/- cells upon depleting UHRF1 by knockdown ( Fig 4E ) . These data suggest that both UHRF1 and UHRF2 contribute towards FANCD2 recruitment and towards ICL repair . However , it is not clear whether these two repair factors affect the recruitment of each other or whether their simultaneous recruitment is needed for full FANCD2 recruitment . To differentiate between these two possibilities , we assessed the recruitment of UHRF1 to ICLs in the absence of UHRF2 , and vice versa . Depleting either UHRF1 or UHRF2 did not significantly reduce the recruitment of UHRF2 and UHRF1 , respectively ( Fig 5A ) . Additionally , over-expression of mCherry-UHRF1 in HeLa UHRF2 -/- cells , did not restore monoubiquitination of FANCD2 ( S6A and S6B Fig ) . Taken together , these data show that UHRF1 and UHRF2 both need to be recruited to ICLs for subsequent FANCD2 recruitment and monoubiquitination to take place , and that the recruitment of either factor takes place independently of the other factor . Both UHRF1 and UHRF2 have been reported to interact with hemimethylated DNA [12] . It was therefore plausible that such an interaction could play a role in the recruitment of these proteins to ICLs . We decided to test this hypothesis directly . Cells expressing either mCherry-UHRF1 or UHRF2-EGFP were treated with 5-Aza-2'-deoxycytidine for 5 days to reduce DNA methylation . The treatment led to a near elimination of DNA methylation ( Fig 5B ) . The ability of the proteins to be recruited to ICLs was then assessed and compared to untreated cells . We observed no significant reduction in recruitment of UHRF1 to ICLs when DNA methylation was reduced , demonstrating that recruitment to ICLs does not depend on DNA methylation , while the recruitment of UHRF2 was slightly reduced . ( Fig 5C ) . To further strengthen this conclusion , we assessed the degree of UHRF1 and UHRF2 recruitment to ICLs in the G1- and S-phases of the cell cycle . The amount of hemimethylated DNA is significant in the S-phase , where replication leads to synthesis of new non-methylated DNA , thereby creating hemimethylated DNA . On the other hand , G1 cells contain much less hemimethylated DNA [18] . We stably expressed mCherry-UHRF1 or UHRF2-EGFP together with a cell cycle marker containing the 110 N-terminal amino acids of Geminin fused with EGFP or mCherry , respectively . The cell cycle marker is absent in G1-cells and present in S/G2-cells [19] . Using these stable cell lines , we could then assess the recruitment of UHRF1 and UHRF2 in G1 and S/G2 cells without the use of any perturbing drugs , which could affect the cells . We observed strong recruitment of both UHRF1 and UHRF2 in both G1 and S/G2 cells , with UHRF2 recruitment of slightly higher amplitude in G1 cells compared to S/G2 cells ( Fig 5D ) . Perhaps a larger fraction of UHRF2 is mobile in G1 compared to G2/S . A previous report suggested that UHRF2 is active in G1 as a cell cycle regulator [20] . The kinetics of recruitment for both proteins was not dependent on the cell cycle . UHRF1 and UHRF2 are recruited to the ICL shortly after DNA damage and this is , in turn , necessary for the recruitment of FANCD2 to the ICL . We also show that depletion of UHRF1 and UHRF2 reduces the degree of FANCD2 monoubiquitination . It was recently found that monoubiquitination of FANCD2 occurs after its recruitment to DNA [21] . Therefore , it is possible that UHRF1 and UHRF2 facilitate the retention of FANCD2 at the ICL , potentially through a direct protein-protein interaction . To test this , we first assessed whether the two proteins interact with each other in vivo . Cellular extracts from HeLa cells expressing either epitope-tagged UHRF1 , or epitope-tagged FANCD2 , were subjected to immunoprecipitation after ICLs had been induced by TMP/UVA . Western blot analysis confirmed the presence of endogenous FANCD2 in the UHRF1 immunoprecipitate ( Fig 6A , lane 4 ) and endogenous UHRF1 in the FANCD2 immunoprecipitate ( S7A Fig , lane 4 ) . These data suggest that some FANCD2 interact with UHRF1 in the cell . We then went on to test whether the UHRF1 and FANCD2 interact directly . To this end , we purified both proteins in their full-length forms from Sf9 cells , and assessed their ability to interact in vitro . We observed a direct and strong interaction between FANCD2 and UHRF1 ( Fig 6B , lane 9 ) . Importantly , there was no interaction between FANCD2 and MBP , used as a negative control , confirming the specificity of the interaction ( Fig 6B , lane 10 ) . We confirmed the observed interaction by the reciprocal purification ( Fig 6C , lane 6 ) . To additionally reinforce the interaction data , we analyzed either UHRF1 or FANCD2 alone , or after they had been incubated together , by gel filtration . These experiments also confirmed a UHRF1/FANCD2 protein-protein interaction ( S7B Fig ) . Additional experiments demonstrated that UHRF2 interacts with FANCD2 in a similar fashion ( Fig 6D ) . Given that FANCD2 is likely to be monoubiquitinated before or shortly after its recruitment to the ICL , we wanted to test whether the ubiquitination status affects its interaction with UHRF1 . We purified either unmodified or monoubiquitinated FANCD2 to homogeneity from HeLa cells , and assessed their abilities to interact with UHRF1 . Both forms interacted equally well with UHRF1 , suggesting that the interaction is not dependent on monoubiquitination of FANCD2 ( S7C Fig ) . Our data demonstrate that UHRF1 and FANCD2 form a direct protein-protein interaction . We next sought to determine which domain of UHRF1 is interacting with FANCD2 . To this end , we expressed and purified mutant versions of UHRF1 from Sf9 cells , each containing either the UBL , TTD , PHD , SRA or RING domains deleted ( S7D Fig ) . Again , a strong interaction was observed between wild type UHRF1 and FANCD2 ( Fig 6E , lane 10 ) . Likewise , all mutant proteins where either the UBL , TTD , PHD or RING domain was deleted interacted well with FANCD2 , whereas deletion of the PHD domain might slightly enhance the binding . However , when we deleted the SRA domain it completely abolished the interaction with FANCD2 ( Fig 6E , lane 10 ) . These data suggest that the function of UHRF1/2 in ICL repair might depend on FANCD2 . To test this possibility , we depleted UHRF2 in cells already depleted of FANCD2 , and asked whether we would observe further sensitization . We observed no further sensitization , suggesting that UHRF2 indeed functions together with FANCD2 in ICL repair ( Figs 6F and S7E ) .
Here we show that UHRF2 , a paralogue of UHRF1 , plays an essential role in the initial stage of the ICL repair process . Simultaneous depletion of UHRF1 and UHRF2 significantly sensitizes cells to ICL forming agents , such as MMC and TMP/UVA . We found that UHRF2 interacts directly with an ICL in vitro , and that it is recruited to ICLs within seconds of their appearance in vivo , using live-cell imaging . We also found that both UHRF1 and UHRF2 are needed for efficient recruitment of FANCD2 to ICLs , and for its subsequent monoubiquitination by the FA core complex . UHRF1 and UHRF2 are relatively homologous , with an overall protein sequence identity of 54% and overall sequence similarity of 68% . However , the two proteins are not redundant for their cellular ICL repair functions , instead we found the two proteins to functionally cooperate in ICL repair . Recruitment of UHRF1 and UHRF2 to ICLs in cells occurs independently . However , recruitment of both proteins is required for normal FANCD2 recruitment and monoubiquitination . We show that FANCD2 is not monoubiquitinated by UHRF1 and UHRF2 , however it is possible that other ICL repair proteins are ubiquitinated by these E3 ligases after recruitment to the ICL . As both UHRF1 and UHRF2 are E3 ligases and have been shown to be important for chromatin remodeling , it is also possible that UHRF1/2 mediate histone modification at the site of the ICL , or recruit the DNA methyl transferase DNMT1 mediating DNA methylation [11] . Additionally , UHRF2 has been shown to directly interact with PCNA , perhaps facilitating PCNA recruitment for subsequent DNA repair [16 , 20] and UHRF1 has been reported to interact with nucleases involved in ICL repair [22] . The various proposed roles of UHRF1 in ICL repair has been discussed extensively elsewhere [4] . We have uncovered a direct protein-protein interaction between UHRF1/UHRF2 and FANCD2 . Interestingly , the SRA domain of UHRF1 , which is only present in UHRF1 and UHRF2 in humans [11] , is required for this interaction . The SRA domains share 75% amino acid identity . The structure of the SRA domain of UHRF1 in complex with hemimethylated DNA has been solved [23–25] as have the structures of the SRA domain of UHRF2 in complex with either hemimethylated or hemihydroxymethylated DNA [12] . In all cases , the protein is shaped as a saddle sitting on the DNA , creating a large surface all around the SRA domain , available for interaction with other proteins , such as FANCD2 . Intriguingly , it was similarly shown that DNMT1 also interacts with the SRA domain , important for its recruitment to hemimetylated DNA [26 , 27] . We found that while recruitment of UHRF2 and UHRF1 to ICLs occur independently of each other , both proteins need to be present at the ICL for full repair . It is possible that a stronger retention of FANCD2 on chromatin is obtained when both proteins are present . Future structural studies will be required to understand the atomic nature of the ICL/UHRF1/UHRF2/FANCD2 interactions . We found that DNA methylation is not a main prerequisite for UHRF1 and UHRF2 recruitment to ICLs . Our data support a model where UHRF1/2 are recruited very quickly to the ICL , facilitating the following recruitment , and importantly , retention , of FANCD2 . If UHRF1 and UHRF2 function to retain monoubiquitinated FANCD2 on chromatin , we would expect some monoubiquitination to take place in the absence of UHRF1/UHRF2 , but we would expect little or no retention of FANCD2 at ICLs , measured by live-cell imaging . This is indeed what we observe . Also , in good agreement , we recently demonstrated that FANCD2 is ubiquitinated after its recruitment to DNA as opposed to before recruitment [21] . Therefore , it is plausible that retention of FANCD2 at the ICL via a direct interaction with UHRF1/UHRF2 , allows for the ICL repair to initiate ( Fig 7 ) . We speculate that UHRF1 and UHRF2 form a type of landing platform for FANCD2 on DNA , ensuring the necessary retention of FANCD2 permitting for its monoubiquitination and activation preventing its dissociation from DNA . It is plausible that the FA core complex is recruited independently by the FANCM/FAAP24 complex [28] . Interestingly , we observe recruitment of UHRF1/UHRF2 to ICLs both in the S- and G1-phases of the cell cycle . It is known that FANCD2 is not recruited efficiently in G1 . We speculate that UHRF1/UHRF2 interact with and mark ICLs during G1 allowing for prompt recruitment and monoubiquitination of FANCD2 as soon as the cell enters S-phase . We previously showed that FANCD2 is recruited to ICLs before it is monoubiquitinated , and that monoubiquitination stabilizes its retention on DNA . In good agreement with these data , a non-ubiquitinatable mutant of FANCD2 ( K561R ) is only slightly enriched at ICLs [21] . Current in vivo single-molecule live-cell imaging of these molecules at the millisecond timescale might allow us to test these hypotheses . A recent report placed UHRF1 functionally together with the majority of FA and other ICL repair proteins , further underscoring a role in these processes [29] . In conclusion , we report UHRF2 as a novel ICL sensor protein . UHRF2 interacts directly with ICLs and facilitates the recruitment and retention of FANCD2 to ICLs . Retention of FANCD2 at the ICL allows for its monoubiquitination , and for the ICL repair process to initiate . It will be interesting to explore whether UHRF1 and UHRF2 also play roles in other DNA repair pathways .
HeLa and HEK293T cells were grown in DMEM ( D5796 , Sigma ) supplemented with 2 . 5–10% FBS . Antibodies used were as follows: anti-UHRF1 antibody ( sc-373 , Santa Cruz Biotechnology ) ; anti-FANCD2 ( sc-20022 , Santa Cruz Biotechnology ) ; anti-α-Tubulin ( 5829 , Millipore ) ; anti-UHRF2 ( sc-54252 , Santa Cruz Biotechnology ) ; anti-p300 ( sc-584 , Santa Cruz Biotechnology ) and anti-HA ( mouse monoclonal antibody clone 12CA5 ) . EGFP-fused FANCD2 and mCherry-fused UHRF1 cDNA were expressed using a derivative of the pOZ-N plasmid [30] . EGFP-fused UHRF2 cDNA was expressed using a derivative of the pOZ-C plasmid . shRNA-mediated knockdown of the UHRF1 , UHRF2 , and FANCD2 genes was achieved by expressing the target sequences 5’-AGATATAACGTTAGGGTTT-3’ , 5'-TGTAAAGGCTGGTGAAAGA-3' and 5’-GAGCAAAGCCACTGAGGTA-3’ , respectively , in the pSuper . retro vector ( Clontech ) . Transfections of plasmid DNA were carried out using FuGENE6 ( Promega ) according to the manufacturer’s instructions . The UHRF1 domain deletion plasmids were generated as above . The deleted regions of UHRF1 are amino acids 14–91 for the UBL domain , 146–296 for the TTD domain , 323–379 for the PHD domain , 427–630 for the SRA domain and 727–806 for the RING domain . The deleted regions of UHRF2 are amino acids 1–78 for the UBL domain , 118–312 for the TTD domain , 419–648 for the PHD domain , 419–648 for the SRA domain and 733–771 for the RING domain . HeLa UHRF2 -/- cells were generated using plasmid pX459 ( Addgene #48139 ) [31] . The targeting sequence used in the sgRNA was: 5’-GTGCCCGTCTTATTGATCC-3’ . Primers , 5’- caccgGTGCCCGTCTTATTGATCC -3’ and 5’- aaacGGATCAATAAGACGGGCACc -3’ , were annealed and introduced into the pX459 plasmid through its BbsI site . HeLa cells were transfected with 2 μg of the resulting pX459 plasmid and selected with 4 μg/ml puromycin after 24h . After another 24h cells were plated at low density and clones were picked after 2 weeks . Clones were analyzed using immunoblot analysis and genotyped by sequencing the gRNA target site of the gene locus . ICLs were essentially prepared as described [7] . In brief , the DNA oligos were annealed in the buffer containing 10mM Tris-HCl pH7 . 5 , 100mM NaCl and 1mM EDTA . 4 , 5′ , 8-trimethylpsoralen ( TMP , Sigma , T6137 ) /UVA ( 365nm ) crosslinking induction was described previously [32] . Interstrand crosslink was confirmed by 8M urea 20% denaturing polyacrylmide gel electrophoresis . Proteins were reduced with DTT , cysteine residues were derivatized with iodoacetamide , and the proteins were separated by SDS-PAGE . Proteins from silver stained gel bands were in-gel digested with trypsin [33] . The generated peptide mixtures were subjected to LC-MS/MS using a hybrid linear ion trap/ FT-ICR mass spectrometer ( LTQ FT , Thermo Electron ) essentially as described previously [34] . MS/MS spectra were assigned by searching them with the SEQUEST algorithm [35] against the human International Protein Index sequence database . Proteins purified from Sf9 cells were expressed using the pFastBac1 vector ( Life Technologies ) with an engineered N-terminal Flag-HA tag . Cell pellets were resuspended in lysis buffer ( 20mM Tris- HCl pH 8 . 0 , 0 . 1M KCl , 10% glycerol , 0 . 1% Tween-20 , 2mM β-ME and 0 . 2mM PMSF ) . Lysates were clarified by centrifugation , and the supernatants were incubated with M2 anti-Flag agarose resin for 2 hr . The resin was washed extensively , and the protein was eluted in the same buffer containing 0 . 5mg/ml Flag peptide , but excluding Tween-20 . Flag-HA-FANCD2 and Flag-HA-FANCD2-Ub were purified as described [36] . 6xHis-MBP was purified from BL21 ( DE3 ) cells using the pET28a vector ( plasmid kindly provided by Dr . Mark Howarth ) following standard purification methods . E1 , E2 and E3 enzymes were purified as described [21] . The recombinant proteins were expressed and purified from Sf9 insect cells as indicated above . 1μg of each protein was mixed in the reaction buffer containing 10μg BSA ( NEB ) , 20mM Tris-HCl pH 8 . 0 , 150mM KCl , 10% Glycerol , 2mM β-ME and 0 . 2mM PMSF in 10μl . The mixture was first incubated at 37°C for 1h for protein complex formation . Protein A sepharose coupled with anti-HA IgG were added subsequently , and the mixture was incubated at 4°C with gentle mixing for 30 minutes . The mixture was then transferred to Micro Bio-Spin Chromatography Columns ( Bio-Rad ) , and washed with the reaction buffer supplemented with 0 . 1% Tween-20 . The proteins were eluted in a buffer contained 100mM Tris-HCl pH 6 . 8 , 100mM KCl , 0 . 1% Tween 20 , 0 . 2mM EDTA , 10% glycerol and 0 . 5 mg/ml HA peptide ( Sigma ) . HeLa . Scramble , HeLa . shUHRF1 expressing Flag-tagged UHRF1 or HeLa . shFANCD2 expressing Flag-tagged FANCD2 cells were treated with TMP/UVA as described . Cell pellets were incubated with Buffer A ( 0 . 1% Triton X-100 , 20mM Hepes pH7 . 9 , 5mM MgCl2 , 10% Glycerol , 1unit/μl Benzonase and 2 . 5mg/ml DSP ( D3669 , Sigma ) ) for 30 minutes on ice , and the Tris pH 8 . 0 was added to the mixture to 0 . 2M to quench DSP . 10 times pellet volume of Buffer B ( 0 . 15% Triton X-100 , 20mM Tris-HCl pH 8 . 0 , 5mM MgCl2 , 10% Glycerol , 300mM KCl and 0 . 2mM PMSF ) was added to the mixture and incubated for 10 minutes for extraction . Lysates were clarified by centrifugation , and supernatant was used for immunoprecipitation . M2 agarose beads was added to the lysates , and incubated for 2 hours . The resin was washed extensively , and eluted with 0 . 5mg/ml Flag peptide . EMSA was performed as previously described [37] with the following modifications: The binding reaction that contained 1 μg of UHRF1 or UHRF2 and 1 nM radiolabeled DNA , was performed in 10μl containing 25mM Tris-HCl pH 8 . 0 , 100mM NaCl , 6% glycerol , 1mM dithiothreitol ( DTT ) , 5ng poly ( dI·dC ) -poly ( dI·dC ) and 1μg bovine serum albumin ( BSA , New England Biolabs ) . For super-shift 2μg anti-HA antibody was added . Cells ( 250–4 , 000 ) were plated in 6-well plates and treated with different dosages of the indicated damaging agents on the next day . For TMP/UVA treatment , the cells were treated with 50ng/ml 4 , 5′ , 8-trimethylpsoralen ( TMP ) for 30 minutes , and irradiated with the UVA dosages indicated . Colony formation was scored after 10–14 days using 1% ( w/v ) crystal violet in methanol . Cells were scraped off the dishes , and centrifuged at 1 , 000 rpm for 5 minutes . Cell pellets were resuspended and incubated in equal volume of Benzonase buffer ( 2mM MgCl2 , 20mM Tris pH 8 . 0 , 10% glycerol , 1% Triton X-100 and 12 . 5units/ml Benzonase ( E1014 , Sigma ) on ice for 10 minutes . The cells were then lysed by the addition of an equal volume of 2% SDS to reach a final concentration of 1% . Samples were heated at 70°C for 2 minutes . The protein concentration was determined by Bradford assay ( Bio-Rad Life Science ) . EGFP-fused FANCD2 , EGFP-fused UHRF2 , and mCherry-fused UHRF1 cDNA were inserted into the pOZ vector as described above . Live-cell imaging was carried out with an OLYMPUS IX81 microscope connected to PerkinElmer UltraView Vox spinning disk system equipped with a Plan-Apochromat 60x/1 . 4 oil objective using Volocity software 6 . 3 for image capturing . EGFP and mCherry were excited with 488 nm and 561 nm laser lines , respectively . Throughout the experiment , these cells were maintained at 5% CO2 , and 37°C using a live cell environmental chamber ( Tokai hit ) . Confocal image series were typically recorded with a frame size of 512x512 pixels and a pixel size of 139 nm . For localized DNA damage induction , cells were seeded in glass bottom dish ( MatTek ) and sensitized by incubation in DMEM supplemented with 2 . 5% FBS and 20 μg/ml 4 , 5′ , 8-trimethylpsoralen ( TMP ) for 30 min at 37°C . Microirradiation was performed using the FRAP preview mode of the Volocity software by scanning ( each irradiation time was 100 ms ) a preselected area ( 50x3 pixels ) within the nucleus 20–75 times with a 405nm laser set to 100% laser power . The mCherry and EGFP intensities at microirradiated sites were quantified using ImageJ with Fiji , and normalized by their intensities before microirradiation . 1 μg of DNA in 200μl of water was added to 200μl of hydrolysis solution ( 100mM NaCl , 20mM MgCl2 , 20mM Tris pH 7 . 9 , 1000U/ml Benzonase , 600mU/ml Phosphodiesterase I , 80 U/ml Alkaline phosphatase , 36 μg/ml EHNA hydrochloride , 2 . 7mM deferoxamine ) . The mixture was incubated for two hours and then lyophylised by SpeedVac . The lyophylisate was resuspended in 1000μl of buffer A and 300μl was transferred into an LC-MS vial for analysis . A sample 100 times more dilute was prepared by dilution 5 μl of the original sample into 495 μl of Buffer A . For the analysis by HPLC–QQQ mass spectrometry , a 1290 Infinity UHPLC was fitted with a Zorbax Eclipse plus C18 column , ( 1 . 8μm , 2 . 1mm 15mm; Agilent ) and coupled to a 6495a Triple Quadrupole mass spectrometer ( Agilent Technologies ) equipped with a Jetstream ESI-AJS source . The data were acquired in dMRM mode using positive electrospray ionisation ( ESI1 ) . The AJS ESI settings were as follows: drying gas temperature 230°C , the drying gas flow 14 lmin-1 , nebulizer 20 psi , sheath gas temperature 400°C , sheath gas flow 11 lmin-1 , Vcap 2 , 000 V and nozzle voltage 0 V . The iFunnel parameters were as follows: high pressure RF 110 V , low pressure RF 80 V . The fragmentor of the QQQ mass spectrometer was set to 380 V and the delta EMV set to +200 . The gradient used to elute the nucleosides started by a 5-min isocratic gradient composed with 100% bufferA ( 10 mM ammonium acetate , pH 6 ) and 0% buffer B ( composed of 40% CH3CN ) with a flow rate of 0 . 400 ml min-1 and was followed by the subsequent steps: 5–8 min , 94 . 4% A; 8–9 min , 94 . 4% A; 9–16min 86 . 3% A; 16–17 min 0% A; 17–21 min 0% A; 21–24 . 3 min 100% A; 24 . 3–25min 100%A . The gradient was followed by a 5min post time to re-equilibrate the column . The raw mass spectrometry data was analysed using the MassHunter Quant Software package ( Agilent Technologies , version B . 07 . 01 ) . The transitions and retention times used for the characterization of nucleosides and their adducts are summarized in S1 Table . For the identification of compounds , raw mass spectrometry data was processed using the dMRM extraction function in the MassHunter software . For each nucleoside , precursor ions corresponding to the M-H+ and M-Na+ species were extracted , and the average of the signal observed from each target ion weighted by response was used for quantification . To utilise the linear range of for each nucleoside , the quantifications of dC , dG and dA were carried out with the diluted samples and quantification of dT and mdC was carried out with the concentrated sample .
|
Fanconi Anemia is a genetic disease where patients typically have congenital abnormalities , develop bone marrow failure and suffer from cancer predisposition . The cells in patients have a reduced ability to repair a type of DNA damage where the two strands of the DNA double helix are physically linked together , and this failure in repair is believed to contribute to cause of the disease . Many proteins are involved in repairing this type of DNA damage in healthy individuals , via a complex DNA repair pathway called the Fanconi Anemia pathway . Here , we report the identification of a new player in this pathway . The protein , called UHRF2 , is able to sense the DNA damage in the genome , and thereby help to initiate the healthy repair of the damage . In addition to improving our molecular understanding of the Fanconi Anemia pathway , in the long term , this new knowledge could have medical implications for diagnosis and therapy relating to pathologies involving this type of DNA damage .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"chemical",
"characterization",
"protein",
"interactions",
"hela",
"cells",
"enzymes",
"cell",
"cycle",
"and",
"cell",
"division",
"biological",
"cultures",
"cell",
"processes",
"enzymology",
"in",
"vivo",
"imaging",
"dna",
"damage",
"immunoprecipitation",
"cell",
"cultures",
"dna",
"co-immunoprecipitation",
"ligases",
"research",
"and",
"analysis",
"methods",
"imaging",
"techniques",
"protein-protein",
"interactions",
"proteins",
"cell",
"lines",
"binding",
"analysis",
"precipitation",
"techniques",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"cultured",
"tumor",
"cells"
] |
2018
|
Identification of UHRF2 as a novel DNA interstrand crosslink sensor protein
|
We sought to evaluate the relationship between onchocerciasis prevalence and that of epilepsy using available data collected at community level . We conducted a systematic review and meta-regression of available data . Electronic and paper records on subject area ever produced up to February 2008 . We searched for population-based studies reporting on the prevalence of epilepsy in communities for which onchocerciasis prevalence was available or could be estimated . Two authors independently assessed eligibility and study quality and extracted data . The estimation of point prevalence of onchocerciasis was standardized across studies using appropriate correction factors . Variation in epilepsy prevalence was then analyzed as a function of onchocerciasis endemicity using random-effect logistic models . Eight studies from west ( Benin and Nigeria ) , central ( Cameroon and Central African Republic ) and east Africa ( Uganda , Tanzania and Burundi ) met the criteria for inclusion and analysis . Ninety-one communities with a total population of 79 , 270 individuals screened for epilepsy were included in the analysis . The prevalence of epilepsy ranged from 0 to 8 . 7% whereas that of onchocerciasis ranged from 5 . 2 to 100% . Variation in epilepsy prevalence was consistent with a logistic function of onchocerciasis prevalence , with epilepsy prevalence being increased , on average , by 0 . 4% for each 10% increase in onchocerciasis prevalence . These results give further evidence that onchocerciasis is associated with epilepsy and that the disease burden of onchocerciasis might have to be re-estimated by taking into account this relationship .
Recent surveys , based on rapid epidemiological mapping of onchocerciasis ( REMO ) , have revealed that the number of people infected with Onchocerca volvulus , the causative agent of onchocerciasis , had been largely underestimated . According to the latest REMO results obtained as part of the African programme for onchocerciasis control ( APOC ) , 37 million are estimated to be infected with O . volvulus in Africa [1] . Accordingly , the Disability Adjusted Life Years ( DALYs ) lost per year due to onchocerciasis is estimated to be 1 . 49 million . This number is threefold higher than the previous figure published 2 years earlier [2] but probably still underestimates the real burden of disease attributable to onchocerciasis . The current DALY estimates account for blindness , visual impairment and onchocercal itching due to skin disease , yet it does not account for significant excess mortality due to heavy infections with O . volvulus [3] . The excess mortality of sighted individuals with high microfilarial loads suggests the operation of insidious and systemic involvement , neurologic involvement having been proposed as a possible candidate [4] . In particular , a link between onchocerciasis and epilepsy has been suspected since the 1930s [5] . Population-wide surveys found significant correlations between prevalence of epilepsy and that of onchocerciasis in Uganda [6] and between prevalence of epilepsy and community microfilarial load ( CMFL ) of onchocerciasis in Cameroon [7] . However , case-control studies evaluating the association between epilepsy and onchocerciasis reached diverse conclusions: significant association was demonstrated in areas where onchocerciasis prevalence exceeded 60% [7] , [8] , whereas no significant relationship was found in areas with lower onchocerciasis prevalence [9] , [10] , [11] . The present work aims at testing the hypothesis that in communities with high onchocerciasis endemicity , the prevalence of epilepsy will clearly exceed that found in communities of low onchocerciasis endemicity . We conducted a literature review of epidemiological studies addressing the issue of the onchocerciasis-epilepsy relationship and performed a meta-regression analysis including available population-based data to quantify the influence of onchocerciasis endemicity on the prevalence of epilepsy .
We searched PubMed and ISI Web of Knowledge up to February 2008 , with neither past time limit nor language restriction , to identify population-based studies reporting on the prevalence of epilepsy in communities for which onchocerciasis prevalence was available or such prevalence could be estimated . We entered the following search terms and Boolean operators , for matches under any field: epilep* AND onchocerc* , epilep* AND country with country being iteratively one of the 11 OCP countries , 19 APOC countries , six OEPA countries , and Yemen . Titles and available abstracts were scanned for relevance , identifying papers requiring further consideration . Bibliographies of relevant articles were checked . Relevant theses and reports were also searched at the library of the Institut d'Epidémiologie et de Neurologie Tropicale ( IENT , Limoges ) . The authors who initiated the study ( SDSP & MB ) also contacted researchers who were requested to provide information and/or data not included in the articles or other documents . Studies were selected if both epilepsy prevalence and an indicator of onchocerciasis prevalence were available or such prevalence could be calculated . Inclusion criteria for subsequent analysis were set to incorporate studies: ( 1 ) carried out following a population-based design; ( 2 ) providing information on the methods used to diagnose epilepsy and onchocerciasis and ( 3 ) in which epilepsy prevalence was assessed in the general population , i . e . both in children and adults . Study communities with a sample of less than 10 subjects were discarded . Two authors ( SDSP , MB ) independently assessed eligibility and study quality , and extracted data . We recorded all basic parasitological and demographic information from each eligible study into a purpose-built database . The extracted data included demographic characteristics of the population examined ( age range and sex ) , recruitment methods , and number and dates of previous community treatments with ivermectin . For onchocerciasis , specific information was recorded on methods used for parasitological examination . For epilepsy , details on sampling procedures , and definition of epilepsy were recorded . To quantify the extent to which epilepsy prevalence is associated with onchocerciasis endemicity across the different studies , a meta-regression was performed . Epilepsy prevalence was defined as the outcome and onchocerciasis prevalence as the explanatory variable . We used a logistic model to assess the relationship between prevalence of epilepsy and that of onchocerciasis . A random effect , capturing between-studies heterogeneity was subsequently incorporated into the model previously outlined . Significance of this effect was tested using the likelihood ratio test [12] . Logistic models were first estimated using the raw data collected during the review of literature . Because we had some reasons to think that some of the published data may have suffered from methodological bias , we developed some correction factors to standardize the data ( see below ) . Logistic models ( fixed-effect then random-effect models ) were then estimated from the corrected data . In addition , in order to test whether the relationship between prevalence of epilepsy and that of onchocerciasis was influenced by a specific study , each study was successively omitted from the whole database and the parameters re-estimated . Parameters of the different models were estimated using the non-linear regression procedure ( NLMixed ) provided in the SAS v8 . 1 software . This procedure provides Bayes empirical estimates of the study-specific random-effect [13] . The prevalence of onchocerciasis was considered to have been measured by a standard procedure if it has been estimated in the general population ( ≥5 years old ) using the onchocerciasis diagnostic method used by the Onchocerciasis Control Programme in West Africa ( OCP ) : this entails taking a skin biopsy from each iliac crest , using a 2 mm Holth-type punch , and incubating it for 24 h in normal saline before searching for the presence of O . volvulus microfilariae [14] . Five of the eight studies included in the analysis [6] , [8] , [15] , [16] , [17] used a method for the diagnosis of infection with O . volvulus deviating from the OCP standard procedure . Table 1 summarises the types of deviation and the rationale for correction of the respective onchocerciasis prevalence values . Correction factors were determined using information from published comparative studies . Details on these standardisation processes are presented in the Supporting Information to this article . In addition , three studies were carried out in areas where ivermectin treatment campaigns had been performed [6] , [18] , [19] , which affected the measure of the prevalence of O . volvulus microfilaridermia . By using published studies on the effect of community treatment with ivermectin we developed a model to determine appropriate correction factors allowing us to estimate the pre-treatment prevalence according to the number of preceding treatment rounds ( details given in Supporting Information ) . In the present paper , we assumed that if epilepsy were associated with onchocerciasis , it might be due not only to the presence of parasites in the cerebral tissue , but also to cicatricial lesions persisting after the disappearance of the parasites after a treatment . This is the case for epilepsy induced by other infectious diseases ( e . g . malaria [20] ) . Thus , we assumed that , in the present study , ivermectin treatments had little or no effect on the epilepsy of those patients who already suffered from this condition . Ivermectin treatments might decrease incidence of epilepsy but this effect would probably be perceptible after a number of years . This is supported by observations made by Kaiser et al . [21] who found that no major change in epilepsy incidence was observed over 4 years after the start of annual ivermectin mass treatment . Four of the 8 studies included in the present analysis referred to the definition of epilepsy proposed by the International League Against Epilepsy ( ILAE ) in 1993 [22] of “two or more unprovoked seizures during the previous 2 years” whereas the remaining 4 studies used deviating definitions ( Table 2 ) . Characteristics of neuro-epidemiological methods as used in the different studies are presented in Table 2 . Sampling procedures , methods for case identification and confirmatory examinations were not uniform and incompletely described in some studies . One study published by Kipp et al . [19] reported a particularly high epilepsy prevalence of 8% in the Kabarole district of western Uganda . However , for the diagnosis of epilepsy this study was confined to a rapid assessment of cases with suspected epilepsy without a more extensive confirmatory examination . A later survey conducted in the same onchocerciasis focus confirmed the diagnosis in only 61 ( 54% ) out of 113 cases with suspected epilepsy [6] . If this ratio of confirmed over suspected epilepsy patients is applied to the survey of Kipp et al . [19] , the crude epilepsy prevalence in the latter study should be revised from 8% to 4 . 3% ( 8%×0 . 54 = 4 . 3% ) .
From the 1752 examined abstracts and research in the IENT library , ten different studies reporting population-based surveys on onchocerciasis and epilepsy prevalence were identified [6] , [7] , [8] , [11] , [15] , [16] , [17] , [18] , [19] , [23] ( Table 2 ) . One study was excluded because it originated from an area where control measures had been carried out over 20 years within the Onchocerciasis Control Programme ( OCP ) and the overall prevalence had decreased to a low endemicity level at the time of the survey [10] . A second study focusing on the clinical description of a series of epilepsy patients in an onchocerciasis endemic area reported only an estimate of epilepsy prevalence in the study area and was also not considered [23] . A flowchart summary of the search is shown in Figure 1 . Data meeting the inclusion criteria consisted of 91 communities from 8 distinct areas ( 2 in West Africa: Benin [16] and Nigeria [18]; 2 in Central Africa: Cameroon [7] and Central African Republic ( CAR ) , [15] and 4 in East Africa: Uganda ( 2 sites ) [6] , [19] , Tanzania [17] and Burundi [8] , [24] ) . The different surveys took place between 1987 and 1997 . Overall , 12 388 individuals were examined for O . volvulus infection and a total of 905 subjects with epilepsy were identified out of a screened population comprising 79 270 individuals ( weighted average epilepsy prevalence: 1 . 14% ) . Epilepsy and onchocerciasis prevalences were assessed from the same population sample in two studies [16] , [18] whereas in the six remaining studies [6] , [7] , [8] , [15] , [17] , [19] this was done on different population samples in each community . Figure 2 represents epilepsy prevalence versus onchocerciasis prevalence , as reported in the documents retrieved during the review of literature , for the 91 study communities included in the analysis . Figure 3 represents the same data after applying the correction factors to minimize obvious biases . Epilepsy prevalence ranged from 0 to 8 . 7% and that of onchocerciasis from 1 . 4 to 100% . The fixed-effect logistic models fitted on observed and corrected data sets indicate a significant association ( all parameters with p<0 . 0001 ) between onchocerciasis prevalence and epilepsy prevalence ( Table 3 ) . Inclusion of a random-effect resulted in a statistically significant improvement of the models ( likelihood ratio tests p<0 . 0001 for both models , Table 3 ) . The random-effect logistic model assessed on the corrected data provided slightly lower estimates than when assessed from observed data , with the respective corresponding odds-ratio: 1 . 042 ( 95%CI: 1 . 034–1 . 05 ) and 1 . 044 ( 95%CI: 1 . 036–1 . 052 ) . These values indicate that , on average , epilepsy prevalence is increased by 0 . 4% for a 10% increase in onchocerciasis prevalence . However , as indicated by varying values of the random effect parameter u ( Table 4 ) , the influence of onchocerciasis on epilepsy differed between the studies . In particular , the random-effect model fitted on corrected data suggested that the influence of onchocerciasis on epilepsy was higher in the Nigeria study; conversely , it was lower in the Cameroon and Central African Republic studies ( Table 4 ) . Sensitivity analysis showed that the significance of the association between onchocerciasis prevalence and that of epilepsy was not affected by omission of any of the studies ( results not shown ) .
The present study was carried out in order to evaluate whether the epilepsy prevalence in communities living in O . volvulus endemic areas is related to the prevalence of onchocerciasis . All available community-based surveys on this subject were used . Throughout 91 communities distributed across 7 African countries , variation of epilepsy prevalence is associated with that of onchocerciasis prevalence . A recent review took an approach different from ours to analyse epidemiological studies searching for a relationship between onchocerciasis and epilepsy [25] . This review examined studies from which it was possible to calculate the relative risk of epilepsy in patients being infected with O . volvulus compared to the risk of infection of inhabitants without epilepsy from the respective study area . Inconsistent results were observed for nine African studies providing sufficient data , with their respective relative risk calculated at a broad range from 0 . 84 to 6 . 80 . The common relative risk of 1 . 21 ( 95% CL 0 . 99–1 . 47 ) for all studies was close to the threshold of significance , suggesting a probable but weak risk of epilepsy in onchocerciasis patients . Possibly , this review did not yield a more pronounced association because it included studies which were not originally designed as case-control studies . In areas of low endemicity , the possible effect of onchocerciasis on epilepsy in such an analysis will also be masked by the relatively higher proportion of epilepsy cases due to alternative aetiologies . A major difference between the review of Druet-Cabanac et al . [25] and our analysis is that we introduced the quantitative dimension of O . volvulus infection by defining the onchocerciasis prevalence in a village as the risk factor . It has been demonstrated that onchocerciasis prevalence is closely related to the intensity of infection through a negative binomial relationship [26] . The mathematical properties of this relationship implies that the higher the prevalence , the higher the proportion of individuals with heavy microfilarial loads in the population . In a previous study [7] , intensity of infection expressed as the microfilarial loads of patients with epilepsy was 2 to 3 fold higher than those of controls matched on age , sex , and village of residence . These results suggest that , as it has been found for ocular onchocercal pathology [27] , intensity of infection with O . volvulus is a key factor in the induction of epilepsy . This would provide a further explanation as to why the results of the previous review [25] gave a weaker degree of significance to the association between epilepsy and onchocerciasis than those obtained in our study . The random effect included in the modelling indicates that the influence of onchocerciasis on epilepsy varied between the studies . This heterogeneity can be due to either differences in the methodology used to assess the prevalence of onchocerciasis and/or that of epilepsy , or to true biological differences modulating the epilepsy/onchocerciasis association . In this respect , there is some indication that the onchocerciasis prevalence from the Nigerian study [18] may have been compromised by inadequate correction . We were not able to determine with precision the number of ivermectin treatments administered before the parasitologic survey , nor the time interval between the last ivermectin treatment and the survey . We assumed that the parasitological assessment had been conducted at least 10 months after the last treatment , which is a sufficient lapse of time for the majority of infected patients to present again with microfilariae in the skin . If the precise interval were shorter than our assumption , the true initial prevalence of onchocerciasis would have been higher than that used in our analysis . However , if one adds arbitrarily 20% to all prevalence rates in this survey , one obtains similar estimates for the random-effect model ( a = −7 . 174 ( S . E . = 0 . 301 ) ; b = 0 . 041 ( 0 . 003 ) ; -2Log Likelihood = 688 . 6 ) , with the specific random-effect of this study being no more significant ( P = 0 . 75 ) . In the Cameroonian study site , a 6-fold increase was found in the mortality rate among individuals with epilepsy compared to that in control individuals [28] . This significant mortality rate may have resulted in an underestimation of epilepsy prevalence when it was assessed through a cross-sectional survey . Should the mortality rate of people with epilepsy be lower in the Nigerian area , epilepsy prevalence in the latter would have been overestimated compared to those obtained in Cameroon . Lastly , we do not discard the possibility that other factors , such as O . volvulus strains with differing pathogenic potential [29] or coinfection with other pathogens , modulate the association between epilepsy and onchocerciasis . The assessment of epilepsy prevalence in the various studies of the present analysis may have been influenced by the occurrence of other endemic diseases known to be involved in the aetiology of epilepsy [30] . Supposing a homogeneous distribution in the area of such a competing aetiological factor , the increase of epilepsy prevalence would be expected to comprise areas of low and high onchocerciasis prevalence to the same extent but would not affect an existing true association between epilepsy and onchocerciasis . In contrast , a co-endemic disease following a local distribution similar to that of onchocerciasis could reinforce ( or weaken ) this association . Neurocysticercosis caused by cerebral cysts of the pork tapeworm ( Taenia solium ) is considered a frequent cause for epilepsy in many sub-Saharan regions [30] , [31] and it has been suggested that , in areas co-endemic for onchocerciasis and cysticercosis , epilepsy patients could be mistaken as having onchocerciasis-related epilepsy when they may be suffering from neurocysticercosis-derived epilepsy [32] . As far as information is available for the sites concerned in our analysis , cysticercosis was found endemic in the study areas in Burundi and Cameroon [7] , [8] , [33] , but case-control studies from both these sites found a positive correlation between epilepsy and onchocerciasis . In one study from Uganda , serologic tests produced no evidence of a significant infestation of the population with T . solium [6] . In view of the results of the present analysis , we do not consider that neurocysticercosis could be a confounding factor that would produce a significant increase in epilepsy prevalence with increasing onchocerciasis prevalence . For this to be the case , it would be required that the distribution of neurocysticercosis ( and thus the pig raising habits of the local communities ) would closely follow the prevalence of onchocerciasis in the study areas of all seven countries involved . An extensive literature comparing the performance of the various diagnostic methods of O . volvulus infection allowed us to develop appropriate correction factors for onchocerciasis prevalence . However , no such information is available for epilepsy that would have enabled us to do so for epilepsy prevalence . Strategies to assess prevalence of epilepsy in developing countries were developed in the 1980s [22] , [30] , [34] , [35] . Ideally , such studies should follow a two-step protocol beginning with a population-wide door-to-door survey in order to ( i ) provide a complete census defining the sampling frame , ( ii ) assess epidemiologically relevant characteristics of the population and ( iii ) identify patients with possible epilepsy by use of a sensitive screening questionnaire . In a second step , patients with possible epilepsy are then subjected to a neurological and medical examination allowing confirmation or rejection of the diagnosis . Only two studies included in the present analysis can be regarded as having followed the recommended protocol [6] , [18] , but even these did not meet all requirements in that they did not use a pre-tested sensitive screening tool for patient identification but relied on a single screening question . This may have led to an underestimation of epilepsy prevalence in these studies . Procedures used for testing the population and identifying possible epilepsy patients , and for confirming diagnosis varied across the other studies and in a few instances these were not clearly specified ( Table 2 ) . Comparison between studies was also made difficult because no uniform definition for epilepsy was used . The relative weakness in neurological methods can in part be explained by the fact that most studies were initiated by researchers involved in onchocerciasis control measures and by the general scarcity of neurological expertise available in the endemic areas [34] , [36] . The diversity of neuro-epidemiological methods may have influenced epilepsy prevalence obtained in the different studies . For instance , the use of community key informants for case identification may have resulted in an underestimation [37] , whereas the lack of a confirmatory examination would have given a falsely high prevalence . Although we have to consider these effects as substantial , we were unable to express them quantitatively and our attempts to adjust epilepsy prevalence data were restricted to the study of Kipp et al . [19] . However , despite disparate neuro-epidemiological methods in use , we found epilepsy prevalence closely related to onchocerciasis prevalence throughout different African regions . It may be expected that this finding would be even more significant if methods with better accuracy were applied . This is the first time that the relationship between the prevalences of onchocerciasis and epilepsy has been quantitatively assessed using available data collected at community level to perform adequate statistical analysis . We found that in areas where onchocerciasis is endemic , epilepsy prevalence increases with onchocerciasis prevalence . This is in accordance with the results of case-control studies and supports earlier anecdotal reports of numerous researchers working in various endemic areas [38] , [39] . Because onchocerciasis is a disease affecting remote areas of countries with limited resources for neurological research , some of the studies carried out so far are influenced by methodological shortcomings . Future studies should use established and comparable protocols for neurological research . Possible research questions to be addressed with available and appropriate methods are: ( i ) the confirmation of the hitherto existing findings in other endemic areas; ( ii ) the role of intensity of infection in inducing epilepsy in areas not yet exposed to antifilarial treatment; ( iii ) the clinical characterization and classification of epilepsy in the various endemic areas; and ( iv ) the effect of long term onchocerciasis control on epilepsy incidence and prevalence .
|
Epilepsy is particularly common in tropical areas . One main reason is that many endemic infections have neurological consequences . In addition , the medical , social and demographic burden of epilepsy remains substantial in these countries where it is often seen as a contagious condition and where the aetiology is often undetermined . For several decades , field researchers had reported some overlapping between the geographical distributions of epilepsy and onchocerciasis , a parasitic disease caused by the filarial worm Onchocerca volvulus which afflicts some 40 million persons worldwide . Here , we conducted a statistical analysis of all the data available on the relationship between the two conditions to determine whether the proportion of people suffering from epilepsy in a community could be related to the frequency of onchocerciasis . The combined results of the eight studies carried out in west , central and east Africa indicate a close epidemiological association between the two diseases . Should a causative relationship be demonstrated , onchocerciasis , which is known as “river blindness” because of its most serious sequela and the distribution of its vectors , could thus also be called “river epilepsy” . More research is needed to determine the mechanisms explaining this association and to assess the burden of onchocerciasis-associated epilepsy .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/helminth",
"infections",
"neurological",
"disorders/epilepsy",
"infectious",
"diseases/neglected",
"tropical",
"diseases"
] |
2009
|
Epilepsy in Onchocerciasis Endemic Areas: Systematic Review and Meta-analysis of Population-Based Surveys
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.